Application Design Showcase #2

The Best Complex, Lightweight, Workflow, and Multi-Platform Applications

Kathryn Whitenton, Kara Pernice, Patty Caya, and Jakob Nielsen

WWW.NNGROUP.COM 48105 WARM SPRINGS BLVD., FREMONT CA 94539–7498 USA

Copyright © by Nielsen Norman Group, all rights reserved.

To buy a copy, download from: http://www.nngroup.com/reports/applications/design
 This page is blank for double-sided printing.

2 [email protected]
Contents

Executive Summary ..................................................................... 10

Overview of the Winners ............................................................. 16

Screenshots are Just a Moment in Time ...................................... 18

Design Trends ............................................................................. 19

Designing for different Types of Users ....................................................... 19

Making Top Tasks Obvious for All ............................................................. 19

Presenting Various Options and Views to Meet Needs .................................. 20

Multiple Paths ........................................................................................ 21

Preferences and Customization ................................................................ 21

Wizards................................................................................................. 21

Linked Controls ...................................................................................... 22

Downloading or Sending Data to Another App to Review and Manipulate ....... 22

Making the Most of Space ........................................................................ 23

Safety Nets ........................................................................................... 23

Help (inline) .......................................................................................... 24

Complex Applications ............................................25

BondWorks™ Advisor Platform .................................................... 26

Overview............................................................................................... 26

Users .................................................................................................... 26

The Application ...................................................................................... 27

Design Process....................................................................................... 39

Usability Methods..............................................................................................39

Lessons Learned .................................................................................... 45

WWW.NNGROUP.COM 48105 WARM SPRINGS BLVD., FREMONT CA 94539–7498 USA

Copyright © by Nielsen Norman Group, all rights reserved.

To buy a copy, download from: http://www.nngroup.com/reports/applications/design
 Team .................................................................................................... 46

GEN2i .......................................................................................... 49

Overview............................................................................................... 49

Users .................................................................................................... 50

The Application ...................................................................................... 50

Design Process....................................................................................... 59

Usability Methods..............................................................................................59

Development Approach......................................................................................63

Timeline ..........................................................................................................64

Lessons Learned .................................................................................... 65

Team .................................................................................................... 66

OEConnection .............................................................................. 67

Overview............................................................................................... 67

Users .................................................................................................... 68

The Application ...................................................................................... 68

Design Process....................................................................................... 80

Tracking the Design Evolution ............................................................................81

Lessons Learned .................................................................................... 85

Project Team ....................................................................................................86

Lightweight Applications .......................................89

CycleStreets ................................................................................ 90

Why They Were Chosen .......................................................................... 90

Overview............................................................................................... 90

Who Are the Users? ................................................................................ 91

The Application ...................................................................................... 91

Innovation............................................................................................. 93

4 [email protected] Table of Contents

 Consistency ........................................................................................... 95

Flexibility .............................................................................................. 97

Information Display .............................................................................. 102

Design Process..................................................................................... 103

Implementation .............................................................................................. 103

JQuery Mobile Framework ................................................................................ 104

Standalone Code............................................................................................. 104

Lessons Learned .................................................................................. 104

Team .................................................................................................. 105

VitreaTeach ............................................................................... 106

Overview............................................................................................. 106

Who are the Users? .............................................................................. 106

The Application .................................................................................... 107

Innovation........................................................................................... 109

Flexibility ............................................................................................ 112

Information Display .............................................................................. 116

Interface Elements ............................................................................... 119

Design Process..................................................................................... 120

Expert Review ................................................................................................ 121

Implementation .............................................................................................. 122

Lessons Learned .................................................................................. 123

Team .................................................................................................. 123

Wakelicious ............................................................................... 124

Overview............................................................................................. 124

Who are the Users? .............................................................................. 124

The Application .................................................................................... 124

Simplicity ............................................................................................ 127

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 5

 Interface Elements ............................................................................... 134

Design Process..................................................................................... 136

Usability Testing ............................................................................................. 137

Implementation .............................................................................................. 138

Lessons Learned .................................................................................. 139

Team .................................................................................................. 140

Workflow Applications ........................................141

Climate Action Planning Tool (NREL) ......................................... 142

Overview............................................................................................. 142

Who are the Users? .............................................................................. 142

The Application .................................................................................... 143

Task Flow ............................................................................................ 144

Flexibility ............................................................................................ 150

Information Display .............................................................................. 154

Design Process..................................................................................... 158

Usability Testing ............................................................................................. 159

Implementation .............................................................................................. 159

Lessons Learned .................................................................................. 160

Team .................................................................................................. 160

eReview (IMF) ........................................................................... 163

Overview............................................................................................. 163

Users .................................................................................................. 164

The Application .................................................................................... 164

Design Process..................................................................................... 172

Using the SharePoint 'Out of the box' UI ............................................................ 172

Feeling the Pain .............................................................................................. 173

6 [email protected] Table of Contents

 Redesigning the Interface ................................................................................ 173

Lessons Learned .................................................................................. 175

Team .................................................................................................. 176

Eventbrite “Create” Workflow ......................................................... 179

Overview............................................................................................. 179

Who are the Users? .............................................................................. 180

The Application .................................................................................... 180

Task Flow ............................................................................................ 182

Flexibility ............................................................................................ 192

Information Display .............................................................................. 195

Interface Elements ............................................................................... 199

Design Process..................................................................................... 201

Workshops with Customer Support ................................................................... 202

Exploratory Interviews..................................................................................... 203

Usability Testing ............................................................................................. 204

Lessons Learned .................................................................................. 206

Team .................................................................................................. 206

Product Suggestion Portal (Hobsons)........................................ 208

Overview............................................................................................. 208

Who are the Users? .............................................................................. 209

The Application .................................................................................... 209

Design Process..................................................................................... 226

Usability Methods............................................................................................ 226

Lessons Learned .................................................................................. 230

Team .................................................................................................. 231

Multi-Platform Applications .................................233

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 7

 Dixio Multiplatform Lookup Technology..................................... 234

Overview............................................................................................. 234

Users .................................................................................................. 234

The Applications ................................................................................... 235

Multiple Media ................................................................................................ 235

Dixio Desktop ................................................................................................. 236

Web ........................................................................................................... 242

Dixio eReader Solution .................................................................................... 243

Design Process..................................................................................... 246

Implementation .............................................................................................. 247

Lessons Learned .................................................................................. 250

Team .................................................................................................. 251

Highlight.................................................................................... 253

Why They Were Chosen ........................................................................ 253

Overview............................................................................................. 253

Who are the Users? .............................................................................. 253

The Application .................................................................................... 254

Innovation........................................................................................... 255

Consistency ......................................................................................... 259

Flexibility ............................................................................................ 263

Information Display .............................................................................. 265

Interface Elements ............................................................................... 269

Design Process..................................................................................... 274

Usability Testing ............................................................................................. 274

Beta Testing ................................................................................................... 274

Implementation .............................................................................................. 274

Lessons Learned .................................................................................. 276

8 [email protected] Table of Contents

 Team .................................................................................................. 277

Mobile Unified Communications (UC) Mobile Client (Verizon) ... 278

Overview............................................................................................. 278

Users .................................................................................................. 278

The Application .................................................................................... 279

Design Process..................................................................................... 291

From Idea to Execution.................................................................................... 291

Usability Methods............................................................................................ 292

Designing to the Least Common Denominator (in PBX systems) ........................... 294

Project Timeline .............................................................................................. 296

Lessons Learned .................................................................................. 297

Team .................................................................................................. 298

Ten Application Design Mistakes ............................................... 300

1. Steps in a task or process presented out of order on a page or

screen ...................................................................................... 300

2. Missing Simple Paths .................................................................. 300

3. Invisible Controls and Cryptic Icons .............................................. 300

4. Hideous Forms ........................................................................... 301

5. Unclear Differences When Comparing Choices ................................ 301

6. indistinguishable Current Selection Status ..................................... 301

7. Branding Features or Menu Commands ......................................... 302

8. Only the Promise of Content ........................................................ 302

9. Views and Modes That Trap People ............................................... 302

10. Bad Dialog ................................................................................ 302

About the Authors ........................................................................ 303

Acknowledgements ....................................................................... 305

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 9

 Executive Summary

The winners for best application user interfaces of the year are:

Complex Applications
• The web-based BondWorks wealth management platform (BondDesk
Group LLC) provides financial advisors with a streamlined system to
trade fixed-income assets.
• The portable GEN2i (Hottinger Baldwin Messtechnik GmbH, Germany)
lets users record engineering data in the field.
• OEConnection (OEConnection LLC) lets parts department employees
at auto dealerships and repair facilities process more than 6 million
original equipment parts transactions monthly.

Lightweight Applications
• CycleStreets Mobile (CycleStreets Ltd. and Anna Powell-Smith, UK)
helps cyclists plan safe, fast routes for cycle journeys using crowd-
sourced map data.
• VitreaTeach (Vital Images, Inc.) is an educational tool that radiology
residents and practicing radiologists can use to save, manage, and
share medical images.
• Wakelicious (Utopian Army, The Netherlands) is an alarm clock that
wakes you up by playing YouTube videos sent by your friends.

Workflow Applications
• The Climate Action Planning Tool (National Renewable Energy
Laboratory) helps organizations reduce their carbon emissions.
• eReview (International Monetary Fund, IMF) integrates complex
business functions into an organization’s intranet.
• The self-service Eventbrite (Eventbrite Inc.) software tool helps event
hosts organize, sell, and manage tickets for events. The Create Event
workflow, which won the award, helps users set up a new event within
Eventbrite’s online system.
• The Hobsons CRM (Hobsons, Inc.) product-suggestion portal tool lets
customers make comments and suggestions.

Multi-Platform Applications
• Dixio (Semantix Group SL, Spain) offers users multiplatform lookup
capabilities.
• Highlight (Cohdoo, LLC) lets users with iOS devices make audio
recordings, mark interesting points, and share recordings with others.
• The Mobile Unified Communication client (Verizon Wireless) lets
business customers link their mobile and business lines.

10 [email protected] Executive Summary

 In contrast to our other design awards, we decided to split the application award into
categories. It doesn't make sense to judge small consumer apps that do a single
thing together with huge enterprise applications that manipulate immense sets of
convoluted data—such as the 4 million different bonds that traders manage through
BondWorks. Workflow applications that integrate many steps or coordinate several
user roles have still different requirements, as do multi-platform apps that work
across several devices.
Solving a big problem is more difficult than solving a small problem. However, if your
app does only one thing, it had better to it really, really well. So, the criteria for
polish are stricter for small applications than for enterprise applications. With
thousands of screens in some applications, they can’t all be perfect, but the flow
between screens and the ability to select the correct feature for any given situation
must be great. So, instead of comparing apples and oranges, we decided to judge
each application style in relation to its peers and select the best designs in each
category.
That we would ever classify a radiology application as “lightweight” defies the
imagination, but it’s appropriate in the case of VitreaTeach because it focuses on one
thing: allowing residents and experienced radiologists to manage images for learning
purposes. We might even call it the Pinterest of medical science. But, of course,
Pinterest itself wouldn’t work for radiologists — partly because of the privacy
concerns for any medical information and partly because of the benefits that domain-
specific features provide.
For example, VitreaTeach both offers special access to cases from the user’s own
hospital and knows something about anatomy and pathology. Interestingly, one of
the things that make the design cleaner than other medical systems is the simple
tagging that provides much of the taxonomy support for the images. Designers
simplified tagging after user research found that full, formal anatomy and pathology
taxonomy hierarchies added too much overhead.

AUGMENTING HUMAN CAPABILITIES

Software’s real goal should not be to simply process transactions in a system where
users are nothing more than data operators who click required buttons to make
things happen. Rather, software should work to augment human capabilities, helping
us to overcome weaknesses and emphasize our strengths.
A vital way for computers to help users is to direct their attention to a smaller
number of important issues instead of overwhelming them with all possible options.
For example, BondWorks originally presented 85 attributes of tradable bonds on the
search results screen. However, search log analysis showed that nearly three
quarters of users were searching on only 10 of the attributes, so the screen was
redesigned to focus on those attributes.
OEConnection proactively takes the next step when a spare part is backordered from
the manufacturer: the software automatically proceeds to search other dealers’
inventory for the desired part. Traditionally, users would have had to issue each type
of search separately from separate application areas. But that’s exactly the approach
that turns humans into robotic button-ushers; better to turn the computer into a
problem-solving tool that understands the domain and lets users expend their
brainpower on higher-level issues.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 11

 As another example, the Highlight application’s basic feature is to make audio
recordings searchable by letting users add a marker when something interesting
occurs. For real-world tasks — such as conducting interviews or sitting in a foreign-
language class — users wouldn’t actually know that “something interesting” was
being recorded until after it happened. To accommodate this common scenario, the
application has an offset feature that automatically places the marker into the
recording a certain number of seconds before the button is pushed. Instant time
travel.
More to the point, Highlight uses the computer’s flexibility to augment human skills
by taking over the tedious job of rewinding the recording when users realize that
something interesting should have been marked. This reduces users’ cognitive load
and frees them to concentrate on listening to the event they’re recording instead of
attending to the device.
Another simple way Highlight augments human capabilities is in its ability to
associate a picture with each recording. This is particularly simple when doing
interviews, where users can snap a photo of the interviewee with the same phone
that’s recording the audio. These photos come in handy later when retrieving
recordings from long lists, because it’s often easier for our poor brains to associate
an interview with the person’s photo instead of relying on a textual description.
We also saw a trend to focus screen real-estate allocations on the information that’s
most important in the current context, while showing less UI chrome. More content,
less application overhead.
Some winning applications had special small control panels for direct access to
important operations that would usually have required a switch in context. This
doesn’t quite equate to a modeless design, because complex applications do tend to
have several distinct stages or areas of workflow and data. But surfacing crucial
features to live above the regular UI architecture is yet another way of empowering
users to act on their insights instead of delaying them by the rigid demands of the
way the application is structured.

LINEAR WORKFLOW AND WIZARDS

We want to empower users to be creative and accomplish advanced things with our
software. But we should also recognize that users sometimes just want to get their
tasks done without having to explore numerous options and new ideas.
To speed users through infrequent or complicated tasks, it’s often good to present a
linear workflow with minimal disruptions or alternatives. Yes, the lack of flexibility
can feel constraining, but it can be faster to just power through all the steps instead
of having to ponder which steps are needed. Also, the cost of too much freedom is
that users have to decide the order of the steps — something that they’re often
happy enough to delegate to the computer.
Wizards are the classic approach to linear workflows, and several winning
applications include nicely designed wizards. The Climate Action Planning Tool uses
the simpler approach of offering numbered steps to guide users. An early design
included a Before You Start step, but user studies showed that people ignored it. In
the final design, the Before activities appear under Step 1: Gather Baseline Energy
Consumption Data. Users just want to get going, so you need to be somewhat
heavy-handed if setup work is required before starting the real activity.

12 [email protected] Executive Summary

 Wizards don’t work for everything. In user testing, Eventbrite developers discovered
problems with a wizard for creating new events in the system. Because users’ mental
models of setting up a new event didn’t include all the necessary steps, they often
dropped off and didn’t complete the wizard. So, the team decided instead to go with
a 2-step process in which users first entered all the information and then customized
their pages.
When creating a linear workflow, take care not to signal to users that they can ignore
subsequent steps — unless it’s true. For example, after the first 6 steps, Eventbrite’s
workflow was as follows:
• Step 7: Change the color scheme on the registration page.
• Step 8: Choose privacy settings.
It might have seemed logical to first fully create the thing that the privacy would
apply to, and then to choose the privacy settings. To some users, however, Step 7
signaled that all the “real” set-up was done and that the remaining steps would be
frivolous decorating. If they didn’t care about the decor or were satisfied with the
default colors, some people simply stopped at Step 7 and never got to the privacy
options.

LIVING IN AN ECOSYSTEM
A major trend among the winning applications is a modularized approach to the total
user experience, with major features outsourced to other applications. Designers
accept that their product will live in an ecosystem of other applications (and
websites) and that users will often prefer to do certain things elsewhere in that
ecosystem.
Relying on other applications to perform work for your app is obviously not a new
idea. Having applications work together goes back at least to the pipes in Unix. But
we see this done more often now than in years past, when a more monolithic
approach was common.
Unless you’re IBM or Google, you’ll never build a spreadsheet that rivals Excel.
Furthermore, most business users already own Excel and know how to use it. Thus,
instead of building spreadsheet-like features into your application, it might be better
to find a smooth way to integrate with Excel through data export and import. Many
data-rich applications among our winners chose this course.
As another example, Wakelicious didn’t implement a complete system for uploading
and storing videos for Internet playback. Instead, it offers a simple integration with
YouTube, which not only saved the team tremendous development effort but also
gives users access to a pre-populated content library from which to choose wakeup
calls to send their friends.
One caveat about relying on other software is that it's often best to modify or
simplify the user interfaces to fit the needs of your users; because of its broader
purview, 3rd-party software often embodies unneeded complexity that it’s best to
hide. eReview, for example, built on top of SharePoint but designed a more
streamlined process than the one provided out-of-the-box.
Although it’s great to leverage existing applications and Internet resources,
applications must also provide their own value-add in their core domain. This was
most clearly shown by the Climate Action Planning Tool, which actually started out as
a regular Excel spreadsheet that could perform the necessary calculations.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 13

 Developing it into a real application to guide facilities managers through the workflow
— and explain and illustrate the complicated concepts involved in reducing CO2 —
made the experience much more successful.

SUPER-TOOLTIPS: INTERACTION DESIGN TECHNIQUE OF THE YEAR

In Application Design Showcase #1, 1 we declared the lightbox to be the interaction
design technique of the year. This time, we saw improvements in a broad range of
interaction techniques but no single new widget that was widely used.
Several applications, however, enhanced an old idea: the tooltip. These “super”
tooltips stretch far beyond their original intent and thus earned the spot as the year’s
most improved interaction design technique. Emerging in the 1990s, tooltips (or
“bubble help”) started off by providing small explanatory texts that appeared as
contextual help when the user pointed to a command button or other interface
element.
The new super-tooltip isn’t actually a “tip,” but we’ll stick to this well-established
name with its long history in GUI development. A super-tooltip is actually an
explanatory pop-up that appears as a rollover effect when users point to an item in
the user interface. This “item” might be a command or a button — as with traditional
tooltips — but it can also be a user-created object, such as a message or product
order. The following screenshot shows a typical super-tooltip visual design:

Example of a super-tooltip from Eventbrite.

In addition to static help text, super-tooltips usually contain dynamic status

information about the object that users are pointing to. This lets users see the
information in the context of a larger overview screen, instead of having to click
through to a separate screen with detailed object properties and editing options.
Super-tooltips are thus particularly suited to complex applications with many data
objects that users must consider together.

1
Available at http://www.nngroup.com/reports/applications/showcase-1/

14 [email protected] Executive Summary

 The super-tooltip shouldn’t include editing options; it should be for information only,
since it appears on hover and disappears when users move the mouse. Users should
also retain the option to click through to a specific screen that’s optimized for
viewing and revising an object’s properties in greater detail.
The entire Dixio application could be viewed as an even more souped-up super-
tooltip. However, that wouldn’t be entirely accurate because Dixio explanations
appear when users double-tap (or long-tap) a word. Thus, it’s not a rollover/hover
effect and has to be explicitly dismissed.

USABILITY METHODS
Most winners extensively employed a wide variety of usability methods, which is
certainly one reason they ended up with such good user interfaces. Iterative design
and user testing were particularly common and frequently used to grind down
complexity relative to early design ideas that seemed good but were too difficult.
Another common trend was the steady reduction in documentation and help text
throughout the iterations, as user testing repeatedly revealed the old lesson that
users don’t read very much.
Simple user testing with a handful of users can often substantially improve product
success, and was used by 85% of the winning projects. For example, an early design
for the Mobile Unified Communication client used the command Add Call for placing a
new phone call while putting the current party on hold. Although this seems logical
enough, it was confusing to users who thought the command was for conference
calls. Changing the command label to New Line did the trick.
One-third of the winners used field studies (also known as ethnography), venturing
onsite to observe users’ natural behaviors in their own environments. From
BondWorks sitting with brokers as they made deals, to OEConnection hanging out in
dealership parts departments, to VitreaTeach watching radiologists at hospitals,
domain-specific applications require domain-specific user research.
Although it seems obvious that domain-specific applications need field studies, most
design projects in the past typically did without them. When we analyzed the first
decade of Intranet Design Annual winners, we found that field studies were used by
only 10% of the winners from 2001–2003 and just 23% of winners from 2007–2009.
Although this represented more than a doubling within a single decade, it’s still sad
that less than a quarter of good intranet teams ran field studies at decade’s end.
Going from a quarter of projects to a third in a few years shows continued progress
for this important method.
In addition to field studies, we also saw substantial growth in a usability method that
was often overlooked in the past: sitting in on support calls. Knowing why customers
call with problems is a prime way to discover pain points — as long as you recognize
that a customer’s description of the problem might be far afield from the underlying
root cause that the redesign must address.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 15

 Overview of the Winners

In this installment of the Applications Showcase we increased the number of selected

winners from 10 to 13 (a clear reflection of the quality of the submissions we
received,) and among the applications selected for inclusion in this year’s report
we’ve identified a few commonalities across teams, including:
• US dominance: A majority of winners this year, 2012, (69%
compared with 50% in the 2008 edition of this report) are from US-
based companies.
• Small teams—big impact: Teams large and small built the
applications featured in this report, but small teams are proving that it
doesn’t take an army to build a great app. Nearly a third of the
winning teams had only two members.
• Made in-house: Fewer teams this year share their accolades with
outside agencies as every app in this report was designed and built, at
least in part, by an in-house team.

MADE IN AMERICA BUT USED AROUND THE WORLD

The United States leads the pack among the winning organizations. US-based
companies produced nine (of the 13) winning applications featured in this report,
while companies from across Europe round out the winner’s circle with one
application each coming from the UK, The Netherlands, Germany, and Spain.
So while the companies represented in this report may not be all over the map
geographically, what users can do with these winning applications certainly is.
The thirteen teams featured in this report built applications with uses ranging from
interpreting advanced diagnostic measurements (GEN2i) to using a YouTube video
sent by a friend as an alarm clock (Wakelicious.) They provide tools for industry
sectors as different as fixed-asset (bond) investing (BondWorks) and a non-profit
organization dedicated to cycling. The teams featured here serve many different
sectors, including: scientific engineering, financial services, retail automotive,
telecommunications, education, government, and NGO and general use.
This year’s winning teams may be mostly based in the US but the reach of their
efforts is felt by users across the globe.

IT DOESN’T TAKE A VILLAGE TO BUILD A GREAT APP

Team size, as usual (and as we find with much of our research,) spans the spectrum
from large (15 members) to small (a single user experience designer,) but a
noticeable trend emerged this year as winning duos proved that two heads really are
better than one (or more.) Two-person teams designed four of the winning
applications (Vitrea Teach, Cycle Streets, Wakelicious and Cohdoo Highlight.) And if
you only have two resources to throw at a project it seems a pairing between a
designer and a developer can yield winning results. It did for these teams.

IN-HOUSE TEAMS LEAD THE WAY

The votes are in and so is the work, in-house that is. Far more teams forged ahead
with the team they have on hand to design their winning application rather than

16 [email protected] Executive Summary

 hiring from outside. All but one of this year’s winners did all or most of the work on
their applications in-house. And even the teams that brought in reinforcements in
the form of outside agencies and consultants did so for only a small component of
the work.
The talented teams featured on these pages prove that it really doesn’t matter what
size team you have or what language you speak. What really matters is that you put
your resources toward good work: serve the needs of the end-user regardless of
whether your technology is mission-critical or surprising and whimsical.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 17

 Screenshots are Just a Moment in Time

We appreciate the designers and organizations with winning applications in this

report. They were generous enough to share their screenshots and stories with us.
We would like to acknowledge that the screenshots and descriptions are from the
version of the applications that we reviewed, and the design of these applications is
ever-changing. We included as many design updates as possible up until the time of
this report’s publishing, but please know that these apps may have already changed
by the time you are reading their case studies. Regardless, we feel that the
discussions about why the designs are so good and lessons about the teams’
processes will carry on for a very long period of time.

18 [email protected] Executive Summary

 Design Trends

We can’t help but notice some design trends and commonalities across the excellent,
though widely different, applications in this report. This section summarizes some of
those patterns.
DESIGNING FOR DIFFERENT TYPES OF USERS
Application designers who have one clearly defined type of user are very lucky. Most
app designers are faced with the challenge of designing for numerous different user
types who have varying needs, tasks, approaches, and knowledge. What the great
app designers in this report demonstrate is the ability to determine and contemplate
the particular user types and the mindset of each at any given point in the UI.
Designers made decisions about how to cater to the different user types in different
situations. Some ways the designers helped the different user groups include:
• Making Top Tasks Obvious for All: Identify essential starting points—
the ones people actually use, not just the ones you want them to use
• Presenting Various Options and Views to Meet Needs: See and
manipulate content in multiple ways, such as: text and images views,
categories, and filters
• Multiple Paths: Offer multiple, but not too many, paths to do the same
or a similar task (This means different methods, not listing the same
command or link to access a feature multiple times)
• Preferences and Customization: Customize the UI and edit
preferences to affect content pushed to the users, and arrangement and
look of UI elements

MAKING TOP TASKS OBVIOUS FOR ALL

One good way to cater to different user types in the same app is to ensure that the
main tasks that any users will want to do are incredibly easy and accessible.
Wakelicious (a social video alarm clock) designers researched and contemplated
before they came up with two buttons for their most important tasks in their app:
Set alarm and Wake a friend. These are explanatory and cut to the heart of the app
in the initial screen.
Cohdoo's Highlight (recording) focuses intensely on the core action of marking a
highlight. The majority of the screen is devoted to the Highlight button so that users
can perform the action without even looking at the device.
Hobson’s Product Suggestions app designers knew that two of the app’s top tasks
are 1) dealing with suggestions that have been made, and 2) writing new
suggestions. Thus the top two tabs in the UI are simply View Suggestions and
Submit Suggestion. These are easy to see, click, and the text is easy to understand.
On a similar note, the Eventbrite app displays two rows of options (tickets) by
default in order to introduce the idea of creating more than one type of ticket.
Sometime users actually need to do something before they can use the app's main
function. The Climate Action Planning Tool app, which supports users with
varying degrees of knowledge about energy use and goals for greenhouse gas

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 19

 emissions, offers an incredibly clear call to action at the start: the Start: Gather
baseline data button.
Verizon’s Mobile Unified Communications app caters to different users by
acknowledging that some people mostly make phone calls from their business phone,
some mostly make mobile calls, and some make roughly an equal number of mobile
and business phone calls. The UI makes it possible for the user to quickly and easily
choose between these (and they can also set up default quick access to one or the
other.)

PRESENTING VARIOUS OPTIONS AND VIEWS TO MEET NEEDS

The IMF's eReview (for managing the review and approval of report drafts) has
separate tabs and workflows for different types of users; one area is set up for the
department that creates a report, and other tabs are designed for departments that
review and comment on reports. Similarly, Dixio (lookup technology) offers the
option to search for a word’s definition as either text or images. So people who are
doing more of a visual task, or those who tend to respond better to pictures than to
words may be able to work in a more helpful and meaningful way because of this
option. Dixio also translates any foreign language words into the language most of
the article is written in.
Sometimes offering the same information sliced in different ways can help users find
what they want. This same app, Dixio, presents search results defaulted to all
results, but offers tabs for each individual dictionary (content source) which the user
has searched in. So either people who are questioning a particular definition or
people who respect one source more than another (even for a particular type of
word) can find their answer. But the users who do not care for one particular source
over another may stick with the default state, all sorted.
Within the view area of Hobson's Product Suggestions, users may sort by sub-
tabs—All Suggestions, Suggested by Me, and Suggested by Co-Workers. These
choices enable people to easily review and compare suggestions by source.
Additionally, designers included filtering options and a few different ways to refine
results, including searching for a keyword, filtering checkboxes, and sorting the list
by other criteria.

The VitreaTeach app (for viewing and sharing medical images) offers hierarchy,
filters, and the possibility for user-generated tags; all of which help radiology
residents to review pertinent, interesting images from old cases.
Messages in the lower left on the BondWorks app’s homepage allow brokers to
keep abreast of news, alerts, and announcements. Advisors may sort this list and
see, for example, all orders for a particular account, or all pending orders, which can
help them keep abreast of current work and plan for the future.
This same app offers a bond comparison tool which allows advisors to choose some
of the bonds they are most interested in and compare them. Designers even
anticipate the user’s next steps and make it possible to buy a bond or print the
comparison table.
The Climate Action Planning Tool app presents the final report after users enter
data, but also presents the users carbon emissions reduction goal as a drop-down list
so they may easily test different goals and see the specific results in a graph.

20 [email protected] Executive Summary

 MULTIPLE PATHS
GEN2i, app for recording engineering data in the field, offers a fun way to scroll
through data values for the standard sequencing using a rotary knob on a touch-
screen. An added touch of whimsy, the knob makes a clicking sound as it spins. But
sometimes users are not in a whimsical mindset: People who just want to enter the
value quickly may skip this feature and instead use a different UI, a more traditional
button encompassing the range of numbers to choose from. For full flexibility, the
app also allows people to enter their own values outside the offered range.
BondWorks is designed for financial advisors and brokers who are working with
bonds either 1) all day every day, or 2) rarely. An Advanced Search link leads to
involved search options which help cater to the different user types, but the simpler
default search doesn’t get in the way for any users.
The OEConnection app for ordering and tracking automotive parts always displays
the current part number at the top of the list, but users may also manually type a
new part number if they know it.
The Highlight app allows editing elements on the Details screen by tapping and
holding them, but this screen also includes an explicit Edit button in the top right
corner for users who have not yet learned the gestures or who prefer using clickable
buttons to gestures.

PREFERENCES AND CUSTOMIZATION

Dixio allows the user to edit preferences to see information in the format he
chooses, and allows people to customize 2 their own glossaries.
Verizon’s Mobile Unified Communications app allows users to customize the
defaults for the main buttons (or widgets) setting a button to the type of most
employed phone call, either Business, Mobile, or Prompt me [to choose business or
mobile] every time.
CycleStreets designers understand that some riders may be cautious and prefer to
always take quiet streets even though a busier street might be quicker. In the
Preferences area of the app, user may easily customize the journey planning process
and set their default route type preference, as well as their normal speed. They may
also turn off the GPS location feature in order to conserve battery life.
The Eventbrite app (for online ticketing platform) allows users to change the visual
appearance (background color, text color, images) so it suits their preferences.

WIZARDS
Wizards and linear processes are here to stay, especially for complicated and
infrequent tasks.
Verizon’s Mobile Unified Communications helps the user get his app set up in the
way he needs it. A wizard process, planned for various user scenarios, makes
customizing this very productive and not overly constraining.

2
For more information about customization, please see our separate report
Customization Usability: 46 Design Guidelines To Improve Web-based Interface and
Product Customization, available at http://www.nngroup.com/reports/customization.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 21

 The Climate Action Planning Tool app offers numbered steps to lead people
through their process. Interestingly in research studies designers learned that the
Before You Start step was ignored as users wanted to just get into the work. In the
current design the important steps that used to be categorized under Before You
Start are now under a section called Step 1: Gather Baseline Energy Consumption
Data. This tested far better and is a very good lesson for other designers creating
stepped UI’s such as wizards: Be cautious with information-gathering screens or
pages as users often think they are information-only pages and not necessary to
their process.
LINKED CONTROLS
Linked controls, a form of progressive disclosure, hide choices users don’t need and
show only those that apply to the current situation or questions the system is asking.
It’s easier for designers, often, to just present everything to the user at once, but
winning app designers do extra work to create categories and give users information
and questions just in time.
The OEConnection app for ordering and tracking automotive parts makes it easy to
choose the right part number. Specifically, designers implemented a linked control
relating the Part # field with the New Part # field. The New Part # field considers the
number in the Part # field as well as other information in the order, and based on
database and inventory knowledge suggests like or better parts when the New Part #
field is clicked. A drop-down displays part numbers the user may possibly want.
IMF’s eReview app changes the display of buttons depending on both which step the
user is in in the process (you can do certain things at certain points) and the user’s
access (based on her role.)
The second button, Quick Call, in Verizon’s Mobile Unified Communications app
changes depending on whether users chose Business, Mobile, or Prompt me [to
choose business or mobile] every time as the first button choice. The second button
is a linked control and only presents choices related to the first button’s selection.
DOWNLOADING OR SENDING DATA TO ANOTHER APP TO REVIEW
AND MANIPULATE
Great designers know they shouldn’t reinvent the wheel, nor should they code in
grand, complicated features when their users already have other apps which support
them. When designers do make these imposing attempts the resulting
implementation of the complicated features are often a half-measure as compared to
the other robust app. Or the new features are fine but add too much complexity to
the app as a whole. Better is to make an export possible, simple, and reliable so
people may take advantage of another jam-packed app.

The Hobsons app does this, enabling users to copy the data (suggestions) to a
spreadsheet file by just clicking the Download Suggestions button. Users may then
use a spreadsheet app to manipulate the information.
GEN2i designers know that their users typically use their app to do measurement
tasks, maybe some viewing and basic measurements, but engineers look for the real
power in another large software package. Thus, the Export Recording option is very
easy to see and use, and offers options to export the full or partial in various formats
such as Word, Excel, or PDF.

22 [email protected] Executive Summary

 A related but somewhat different feature from downloading data from an app
appears in the Climate Action Planning Tool app. Designers derived a
downloadable worksheet to help users gather information that is essential to getting
their task done, and getting the most value out of the application. Users must
consult various other people about the different energy sources on campus. Thus,
designers created a worksheet that is optimized to fit the way users perform the
task, looks like the online form in the app which makes it easy to transfer the data
into the application later, and is a PDF so the look remains intact when it is printed.
This same app does allow saving and printing the final report results, based on all
the data the user input in the app, as a PDF document as well.
MAKING THE MOST OF SPACE
Using screen real estate effectively is a skill that the winning designers in this report
all have. These designers prioritize the tasks along with what needs to be
communicated in the allowable space.
The Highlight (recording) app designers chose to not include a bottom navigation
bar in their mobile app in order to keep a strong focus on the primary task of each
screen.
Cyclestreets also relegates the global navigation options (such as switching
between different types of tasks) to a Home screen. This leaves more space on the
map display screen for actually viewing the map, and for controls that let users
quickly flip between different versions of a route.
VitreaTeach designers knew that students were very interested in the text
accompanying medical images, not just the images themselves, so the screen real
estate is balanced between images and their accompanying descriptive text.
Replacing content in the same area is one way to save space in a dashboard, or any
UI. GEN2i changes the content in the lower right rail depending on what is selected.
Sometime layout is critical to understanding functionality. Eventbrite presents the
steps in its Create workflow into a vertical line, to make the sequence of steps
obvious and help simplify a complicated process.
SAFETY NETS
The Gen2i app automatically saves data it collects, as does Highlight. Designers of
these apps are of the mind that if the user takes time to set up a collection or a
recording then they meant for it to be saved. So why make people do additional
work to keep their data? In these particular cases, unless the system would slow
down or run out of space there is no reason to force the user to save.
Highlight also anticipates that many users will realize they want to mark a highlight
only after they experience the action, so the app allows users to automatically mark
all highlights for a few seconds prior to the important event.

Wakelicious has a backup alarm clock built into the application. It is turned on by
default, so that even if the app is disabled or loses an internet connection, users will
still be woken up on time.

eReview manages collaboration among different departments, and everyone needs

to be informed of the project status. The system manages this not just by making
the status and task assignments visible, but by proactively sending out notifications
when tasks are assigned or completed.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 23

 OEConnection designers realized that users may not always remember to check for
conversion opportunities, or parts within an order that they aren't specifically
requested to fill, but that they could take the initiative to fill. The app places this
information front and center within the order details screen, so users won't overlook
opportunities.

HELP (INLINE)
Most applications focus on ensuring that controls and labels are self-explanatory, and
that users don’t need to turn to online help. Wakelicious designers believe this, but
also make help accessible for users who seek out the help icon. Initially the app
offered an abundance of help, but testing showed (a very common usability finding)
that users didn’t read the prominent instructions, so they were shifted into an
auxiliary role. The few areas of the app where users are likely to be interested in
learning how it works rather than just using it—like the Settings screen—have brief
explanations right next to the relevant controls. Many great applications give bite-
sized, accommodating assistance in-line in this manner rather than as huge help
files.
Bondworks includes hover-activated tool-tips next to search fields, to quickly
explain the purpose of each field.
The Climate Action Planning Tool app offers a What’s this? link after various
questions in the app, such as Buidlings_% compounded over 40 years. The link
opens a pop-up box describing the Buildings topic as well as the typical growth rate
per year for that item.

24 [email protected] Executive Summary

 Complex Applications

Applications in this Complex category are those which are usually meant to support
involved processes or complicated problems. Some tasks and content are technical,
mechanical, or financial in nature.
The users of complex applications are often highly skilled in a particular domain,
such as engineering or finance, and may be willing to learn some of the nuances and
intricacies of a deductive application design because of the luxuries available with
more advanced and powerful features. This doesn’t mean, however, that the app
should not still follow good design principles and be as simple to use as possible.
One of their great common problems in complex applications, no matter what the
domain, is the need for seeing or having the ability to access large amounts of data
all at the same time.
UI’s that often help support complex activities include:
• Dashboards and portals
• Using the same space to rotate through different content
• Involved, advanced search capabilities
• Possibilities for viewing data in different ways
• Exporting data to other applications
• Graphical displays of data.
Winning applications in this category include:
• The BondWorks™ wealth management platform, a web-based
application providing financial advisors a streamlined system to trade
fixed income assets.
• GEN2i, a portable device for recording engineering data in the field.
• OEConnection allows parts department employees at auto dealerships
to quickly process parts orders sent by collision body shops, repair
facilities, consumers and other auto dealerships.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 25

 BondWorks™ Advisor Platform

Organization: Design team:

BondDesk Group LLC is a leading fixed-income All design and usability efforts were handled
technology firm. internally. Two agencies were consulted with at the
beginning stages of the product redesign: User Think
Application: and Momentum Design Lab.
The BondWorks™ wealth management platform
is a web-based application providing financial Members:
advisors a streamlined system to trade fixed
income assets. In-house: Kamaryn Tanner, Product Manager; Ken
Hoffman, Product Manager; Kelly Arkles, Team Lead,
Headquarters: Mill Valley, CA (USA) User Experience Team; Brad Taylor, Senior
Designer/Frontend Developer; Joe Dakroub,
www.bonddeskgroup.com Designer/Frontend Developer; Victoria Whitehorne;
Technical Writer and Christina Matsoukis,
Designer/Frontend Developer.

OVERVIEW
Main Function: The BondWorks™ Wealth Management Platform gives financial
advisors the tools to carefully evaluate the best investment choices for each client
they serve, efficiently execute trades and manage clients’ fixed income planning.

Why it’s Unique: BondWorks enables advisors to find suitable bonds and provide
trade suggestions clients. Intuitive design and wizard tools makes it easy for
infrequent users to buy or sell bonds confidently, and sophisticated features appeal
to those advisors who are focused on managing large, complex fixed income
portfolios for high net worth clients.
The Organization: BondDesk Group LLC is a leading financial technology company,
providing innovative, enterprise-wide fixed income solutions to a community of over
2,000 investment firms including many of the top broker-dealers and buy-side
institutional accounts in North America. BondDesk’s flagship application, the
BondWorks™ Wealth Management Platform, enables over 25,000 financial advisors
and asset managers to trade bonds with confidence and gather new assets.
Design Philosophy: A look, listen and learn approach is how the BondWorks team
took their bond trading application from functional to fabulous. By taking a
meticulous and sensitive look at how their users do the work of trading bonds, the
BondWorks design team was able to streamline and optimize functions big and small
across the bond trading lifecycle and across their application. In helping the users
perform daily tasks with ease they ultimately helped their own bottom line, paving
the way for increased trade volume and increased profits.

USERS
Approximately 25,000 financial advisors, who trade bonds on behalf of their clients,
use the BondWorks application. The application is optimized to help make bond
trading accessible for financial advisors who don’t necessarily trade bonds every day
as well as to serve advisors whose primary business is built around fixed income.
The application has been designed to streamline processes for power users and give
confidence-building tools to advisors for who trade less frequently. To accomplish
this, BondWorks offers a variety of tools to facilitate research, workflow and trading,
including the following:

26 [email protected] Executive Summary

 • BondWorks Ladder Wizard: A guided workflow automates this
historically complex, manual process.
• Smart Search: Advisors can use a powerful basic query screen, an
expanded detailed query screen, access to the last 25 recent searches,
and a natural language search engine.
• Bond Comparison Tool: Advisors can perform side-by-side
comparisons of up to five bonds in an easy-to-read format.
• Monitoring & Alerts: BondWorks tracks the status of key events and
allows advisors to create watch lists to monitor availability of bonds in
the system.
• Transparency into available offerings: Advisors have maximum
transparency into their available inventory with tools such as a scatter
plot; access to the last 25 recent searches; rollovers with more bond
detail; toggling between table and graph views; and grouping based
on rating, asset class, and optionality.
• Customizations: Advisors can select their own default homepage;
customize the alerts they receive; and add, delete, or remove columns
from their search results.
• Order and Bid/Offer Wanted Summary: BondWorks provides users
with a trade blotter that shows all open and recently placed orders.
Color-coding makes it easy to check the status of each order.

THE APPLICATION
BondWorks takes what could be an extremely complicated and dry application—an
advanced retail fixed income trading system—and turns it into a helpful, focused tool
that truly aids financial brokers and advisors.
The homepage focuses on what is undoubtedly the most important and common task
for financial brokers and advisors: searching for bonds. Ancillary tasks are possible
too via the top horizontal and left-side menus. While there is a lot on the homepage,
designers subtly made the visual focus on the main task by adding a large amount of
white space to each side of and below the search fields.
The bond market is a relatively complicated one; there are less than 10,000
companies listed on the stock exchanges, but about 4 million bonds available for
trading. Choosing the best bonds thus requires detailed information and thorough
research. Knowing this, designers created the Search for Bonds area on the
homepage, which includes more specific search types, housed in the following tabs:
Criteria Search, Yield Search, State Search, Quick Picks, and Saved Searches.
The default tab, Criteria Search, displays ten fields, many of which offer drop-down
selection when applicable to expedite the task of searching. The remaining fields
allow brokers to enter freeform text values for quantity, price, yield to worst, name,
maturity, and coupon, as these are far less predictable choices. Each field question
has a tool-tip link, which can easily be ignored or displayed, depending on the user’s
needs.
An Advanced Search link leads to even more involved search options. Given the
app’s user types, financial advisors and brokers who are working with bonds 1) all

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 27

 day every day, and 2) rarely, this UI helps cater to both of these but and doesn’t get
in the way for either.
In addition to searching, the homepage provides messages in the lower left, so
brokers can keep abreast of news, alerts, and announcements. The section below
search criteria on the homepage summarizes the most recent orders, presenting the
information of interest in the table, such as: order number, status, transaction,
quantity, issue description, and account number. Advisors can sort this list and see,
for example, all orders for a particular account, or all pending orders. This feature
helps them keep abreast of current work and plan for the future.

The homepage is the entry screen for financial advisors that enables them to
view and access key decision making elements for searching, messages, orders
and tools.

28 [email protected] Executive Summary

 BondWorks provides advisors with a powerful basic query screen, an expanded
detailed query screen, access to the last 25 recent searches, and a natural
language search engine.

The search results page displays very detailed content in a sortable table. Users can
select the number of results to see on the page and use pagination buttons at the
bottom to see more results pages.
In the upper right, users can toggle the data view from a table to a graph.
Additional buttons in the upper right allow users to access their 25 most recent
searches, construct a new search, modify the current search, or save the search.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 29

 BondWorks provides advisors with maximum transparency into their available
inventory with tools such as access to the last 25 recent searches; rollovers
with more bond detail; toggling between table and graph views; and grouping
based on rating, asset class, and optionality.

Clicking the graph button allows users to view the search results in a scatter plot,
which is often easier to consider than tabular format. The graph gives brokers a
better sense for how a bond stands in the market in relation to other bonds.

BondWorks provides a scatter plot view for analysis.

In the chart, users may choose a particularly interesting section, draw a box around
it with the cursor, and add bonds within that section to a reminder list. They are then
offered options regarding the watch list. This is just one more feature designers
added to help advisors analyze bond activity.

30 [email protected] Executive Summary

 BondWorks tracks the status of key events and allows advisors to create watch
lists to monitor availability of bonds in the system.

The user can switch back from the scatter plot to the tabular view by clicking the
Table button in the upper right. Advisors may then choose bonds they are most
interested in from the table and compare them. Clicking the checkbox to the left of
the result enables the Compare button in the upper left. This opens a dialog box that
lists the selected bonds and compares them, displaying green up and red down
arrows to indicate relative performance. Designers anticipated the user’s next steps
by enabling the advisor to buy a bond from here, or print the comparison table.
Closing the dialog box returns the user to the search results page.

Advisors may perform side-by-side comparisons of up to five bonds in an easy-

to-read format.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 31

 After buying a bond, an advisor may choose to check its status to ensure the order is
in the system. The Order & Bid/ Offer Wanted Summary (sometimes called the
“Blotter”) page gives details about the transactions which have posted or are in
progress. The row just above the table gives a snapshot of activity, including the
total number of orders, numbers pending, filled, rejected, cancelled, and expired.
This table affords action as well as passive reading. It’s not just for sorting as
advisors can cancel pending orders listed in the table.

BondWorks provides users a trade blotter that shows all open and recently
placed orders. Color-coding makes it easy to highlight the status of each order.

Designers wanted users to be able to access what they need from anywhere in the
UI without disrupting search results, charts, or currently displayed data. So they
implemented the EasyNav Toolbar, which appears in the upper right corner of every
page in the system. The toolbar includes icons for tasks such as: reading system
messages, reviewing pending orders, checking for watch list alerts, opening a list of
favorite screens, and browsing the list of system-defined short cuts. Designers chose
to invoke the action on click instead of on hover because they were concerned about
data disappearing when the mouse was moved. Advisors deal with a large amount of
data so an error here could be a disaster. The EasyNav Toolbar provides access to
important system information and, more importantly, allows users to see that
information without leaving the screen they are currently using.

32 [email protected] Executive Summary

 BondWorks provides users with the EasyNav Toolbar in the upper right-hand
corner of every screen in the system. With a single mouse click advisors can
read system messages, review pending orders, check for watch list alerts, open
a list of favorite screens, or browse the list of system-defined short cuts. Not
only does the EasyNav Toolbar provide access to important system information,
it allows users to see that information without leaving the screen they are
currently using.

Another task brokers need to do is build a bond ladder. 3 Historically this was a very
complex and manual process. Thus, BondDesk designed a wizard to ease advisors
into building ladders for clients. The wizard takes users through the process, step-by
step, asking questions about tax, bond type, credit rating, and ladder structure.
Originally designers created the wizard to give confidence to users who don’t
consider themselves bond experts. It worked, but the “big bond guys” use the wizard
as well. It makes sense because the wizard consolidates several pieces of disparate
information easily and quickly.
In the wizard, each step follows important guidelines about good wizard design,
including the following:
• Heading at the top naming the reason for the wizard itself, in this
case, Ladder Builder.

3
A bond ladder, according to investopedia.com is “A portfolio of fixed-income securities in
which each security has a significantly different maturity date. The purpose of purchasing
several smaller bonds with different maturity dates rather than one large bond with a single
maturity date is to minimize interest-rate risk and to increase liquidity. In a bond ladder, the
bonds’ maturity dates are evenly spaced across several months or several years so that the
bonds are maturing and the proceeds are being reinvested at regular intervals. The more
liquidity an investor needs, the closer together his bond maturities should be.” BondDesk
designed a wizard to ease advisors into building ladders for clients lessening the need for
liaison traders to perform this function manually. The wizard approach provides an easy to use
automated tool in which a user can find bonds to include in a ladder based upon specific client
and bond characteristics.
http://www.investopedia.com/terms/b/bondladder.asp#ixzz1qWstiOjC

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 33

 • Clear status bar at the top which includes how long the process is;
the actual steps involved; where the user has been in the process, is
currently, and where else he will go. In this case visited steps are
green, current is blue, and future are light grey.
• Page/task name and heading, noting the main things the user will
do on that particular screen.
• Clear questions, labels, and fields.
• Clear ‘advance’ and ‘back’ options, in this case Previous Step and
Next Step buttons in the lower right.
• Plus, of course, a way to cancel out of the process in case the user
entered the wizard by mistake.

Step 1. Tax selections

34 [email protected] Executive Summary

 Step 2. Bond Type selections.

Step 3. Credit rating selections.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 35

 Step 4. Ladder structure selections.

Once the wizard questions are completed, the system selects bonds that meet the
entered criteria and organizes them by year. At this point, the advisor can get a
sense of the cash flow the client might receive, and generate a PDF report to give to
the client.
(Note that even though BondWorks is a winner in the complex applications category,
the need for an advisor — the primary user — to interface with clients means that it
also has advanced workflow elements. The clients can be considered secondary users
in some ways and BondWorks needs to support this multi-person workflow. For more
on workflow, see the category of workflow application winners, starting on page
141.)

36 [email protected] Executive Summary

 Step 5. Results give brokers a sense of what the client’s cash flow could be.

On the results page, brokers may buy bonds by clicking the Buy button to the left of
each item.

Step 6. At the end, brokers can move forward and buy bonds, using the BUY
buttons on the left in the table. The Ladder Builder wizard lessens the need for
liaison traders to perform this function manually. The wizard approach provides
an easy-to-use automated tool through which a user can find bonds to include
in a ladder based upon specific client and bond characteristics.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 37

 Another feature that helps new and advanced users is the Did You Know carousel.
This carousel provides an easy to use and easy to read help option. The beauty of
this interface element is it makes people aware of new functionality in the app. It
also provides quick access to documentation on more complex workflow tools.
Context-sensitive online help is also available to assist users.

Did You Know Carousel: In addition to standard online help and context
sensitive help, BondWorks uses the Did You Know carousel to quickly highlight
new features added during major releases and to highlight more advanced
workflow tools.

Any user can be aided by good customization features; BondWorks provides the
ability for users to change their profile in Settings, select a default homepage,
choose a default search tab, and select the number of search results to display per
page. There are several other configuration options which help expedite work and
meet the needs of individual advisors, such as the ability to choose which alerts to
receive and which columns display in the search results.

38 [email protected] Executive Summary

 Users can change profile settings to select a default homepage and display a
particular search tab by default, among other things.

With a good mixture of inductive and deductive UI elements, BondWorks leads

experienced and less experienced advisors and brokers through the processes of
researching, buying, and communicating about bonds. An ingenious dashboard and a
thorough wizard, among other features, increase user confidence and productivity.
Finally, anticipating the user’s next step, focusing page priority, and embedding
needed albeit unexpected features makes this app a world-class winner.

DESIGN PROCESS

Usability Methods
The most recent product redesign resulted from two primary drivers:
• The application was looking a bit dated and it was time for a visual
refresh, and
• The company wanted to streamline many of the bond trading
processes to help traders increase their volume, and thus increase the
company’s bottom line
The overall goal was to redesign the application from the ground up. “We really
wanted to do a re-design from the ground level,” says Product Manager Kamaryn
Tanner, “To stop and think about what tasks brokers are trying to accomplish we
wanted to go back to fundamentals. We didn’t want to just put a pretty face on
what was already there. We wanted to re-do the information architecture
throughout the application so we started at the task level.”
This required a number of research methods so the team could get a complete
picture of what they needed to do to greatly improve and enhance the
application. These included:

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 39

 • Watching users work: Because it’s difficult to get in front of actual
brokers to do task-based testing the design team did the next best
thing, they spent time sitting with brokers as they went about their
daily work. That exercise proved pivotal in learning where the
roadblocks were in the application and where the design could support
streamlining of frequently used or difficult processes. “We actually
talked these guys into letting us sit at their desk with them,” says
Tanner. “And as they took client calls they let us listen in. We also
watched what they did to sell these bonds and use the software.”
These sessions helped the team identify where and when the advisors
turned to paper and pen as part of their regular process. Those paper
moments highlighted opportunities for improvement in the application,
to provide functionality to replace manual processes. “For example,
they would write on Post-it Notes and stick them on their monitors,”
she says. “We got a good flavor for what their day was like and what
activities they were having trouble accomplishing using their current
software.”
Another key finding from these observation sessions was the
realization that financial advisors work in an environment where
interruption is the only constant. The team took this reality to heart
when designing the new version of the application, building in support
mechanisms so brokers can be interrupted and still be productive.
Observing users also helped the team map out important processes,
such as the steps needed to build a bond ladder wizard.

40 [email protected] Executive Summary

 The process flow for the bond ladder wizard enabled the design team to identify
all elements needed before coding took place.

• Focus groups: The team enlisted the help of an outside firm to

conduct traditional two-way mirror focus groups. These sessions did
not focus on how the clients use BondWorks to do their work, but
rather helped inform the team about the overall challenges of what it’s
like to work with clients to sell investments and where bonds fit into
that investment big picture.
• Sitting in on support calls: Because Tanner had led many of the
trainings for the BondWorks application prior to the redesign, the team
was able to glean user pain points from the things she heard in those
sessions. “It became clear that when I had to explain a little bit too
much about something, we had a usability issue,” she says.
• Checking user stats: Doing an analysis of the user logs was an
ongoing part of the design process. The BondWorks team is fortunate
in this area as they can literally see everything their users do in the
application by studying user logs. Analyzing this data truly informed
some of the decisions the team made about specific functionality.
The design of the Smart Search is a good example of this. Searching is
one of the most common things users do in BondWorks. “In the month
of November, for example, 180,000 searches were performed,” says
Tanner. “It’s by far and away the most common activity that’s done so
we spent a lot of time looking at the logs to determine how users
search so we could help them in that area.” It’s important that the
application deliver finely tuned search results, otherwise the user will
be overwhelmed with results. “Someone who searches for ‘bonds’ and
gets 5,000 results is overwhelmed and is quickly out of there [the
application].” So the team focused on optimizing the search
functionality to serve the less frequent user.
Bonds have many attributes that are important when choosing which
one to purchase. The original application presented approximately 85
of those attributes on the search results screen. After studying the
search logs, the team realized that nearly three quarters of all users
were only searching on some combination of ten attributes. The team
simplified the search screen in the new interface to focus on these ten
key attributes. “That helps us present a search function that is a lot
less intimidating,” says Tanner.
• Surveys: Because the team employed beta users and pilot users while
they were designing and building the new application, they had a pool
of users at the ready who had knowledge of how the new application
compared to the old. At the end of the trial period, these users
completed surveys, rating the new product against the old. The new
application scored high, with over 90% of respondents saying the new
app was better. The respondents were also asked about specific
features, ease of use, etc., to gauge specific details to support their
overall opinions.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 41

 • Paper prototype testing: The first round of low-fidelity wireframes
were simple paper prototypes where the design team asked people to
be potential users and try to accomplish tasks with the prototypes. As
they made decisions and clicked on the screens, they were encouraged
to think out loud and share their thought process. These early
wireframes served as a good starting point for the team to determine
user flows and information architecture. The team also used these
early sketches to identify how important features would work.
• Wireframes: Higher-fidelity wireframes were developed later in the
process to solidify interface design elements and properly allocate
screen real estate. “We learned a lot, especially during that beta
phase,” says Tanner. A good example was an early wireframe comp
that had a lot of additional space in the header that didn’t need to be
there. In place of the white space, the team decided to pull other
elements up the page, to give them more visual prominence and to
distinguish the header space in the application from the various tool
bars and browser elements that are often present when the
BondWorks application is presented inside a company’s intranet
window.
Another finding realized during wireframe testing was that users did
not see the Smart Search box. By maneuvering elements around the
page in later designs, the team was able to give the feature more
prominence and make it more noticeable.

42 [email protected] Executive Summary

 Used mainly as a guidance model, the team’s design process was based on
achieving as many of the deliverables seen in this screen as possible. This
model outlines what was needed to achieve the new design.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 43

 This is an example of a low-fidelity wireframe used during the early task
analysis phase.

This is a high-fidelity design comp created during the design process to fine
tune a feature.

44 [email protected] Executive Summary

 • High-fidelity prototypes: During the beta phase, the design
team also introduced high-fidelity, interactive prototypes (or
simulations) for the order ticket flow, primarily using iRise (a
sophisticated prototyping tool). “This was an important step,” says
Kelly Arkles, User Interface Team Leader. “It allowed for us to
better visualize—for both developers and users—the overall
experience and process flow. It helped us find potential holes in the
design and improve the experience prior to any code being
created. For all new functionality and customized order ticket flow,
simulations are now required as it has proven to enhance the
design process, communications, and overall user experience while
also eliminating unnecessary rework.”

LESSONS LEARNED
BondWorks team members share a few of the lessons learned from the redesign
effort:
• Know your audience. The product team made a technology decision
upfront during the redesign: to provide browser support for Internet
Explorer 7 and above. That single choice presented a world of
challenges and a significant amount of work when they found
themselves forced to go back and retrofit the app for IE6. “We made a
decision that we were only going to support IE7 and IE8,” says Tanner.
“That actually affected a lot of things we did and the way we did them.
That was all fine and good until we needed to bring on a very large,
new client that was using IE6. In retrospect, I think we were maybe
just a little too aggressive in assuming that everybody would have
IE7.” The simple truth is that clients in the financial space are slower
than other industries to adopt new technology; assuming they could
uniformly eliminate an antiquated browser technology from their
product was unwise and costly.
• Take a product approach. Though a long-time coming, the team
decided that together with the redesign they would move BondWorks
from a client-based application to a product-based application. This
decision has saved the company both time and money and will pay off
long into the future. “What we had before was several custom versions
of our applications,” says Tanner, “The amount of custom code we had
was very hard to maintain and was definitely a testing issue for us. We
had to expend an awful lot of resources every time we did any work
because many clients’ code was different.” Going in a different, more
unified direction made sense, but was not universally supported — at
least at first. “Many believed our value was in customizing systems so
we spent a number of years demonstrating that we could provide even
greater value through innovation in the marketplace and a deeper
understanding of our clients’ business challenges.” A product approach
has not only saved the team countless cycles of reconciling code
changes specific to custom client versions, but it has resulted in a far
better product. “We designed a better overall application than any of
the customized versions our clients had,” she says.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 45

 • Know your design landscape. When designing a screen
environment that (often) lives inside of another screen wrapper (as is
the case with intranet apps such as BondWorks) it’s important to
really, truly understand how the window-inside-a-window affects
design size and the use of screen elements. “What we had planned
and designed for was 1024 x 768. But because BondWorks can be
pulled into a browser through various intranets and applications, we
couldn’t always accurately determine how much toolbar space or
screen real estate is taken up by that portal,” says Tanner. “And that
wasn’t something that we found until into the beta stages when a lot
of the design was determined. We had to determine how the elements
would fit in the screens at a specific size.” The solution they found to
accommodate the various unknowns of client environments (and still
maintain the integrity of their own application design) was to establish
a viewport size and make that a required condition of incorporating
BondWorks into a client’s portal environment. “What it came down to
is that we require the viewport size to be 1000 pixels wide by 625
pixels for our application and that doesn’t really affect the
functionality. It just affects the display,” she says.
Arkles adds, “After we made this discovery, simulations for customized
client order tickets have improved. We ensure that each simulation
reflects the viewport size so that clients can visualize how their
requests will affect the user’s view and interaction with the order ticket
flow prior to coding and the UAT process. This also helps us with new
functionality development as we can see how the worst case scenario
with viewport size affects overall layout of screen elements.”
• Choose technology wisely. Sometimes a simple technology choice
can lead to unplanned dependencies, as was the case with the
JavaScript library the application is built upon. “We chose a JavaScript
library that, at that time, was not at a mature level,” says Arkles. “We
were closely tied to the library releases and upgrades. The upgrade
process wasn’t as smooth as it could have been. Now it’s a very
mature library. It’s definitely gotten better but taking the approach
proved to be a very big challenge. Also, our developers had little
experience using the JavaScript library. There was a definite learning
curve to get to a more expert programming level.”

TEAM
“Hunting around for information really annoys advisers.” This quote identifies the
underlying principle for the application—large amounts of information need to be
easy to find and use. This basic idea combined with key usability concepts drove the
product and design teams to create an application that helps users get their job done
rather than get in their way. When suggesting ideas, all design contributors ask the
question, “Does this make the user’s job easier or harder?”

46 [email protected] Executive Summary

 Pictured (left to right): BondDesk Product Management Team members
Kamaryn Tanner and Ken Hoffman.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 47

 Pictured (left to right): BondDesk UX Team members Kelly Arkles, Brad Taylor,
Joe Dakroub, Victoria Whitehorne and Christina Matsoukis.

48 [email protected] Executive Summary

 GEN2i

Organization: Design team:

Hottinger Baldwin Messtechnik GmbH (HBM)
The GEN2i application was designed in-house by the
Application: company’s software group, in the Netherlands. The
Portable device for recording engineering data in team was comprised of six engineers and one
the field interaction designer.

Headquarters: Members:
Darmstadt, Germany In-house: Klaus Lang, Product Manager GEN2i;
Edwin Schild, Software Development Group Leader;
www.hbm.com Marco Dijkmans, Software Engineer; Coen Van
Dongen, Software Project Lead Engineer; Panagiotis
Georgiadis, Software Engineer and Cees de Vries,
Interaction Designer

OVERVIEW
Main Function: The goal of the GEN2i project was to create a touch screen
operation for a portable measurement system (comparable to a digital storage
oscilloscope.) The primary function of the application is to control the system,
acquire data, store data and provide basic analysis and report generation, both in
the field and in the lab. The measurement activities of this instrument can range
from airplanes to space shuttles to cars to heating: whatever you can imagine, as
long as it produces physical signals. It’s designed to record all kinds of physical
phenomena, such as temperature, pressure, power, force, etc to help engineers
diagnose problems. If something produces a physical signal GEN2i can measure it.
So if an engineer in the field can register all that information at the same time, he
can likely see what’s going on, what the relationship is between the various
components, and deduce what’s wrong.
Why it’s unique: The GEN2i application interface is remarkable for its flat structure.
The engineers have achieved a tight focus on core features, all accessible from one
screen and have made effective use of space in the interface. Though they had to
design for many controls, they still managed to reserve most of the screen real
estate for the results display. And to top it all off they’ve made it possible to
completely control the interface with one hand because the apps’ users are often in
situations where they're using their hands for other things. In fact, users can control
GEN2i with just one finger.
The organization: The Hottinger Baldwin Messtechnik (HBM) mission is to help its
customers optimize the product life cycle. HBM is a business-to-business company
with a product range that covers sensors, transducers, strain gages, amplifiers, and
data acquisition systems as well as sophisticated software for testing and analysis of
equipment. The potential fields of application for its products can be found in every
branch of engineering and industry in both virtual and physical test and
measurement.
HBM has approximately 1,500 employees worldwide and is a wholly owned
subsidiary of Spectris plc, a group specialized in precision instrumentation and
electronic controls. HBM has 27 subsidiaries and sales offices in Europe, America,
and Asia. HBM also has representatives in another 40 countries around the world. In
addition to headquarters in Darmstadt (Germany), other HBM production facilities
are located in Marlborough (USA) and Suzhou (China.)

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 49

 The group that built GEN2i, HBM Netherlands BV, is a small (approximately 20
employees) engineering group, part of the Germany-based HBM company
(www.hbm.com.)
Design Philosophy: The team behind GEN2i have achieved a nice, tight interface
solution to solve a very complex problem. When designing for expert users it’s
sometimes difficult to simultaneously accommodate new users as well. The
application has done both well, while keeping the focus firmly on the needs of the
experts. New users will have a steeper learning curve with GEN2i but experienced
users will appreciate the quick access to controls without deep menus to wade
through.

USERS
The GEN2i instrument has been sold to approximately 500 users worldwide since its
debut in the market a year and-a-half ago.
The interface was designed specifically for engineers who do troubleshooting. These
users come from two primary groups: technicians in the field and engineers in the
laboratory:
• The technicians have a measurement job to do, for example trouble
shooting an engine (ranging from a simple motor to a space shuttle) or
any other device. The technicians need to acquire data easily so they
can repair something that is broken.
• Engineers use the instrument to measure, analyze, archive and
report.
The users are often working in a cramped, hectic environment with a lot of noise so
the GEN2i interface is critical to these users who need to get to what they need
without taking the time to read the user manual.

THE APPLICATION
The GEN2i instrument is designed for particular types of users: smart people who
know what they want. They care more about the data than they do the app it is
encased in. Designers recognize this and present a UI that gets out of the user’s
way. A dashboard is in order given the user type and what they need for this app:
status information and the ability to record and watch data interactions.
Understanding that the main section with the data meters is very important,
designers dedicated most of the UI to this, but still offer other toolbars and panels so
users may manipulate data. They also, when possible, economize on space by
making areas multipurpose, replacing content with new and different content to
make the most use of the dashboard real estate.
Just after startup the main application shows all functionality, and the dashboard UI
groups the information so the users may scan and pick out what they want. The data
begins showing and collecting right away so users can start collecting data. Each
section in the dashboard uses backgrounds and lines to set themselves off from the
main content area and from each other. In general, the action buttons are accessible
but take little space away from the main content area. Given that even on a this
small, portable measurement system you may view up to 64 channels, real estate in
the content area is at a premium. The data the users care about takes up the bulk of
the dashboard, the center section. This area is used to play live data or play back
recorded data.

50 [email protected] Executive Summary

 The top horizontal bar with the labels Max and mV presents the real-time values of
the channels. These meters correlate with the channels presented in the main
content area in three ways, including:
1) name, for example, A1 is A1 in both areas,
2) color, A1 is yellow both in the top horizontal area and on the left, and
3) position, A1 is first on the left in the horizontal bar and first from the top in
the vertical bar.
These three criteria can truly help the users orient themselves in a busy dashboard
environment.
Users may add and remove, and drag and drop channels within the main content
area, and tap on a channel or meter to configure it. When an object is dragged an
information balloon pops up to give user feedback. Once the user has the
configuration she desires she may save it via the Save setup button, which is third
from the top in the left-most toolbar.

Dragging an object opens a yellow information balloon for feedback.

If a user is sitting in front of the instrument and is left-handed, then she is always
obscuring the more important parts of the instrument. Switching to left-handed
reverses the order of the vertical panels—everything switches around.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 51

 In addition to small customization, the app allows users to switch from righty to
lefty view.

52 [email protected] Executive Summary

 Users may switch to the lefty view, which swaps the panels, the right for the
left.

Many times engineers want to review the activity that has been inspected, but since
data consumes space and recording can make the system lag, users may choose to
record as they see fit. Record and playback options appear in the upper right of the
right rail. The red arrow button with the word Record under it makes it easy to start
a recording, and the green square Stop button to stop it. The trigger icon allows
users to generate an event that finalizes specific types of recordings. This enables
people to capture the events they care most about without having to record, or
review recordings from, 24/7.
And since recording is one of the most important tasks and are used to review the
recorded data, the standard play and replay controls appear in the bottom,
horizontal bar of the dashboard.
The blue rotary knob in the upper right area was originally a physical knob designed
to scroll through the data and set the sample rate. But in the final version the user
can swipe through the data and use the rotary knob icon to scroll through the
sample rate values. As it does it makes a clicking sound to pull the metaphor fully
through. This, the designer says, makes people smile when he demos it. The values
above the knob change as it scrolls. Those who don’t want to use the knob may click
a button, or select from preset values.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 53

 The lower right section displays a blue bar with arrows. This shows the selected
channel, and the arrows allow users to scroll through and choose a different channel.
The set of mostly blue-bannered buttons displays settings for that channel, for
example: mode, coupling, input, span, offset, scaling, filter, and more.

The main application after startup makes all functionality readily available
(except for some commands on the left-hand side.) The top bar is status
feedback. The main content area houses a selection of available channels and
can be configured by drag and drop.

The section in the lower right rail changes depending on what is selected. Replacing
content in the same area is one way to save space in a dashboard, or any
environment. The user may select one of the settings buttons for the selected
channel, at which point the blue square buttons representing settings go away and
are replaced with information related to that selected setting and channel.

54 [email protected] Executive Summary

 When a user taps a settings button in the settings panel (lower-right group of
the main window) the panel rotates to show the options for the selected
setting. The same is true for the acquisition control group. To keep operation
straightforward and minimize the user’s cognitive load you can modify only one
setting at a time.

Some of the commands in the left-hand toolbar (or menu) offer are a top-level item
segueing to more options. For items that have more options available, the
commands fold out. This progressive disclosure is one more way the designers
refrained from taking space away from the content area. To ensure that people can
close the fold-out item, designers added a red “x” in the upper right.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 55

 When the user taps a command which has more options on the left-hand side,
an options list will fold out, judiciously hiding commands until users ask for
them. (This toolbar is the only location that may require two taps to get to the
information.)

A user may want to look at a lot of data at once in numeric rather than graphical
meter format (sheet-based rather than a channel-based.) They may view all settings
(organized on the x-axis) by all channels (organized on the y-axis.) The tabular
format allows this channel-based viewing, and users may set and compare items.
With so many of each they cannot all fit on the small display, so users may scroll
vertically as well as horizontally when needed. The unconventional scroll control on
the right and bottom offers arrows to move, and clickable dots which also convey
status.
Users may read and compare these items, and it is also possible to easily copy the
settings in one channel to other channels via this matrix. Buttons below the table,
copy to, copy, and paste make this a snap.

56 [email protected] Executive Summary

 This matrix allows users to see an overview of all settings and channels (up to
64,) and/or modify settings. Users may easily copy settings from one channel
to other channels.

The underlying software behind the app is called Perception, which is a Windows
program, and users may swap between it and the app. Using Perception users may
access more functions to manipulate their data.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 57

 The Perception program (an underlying software for the app) allows for much
more functionality for the users.

At first glance this application appears to be quite complicated, but after a short time
the target users become accustomed to it. Each section has a compartment that
people can easily scan to, and controls for manipulating it are handy. Data is shown
real-time or can be recorded and edited. Recordings may also be exported to ensure
that data can be best analyzed and shared, no matter the computer.

58 [email protected] Executive Summary

 The users may export the data in a variety of formats.

DESIGN PROCESS

Usability Methods
The first thing you have to know about the design process for GEN2i is that the
design team was designing for a conservative market so the target audience was
looking for evolution not revolution from the existing product.
“It’s a very conservative market where we are operating in,” says Interaction
Designer Cees de Vries, “and that means that you can’t revolutionize, you can only
make an evolution of what you’re doing. So we had to put most of the legacy stuff
already in the new instrument, but also find a way of making it easier for people to
use.”
The previous version of the instrument was a result of the knowledge the team had
gained over the years from designing other software programs. This collective

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 59

 knowledge heavily influenced the mindset of the engineers when they embarked on
this project. “The more you create these kind of software packages, the more
questions you get over the years,” says de Vries. “And then you already have a kind
of image of what it should look like and what the feature set would be, in a
minimalistic way.”
One of the clear mandates in moving from the existing instrument to the new design
was to improve the information architecture and optimize the operation flow.
For example, “When using the knobs in the old version we had to go through a multi-
level menus, so if you search for a certain function which that could take you three,
four, or five steps, basically,” he says.
With a poor IA structure, users often ended up in the wrong menu and that was
something the team especially wanted to improve. The team accomplished their
goals through a number of design activities, focusing on optimizing the operation
flow. These included the following preliminary activities:
• Heuristic evaluations: The team conducts heuristic evaluations on a
regular basis with Product Marketing and other stakeholders.
Information from these regular meetings informed many of the
decisions the team made with the new design.
• Books, reports, and other resources: Learning is a continuous
activity for the design team. They frequently consult books, report and
other user-focused research to improve their knowledge and apply that
learning to the products they design. And in this case, books and
research played a particularly important part in their decision-making.
“What really contributed to the decisions was what I read in books,”
says de Vries. “That seems a little bit simplistic but I really like what
Tufte is doing. I came up with the flat interface, and the flat interface
was basically given by the iPhone, but even more so by the books from
Tufte where he explains why you should have flat interface and what
the possibilities are. And that influenced the whole design from start to
finish.”
• Sitting in on support training: During the pre-release period the
team had an opportunity to listen in on support training and gather
user feedback.
• User Comments: The company regularly collects user comments and
incorporates that feedback into product releases. These comments are
collected through a database called TestTrack.

During the design project the team used wireframes and beta testing to ensure that
the engineers had a solid design direction to work from and that the product was
well tested prior to launch.
• Wireframes: de Vries created prototypes in Photoshop to give the
software engineers a guide for designing the instrument. The
wireframes ensured that the engineers didn’t have to figure out the
basics.

60 [email protected] Executive Summary

 • Updating the IA: The team used the existing information architecture
as a jumping off point for their work. “What I did was, I took the old
instruments, we are replacing not only the one instrument but also two
other ones,” says de Vries. “And then I went through their information
architecture and just by comparing those two to three models, to see
what were the typical overlays, I figured out what would fit better.”
• Beta testing: Beta testing was conducted with a small group of sales
and marketing staff. These participants had never used touch interface
instruments before and they weren’t familiar with this new product.
The design team thought it would be beneficial for them to have
firsthand knowledge of the product before it was launched so they
were a good group to test with. The testing was done a little earlier
than traditional beta testing, with approximately 50 percent of the
software working when the tests were conducted. And the testing was
very informal. Participants just given the freedom to use the device
and provide feedback with very little direction or instruction.
• Usability testing of design prototypes: The team tested with
internal users (normally not recommended) but their testing had an
interesting twist. They used the hardware engineers to test the
software features. And for these engineers testing software was “a
whole new world” that they were unfamiliar with. “It’s a good idea,”
says de Vries, “to have them come out of their hardware knowledge
and go into the software.”
The testers were presented with different feature sets and one group
would test one version and the other group an alternate version. Then
the design team could see what went wrong and where. These
sessions were informal, more of a discussion so the design team could
not only observe how the participants interacted with the controls, but
also hear how they were thinking and how the system did/didn’t match
their own ideas of how it should work.

The illustrations below show how the interface design progressed from concept to
execution:

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 61

 The settings panel. Initially there was no panel.

Later, the design evolved from general to detailed.

62 [email protected] Executive Summary

 The acquisition panel also evolved from general to detailed. For this part of the
design, the team borrowed a number of components from the underlying
Perception software, such as the rotary knob, pixel display and gauges (colored
progress bars.) Note: All controls are “handmade”.

Marketing and sales team beta testing the GEN2i.

Development Approach
The team employed a modified Agile development method, where the team
introduced new features every few weeks (rather than every week in a traditional

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 63

 Agile environment.) The team sat together, set goals for a one three week period,
identified roadblocks, etc. and kept up this schedule throughout.
They worked at a very fast pace, with something happening every day—daily
builds— so the team had a new functional version of the software to review and test
nearly every day. And that’s a big plus for improving the product.
“You can look at the new features, the new implementations and decide right on if
it’s going on the right direction or the wrong direction,” says de Vries. “So you have
much faster turnaround with respect to implementation of features, the usability of
features, and you just try it out as soon as you get a new version and you see what’s
going on, and then try it out, if it works, if it doesn’t work.”
But that pace can be exhausting as it is productive.
“It’s a little bit more what I would say hectic,” he says. It’s very, very fast
programming.”
It may have been a lot of work, but that work paid off.

Timeline
The project took approximately one year from feasibility study to release. They
started with a lot of background information because the legacy product was used as
a foundational feature set. Then, that feature set was prioritized to accommodate the
tight timeline. Features were added and tested nearly every week until the release
date. Thought the project took a full year, the development cycle was only three
months and is outlined as follows:

64 [email protected] Executive Summary

 PROJECT DEVELOPMENT TIMELINE
Milestone Hardware Software UX
Date
Q2 2009
• Feasibility study • Design strategy
design document
Q3 2009
• Box prototype 0 • Profile
• Task analysis
• Information
architecture
Q4 2009
• Box prototype 1 Implementation • GUI Design
• GUI
Implementation
Q1 2010
• Box 0-series • Release • Finalize
candidate
• Release
• Shipment

LESSONS LEARNED
De Vries shares a few lessons learned from the redesign effort:
• Do IA up front. “I'm not completely satisfied with the IA, because we
did not do any research on that part. For other projects I did research
and then you see how relatively easy it can be. You can't get
everything in perfect but you can get 80 percent.”
• Do user testing early. “As long as it's everybody's personal opinion
you will never get a verdict, not for all features, but for features that
are important or that you can't compromise I suggest user testing. I
always find it very surprising to see what people struggle with.”
• Test with the people you have. “I use in-house [testers] depending
on the situation, because it’s hard to find people in this specific niche
area. They [the in-house testers] always like it; it's only an hour they
don't need to work.”
• Get buy-in from marketing and sales. “With this instrument, which
is relatively new, why should they try to sell this thing which is a little
bit risky. It's their bonus on the line. If you have buy-in from these
guys and they like it you’re all set. These are the guys that need to
prove that my design is right and not a shipwreck. These are the guys
that need to sell the instrument and get a bonus out of it. It is like
features: a feature not found is a feature that’s not there. When I
design a perfect instrument and it doesn’t get sold it’s worthless.”

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 65

 • Don’t skip the testing. “Do your card sorting or other technique to
get the information architecture right. I did not do this for this project
and that lead to numerous discussions without a defined outcome. Do
user testing as soon as you have wireframes or prototypes. This is also
to avoid long discussions with stakeholders about the implementation
of a feature.”

TEAM
The software group that designed the application is based in the Netherlands is
comprised of six engineers and one interaction designer. The team’s direct manager,
Peter Ackermans also played a pivotal part in the project by giving the design team
“room to breathe.”
Cees de Vries is currently a Usability Manager within HBM; during the development
of the GEN2i interface, he was HBM’s Interaction Designer. de Vries has a bachelor’s
degree in Computer Science. He originally started as a hardware/software
development engineer, and worked his way through various job roles. Five years ago
he started to develop himself in the field of design and usability. His personal ‘hero’
is Edward Tufte, who inspired him to develop the interface as a flat interface where
information becomes the interface. His favorite quote is “Don’t make me think,” by
Steven Krug.

Pictured: Members of the application design team (left to right) Marco

Dijkmans, Edwin Schild, Cees de Vries and Coen van Dongen.

66 [email protected] Executive Summary

 OEConnection

Organization: Design team:

OEConnection LLC is a strategic venture User Experience designer Shawn Kenney led the
chartered by Ford, GM, and Snap-On Business design and usability efforts, in collaboration with
Solutions in 2001 to leverage and implement other design team members who played various roles
technology solutions that help automotive at different points in the design process.
dealers manage their parts inventory and sell
more OEM parts.
Application:
OEConnection provides a user-centric workflow
that allows parts department employees at auto
dealership quickly process parts orders sent by
collision body shops, repair facilities, consumers
and other auto dealerships.
Headquarters: Richfield, OH (USA)
www.oeconnection.com

OVERVIEW
Main Function: As the automotive industry’s largest original equipment (OE) parts
marketplace, OEConnection connects more than 32,000 North American buyers and
sellers to seamlessly market, manage and move an average of over 6 million original
equipment parts transactions monthly. OEConnection is not just a single application,
but also a gateway to five integrated technology solutions that facilitate various
aspects of the parts procurement and inventory management workflow for
automotive repair shops and parts dealers.
Why it’s Unique: When serving a user population that will move back and forth
between related and linked applications it’s important to create a frictionless
experience. That level of seamless integration between products is precisely what
OEConnection achieves. Using the suite of products that combine to create the
OEConnection experience is fast and simple, and from a business perspective the
technology is also scalable.
The Organization: OEConnection is both the technology and the organization
behind the platform. It is an award-winning technology leader and innovator of
original equipment (OE) replacement parts solutions. OEConnection serves the
automotive and heavy-duty truck markets and beyond. OEConnection equips some
of the world’s largest equipment manufacturers, their franchised dealers/distributors,
and their customers with timely and accurate online parts marketing, procurement,
and wholesale management solutions.
Design Philosophy: Workers in parts departments work in chaotic environments
filled with competing priorities and a world of distractions. One of the core goals of
OEConnection is to make sure these users can seamlessly attend to all the things
competing for their attention while still allow them to complete the task at hand:
procuring parts.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 67

 USERS
OEConnection serves 32,000 users who primarily work in the parts departments of
automotive dealerships and collision and repair facilities representing 18 automotive
manufacturers across North America.
What users do with the tool depends somewhat on which application they are
working with at any given time as they move back and forth across the platform, but
the primary tasks are looking for and ordering parts and fulfilling parts orders

THE APPLICATION
It’s utterly unbelievable how helpful OEConnection is. It is testament to how studying
user tasks and mental models can inform an amazing application, in this case, for
automotive dealers managing their parts business. These professionals, if nothing
else, need to keep a thumb on the status of their parts and parts orders. Help with
this begins on the OEConnection home screen, which enables users to track, review,
and search easily and quickly. Each order is listed in a clear table with various
important criteria as column headers, including: status, order type, submitted by,
vehicle description, salesperson, and last modified. Pagination under the table allows
users to see more orders, and the search box above the table enables users to
search for something specific they recall.
Users may show or hide an additional column for an order’s Transaction ID, a unique
ID assigned to the order by the software that the Customer Care Department uses to
assist end users with troubleshooting an order.

The common order overview screen is the first screen the user would see upon
log in.

68 [email protected] Executive Summary

 On an order details page, the user may review many different fine points regarding a
specific order, and he may also perform a number of actions on the order. The VIN
number, car year and model, and shop are all displayed at the top of the page,
clearly denoting which car the information on the rest of the page concerns. The
horizontal list of links below these headings allows the user to take any number of
actions on the particular order, including: save, print (which displays a list or reports
to print along with a quick print of the most commonly used reports, as well as an
advanced print option that mimics the print dialog found in a previous version of the
app,) add part, locate, view returns, cancel order, respond to buyer (which transmits
any edits performed by the dealer back to the buyer,) and submit/ commit (which
completes the order by walking the dealer through a series of verification steps
depending on the program promotions he is eligible for.) Having these tasks all in
one place greatly expedites the dealer’s work and increases his confidence level.
The content area of the page is basically divided into three sections—Conversion
Opportunities, Buyer Requests, and Response to Buyer—and are expandable, but
collapsing by default smartly makes good use of screen real estate. The plus sign
buttons to the right are obviously for expanding the section. Within the three
sections in the content area the information is about the part, program, shop cost,
dealer cost, and part catalog information, among other things. Designers packed the
area with content, but laid it out in such a way that it is understandable and
manageable.
For each of the three sections a number appears next to the group heading. This
assists the user in understanding the work queue remaining on this order, and he
may quickly assess whether a part needs to be fulfilled.
The Conversion Opportunities section shows which parts on the order were not
specified as Original Equipment Manufacturing (OEM) and were intended for a third
party to fulfill. It’s important to keep the OEM information up to date, as the parts
represent upselling opportunities; that is if the parts are converted to the OEM
version of the part.
The Buyer Requests section shows the parts that were specifically requested by the
buyer to be filled by the dealer with genuine parts.
The last section, Response to Buyer, comprises the list of parts that the dealer has
verified to be correct for the vehicle being serviced, and presents the required
information to price the part (as the seller wouldn't want to spend time pricing a part
they won't actually be selling.) The Buyer Request and Response to Buyer sections of
this workflow are also common to other transaction based workflows.
Users may access notes and photos that were sent with the order, as well as another
piece of important content, the RO number, which the shop provides. If a customer
calls with this number it makes the order easier to locate.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 69

 The order details page displays the pertinent information about a particular
order, and enables the dealer to take several important and common actions.

The table and page layouts in OEConnection are well prioritized and limit clutter.
Headings and sections help with this, as does consistency. Designers understand the
importance of knowledge transference and flow throughout the app. Knowing that
users acclimate more quickly and successfully in UIs that have some semblance of
consistency, the team created a persistent menu across the top, consistent headings
across the top, and repurposed tables with varying content within each of the
transactions workflow processes.

70 [email protected] Executive Summary

 Menus and layout remain consistent across the transactional workflows. Any
nuances in data display exist due to special considerations based on the source
of the transaction, such as the requirements for transaction financial totals.

The Print command offers several options including Quick Print and Advanced Print.
Advanced Print offers the options to select multiple reports, and to hide pricing in the
printout. In the previous version of CollisionLink (v.4) these options were available
under a command call Print. This was confusing and the new name has alleviated the
misperceptions. The new command name maintains the ability to toggle pricing
options, support legacy users, and also prevents any semantic confusion about what
the difference may possibly be between a Print and a Quick Print command.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 71

 The print commands offer the ability to print without prices, among other
things. The Quick Print and Advanced Print commands certainly appease legacy
users.

Part numbers are extremely important when ordering automotive parts. If a part
number is right then everything falls into place. But if the user is unsure of the part
number then it can laden the order process with issues. Often times an order comes
in and the part requested is incorrect, for example, a shop requests a hood, but the
hood part number is for the wrong model. The system avoids these issues by
predicting them and aiming to reduce them. Specifically, designers implemented a
linked control relating the Part # field with the New Part # field. The New Part # field
considers the number in the Part # field as well as other information in the order,
and based on database and inventory knowledge, suggests better parts. When the
user clicks the New Part # field a drop-down displays part numbers the user may
possibly want. As an added convenience, the current part number displays at the top
of the list. Users may also manually type the new part number if they know it, and
the system will add it to the order.

72 [email protected] Executive Summary

 A linked control offers a more correct part number based on the requested part
number as well as other information in the order.

In many cases the user will want more information about a recommended part, or
even to compare the recommended part with the current one. Selecting the
Additional Information link from the bottom of the drop-down opens a matrix with
more details about each part, and the possibility to select one.
This information should cover most peoples’ needs, but in the event that the user
needs more information than what is in the table, clicking the part number in either
the Buyer Requests or Response to Buyer section will open a dialog that provides all
the information that is available but that was hidden by default. This type of
convention has been extended throughout the app so the user can always expect to
click on a part number for detailed information. And the approach itself, hyperlinking
to information, allows OEConnection to provide information that they may
infrequently reference, and it allows room in the UI to expand with new features
without cluttering the main details page.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 73

 In the event that the user needs more information than what’s in the table,
clicking the part number in the Buyer Requests or Response to Buyer section
will open a dialog that provides all the information that is available for them but
that is also hidden by default.

In another attempt to save on screen real estate, designers created the Info Center
which shows information (that used to appear constantly in the main content data
table in a past version of the app.) The categories are separated into tabs—VIN
Details, Order Info, Notes, Photos, and Event Log—which are linked to from the
menu on the details screen. This allows the user to select the information he wants
and be taken directly there.

74 [email protected] Executive Summary

 The Info Center helps reduce clutter in the main content area.

OEConnection, in this case the D2D Express app, thinks of everything and doesn’t
just drop the user when it faces any backend issues. Namely, when a part is out of
stock or backordered from the OEM, the system automatically searches other
dealers’ inventories for that part.
Given the ever-moving inventory, this system poses a level of urgency to the
ordering process: These offers are first-come, first-serve and have limited time
duration. The Overview screen provides enough information to the dealer to allow
him to quickly and confidently decide whether he can commit to selling the part in
question. And once the dealer commits, the system provides the necessary shipping
details so he may quickly box up the part and ship it to the appropriate destination.
Leaving no stone unturned, the omni-present notification system, along with e-mail,
and text alerts, helps support this process. So users will never miss an offer.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 75

 When a part is backordered the system queries all the dealers until it finds the
part and allows them to sell it on behalf of the OEM to a dealer in need.

The most common user task is simply to locate parts, so other features may be lost
on them or get in their way. Given this, designers devised the D2DLink console. The
Locate task is the main area of the screen, surrounded by Opportunities, Discounted
Parts, and Inventory Updating.

76 [email protected] Executive Summary

 The D2DLink console makes Locate [parts] the main area of the screen,
surrounded by Opportunities, Discounted Parts, and Inventory Updating.

When the user finds a part he wants the system predicts and expedites his next step,
like a busy, smart elf. Knowing what information the dealer needs in order to move
forward, the app presents the following items to him in a neat little package:
• A view of his own inventory to verify that he doesn’t already have the part in
stock,
• Primary Distribution Center results, (PDC,) which shows the current stocking
levels at one or more OEM warehouses,
• Parts that are available in the dealer community, sorted by the dealership’s
favorite trading partners, then the remaining dealers sorted by proximity.
After this appears a list of parts that the dealership has decided to buy electronically.
To boot, the system can order multiple parts both in quantity and part #’sr/sku’s
from multiple dealerships and send those to our system in bulk. And, multiple part
locate results can remain accessible simultaneously, which reinforces the notion of
having multiple transactions open in other workflows. This is important because
often times dealers are looking up a part for a walk-up customer at a counter, while
looking up a second part for a customer on the phone, while looking up a third part
on behalf of their own service department. The ability to maintain these queries
without having to start and stop is vital to the parts department’s efficiency.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 77

 The system predicts and expedites the user’s next step, like a busy, smart elf.
Knowing what the dealer needs to know before he can move forward, the app
presents important information to him in a neat little package.

Continuing with momentum, the dealer may select Create Order from any set of part
locate results, then verify their billing and shipping information, verify the correct
parts and part quantities, then submit the order. Once submitted, a new transaction
is created and sent to the selling dealer.

78 [email protected] Executive Summary

 Create Order leads to verifying billing and shipping information, verifying the
correct parts and part quantities, then submitting the order.

To construct a search, users may access the locator, which is available in all
workflows in the same location in the upper right area of the screen. Once a part has
been entered the user is redirected to the D2DLink workflow and a new subtab with
results. The ability to search for a part at any time no matter what the current focus
is is compulsory to the users’ success. Thus, designers made it possible to initiate a
search to locate parts from CollisionLink. This automatically redirects the user to
D2DLink and creates a subtab for the part number entered. This type of
accommodation shows a true understanding of user workflows, and does not limit
the user based on the app or view he happens to be in when he wants to initiate a
task. Additionally, designers implemented the safety net of auto-save as a
background task.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 79

 To construct a search, users may access the locator, which is available in all
workflows, in the same location near the upper right hand corner of the screen.

OEConnection designers were disciplined and thoughtful when considering screen

real estate. They frequently prioritized pages and workflows based on users’ tasks
and mental models. They made mindful tradeoffs in many cases where they chose to
hide information by default, eliminating clutter on the screen, but making the deeper
information available and easily accessible with the look of an eye and the click of a
mouse.
As for workflow and considering users’ next steps, OEConnection is like the fairy
godmother of apps, making suggestions and offers, and doing all the backend work,
coming close to perfectly automating everything for the user.

DESIGN PROCESS
Prior to the most recent redesign the OEConnection products were “really hard to
use”. “They were actually kind of brutal,” says Shawn Kenney, User Experience
Designer. So, it was clear that in order to really fix what was broken would require
the company to re-build from the ground up.
One of the first steps in the process was to achieve a level of integration across the
products that had not been successful before.
• Site Visits: Hanging Out in the Parts Department: On site visits played a
major role in the product redesign. And these site visits happened at various
stages of the design process.

80 [email protected] Executive Summary

 Prior to Kenney joining the design team, the company’s development
manager, business lead, and product specialists went on the road to visit
dealerships around the country to “hang out” and watch how the parts
departments work. Through these visits they were able to get a really good
understanding of the use cases, decision tree, what's important to users, and
what steps a typical user needs to take to perform a task. These initial
insights contributed to the CollisionLink tool.
Site visits also contributed to shattering the team’s misconceptions of how the
users work. “Observational learnings are the best,” says Kenney. “We had a
misconception that everybody had a unique process but what we found that
they're mostly the same, with maybe only a few geographic variations.”
This was helpful because it contributed to the task analysis for the tool, but
also because finding homogeneity across the dealerships meant the team
could use almost any dealership’s staff as a proxy for the greater user base.
“For the most part we can find very specific dealers, observe them, and
validate against a broader set if we need to,” he says.
• Relying on Experts. Kenney was brought to the team as much for his
expertise as for his ability to execute on design concepts. That expertise
played heavily in the design process. “One of the things I brought was an
understanding of the basics of what works from an interaction standpoint,” he
says. “That we’d been able to layer that on top [of the application] but we
don’t spend a lot of time with taking a design, putting it in front of a user,
setting up cameras, nodes on their forehead watching them, etc.” The team
relied instead on equal parts expertise and iterative prototyping. For as much
as Kenney was a usability expert, the team felt the users were the real
experts on whose advice they should rely.
“We can for the most part take paper prototypes of the design and set it on a
table in front of them [the users] and actually have them explain back to us
what they are seeing. And a lot of times we don’t even have to have to
explain it to them, their eyes just light up and they ask when they can have it
because you can see that the light bulb went off when you talk with them. So
we try to spend as much time as possible truly understanding how to think
about their job.”
• Using Prototypes: Kenney used Flash prototypes to demonstrate any radical
design changes, such as those in CollisionLink. “We actually made a Flash
prototype demo to seven dealerships and had them play with it to make sure
we weren't going too far astray,” he says.
Paper prototypes were used throughout the process as a tool they could put
in front of users to validate design choices. The prototypes were used to help
both the internal team and users understand how the pieces of the unified
product fit together. And they really worked. “They allowed us to take
inventory of concepts we may have overlooked, and discover new feature that
we could introduce.”

Tracking the Design Evolution

The following screens illustrate the products as they progressed through various
design states as the project progressed:

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 81

 This is a view of a legacy CollisionLink order screen.

82 [email protected] Executive Summary

 This screen shows an example of a wireframe created to show the transactions
details in CollisionLink. At this point in the design, the team was still using a tab
paradigm to isolate the state the transaction came in as, and then a worksheet
view that contained the changes being made. During the prototype phase they
learned there was no need for the dealer to see the state the view came in as,
so that was eventually dropped. Eliminating unneeded features is one of the
best ways to increase usability — and also obviously saves on development
costs.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 83

 This screen shows the results of a design study conducted by an external
agency, prior to Kenney’s arrival at the company. It was aimed at the D2DLink
parts locator. The company had been looking for a design framework to use for
the new CollisionLink and used this as a starting point. Some basic elements
were carried over from this design while some concepts such as “Find it Fast”
and “My Dashboard” were shelved, and the tabs, icons and other visual
elements were created new.

84 [email protected] Executive Summary

 This screen shows an early attempt to merge together the third party design
study with the new workflows Kenney and his colleagues developed. When this
screenshot was captured they had yet to integrate application specific branding
or the idea of multiple products running side by side, the ability to have
multiple orders open at the same time or what would eventually become the
organizational strategy around the left nav.

LESSONS LEARNED
Kenney shares his experience as the “new guy” joining a team within an organization
that had well-established ways of getting things done:
• First unify - then act. When Kenney came on board the design philosophy
of the organization was in transition. Kenney’s addition to the team was not
merely the addition of a new skill set, but the addition of a person who would
have to facilitate a change in the way people were used to doing things before
any real design change could happen. “We had everyone with their own ad
hoc processes and it was based on whatever experiences they’ve had,” he
says, “whether it was in use here or at other companies and no one to get
them into alignment. Not only to just be consistent in their own process, but
understanding the decisions that they make impact other teams.” So, that
responsibility fell to him and he learned a lot about the politics of change
through this process.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 85

 “When a company is going through cultural change - before you do anything
understand the level of communication and training required and get all your
pieces in alignment before you begin to design and develop because that
change isn’t going to happen immediately and it’s going to take time on a
person to person level. You can do all the design/usability work you want but
you’ll essentially be fighting against people who really should be your allies
and you probably won’t get too far.”
His advice: slow down. “Slow the process down just to get people on the
same page.”
• A rush to market doesn’t save time. “So, there’s always going to be this
desire to get to the market as fast as possible but if speed is your primary
driver to market you are going to cut corners,” says Kenney. “You are going
to miss steps, things are going to slip through the cracks. And if you don’t do
all the right steps up front, you’re going to have to do them at some point,
typically those things come up in the development process because
developers will see the holes that the product teams have missed and they
are either going to go to the product team for answers or they are going to
come up with them themselves. You’re introducing re-work, burned bridges
and hurt feelings, so there’s that whole human dynamic too, that you can
avoid by just taking your time up front.”
• Your place in the org chart affects your ability to affect change.
Kenney learned the hard way that a lot of decision-making hinges on the
perception of who’s driving the change. When he joined the company he was
placed in the technology department, which is okay except at OEConnection
there is a perception that the business units say what the need is and the
technology department determines the “how”. “It proved to be very
inefficient,” he says. “One, because our business units didn’t really like
someone in the technology unit basically changing the requirements on them
and introducing a new level of requirements around users. And then also I
was in a position where I was actually trying to hold my own management
accountable for the quality of the product going out the door which they didn’t
really appreciate either.”
Currently he reports into the company’s Chief Commercial officer, which he
says, allows him to do two things, “One, it puts me on an even playing field
with our business leads,” he says. “That way I can actually work with them as
a peer and have in-department discussions when I need to have those. Also
since I report directly to their boss so, in the event that we have walls that we
need to have removed I now have someone who’s in a position to remove
those walls for me and I’m not trying to negotiate or twist arms or bribe or
throw my hands in the air and call it a day because I can’t remove those on
my own.”

Project Team
Kenney was the primary designer and user experience lead for the OEConnection
redesign project, but other team members performed specific functions throughout
the process. Here is a list of the team members and their roles:

86 [email protected] Executive Summary

 • Shawn Kenney, User Experience Designer: Responsible for disseminating
team feedback, participating in onsite user research, identifying user mental
models, creation of wireframes and workflows, facilitating design sessions,
creation of mockups and functional prototypes, participating in validation
sessions, creating usability & design standards, and ensuring compliance with
existing company usability and design standards.
• Chris Fuss, Project Manager, FordParts: Responsible for project goals and
timelines, communicating business objectives, identifying test markets for
workflow validation and organizing validation sessions.
• Erik Enright, Project Manager, CollisionLink: Responsible for providing
feedback gathered from users in field, maintaining project timelines,
scheduling design sessions, communicating design directions to executive
leadership.
• Jeff Toom, Product Specialist, CollisionLink: Responsible for providing
feedback gathered from users in field, contributing OEM requirements to
project
• Elizabeth Fontes, Product Specialist, CollisionLink: Responsible for
providing feedback gathered from users in field, contributing reporting
requirements to project
• John Haluch, Account Services Representative, CollisionLink:
Responsible for contributing dealer feedback provided by users of legacy
applications, understanding of existing user pain points.
• Carol Mandato — Product Manager, D2DLink/D2D Express: Responsible
for project goals and timelines, communicating business objectives,
participating in onsite user research, participating in design sessions,
communicating design philosophies and direction to executive leadership,
identifying user groups for workflow validation and organizing validation
sessions.
• Erica Holvey, Product Specialist, D2DLink/D2D Express: Responsible for
onsite dealer user research, documenting existing user pain points in Legacy
applications, participating in and contributing industry knowledge to design
sessions, validating designs with users.
• Sara Roubenes, Product Specialist, D2DLink/D2D Express: Responsible
for onsite dealer user research, participating in and contributing industry
knowledge to design sessions, validating designs with users.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 87

 Pictured: OE Connection team members (front row — left to right) Carol
Mandato, Elizabeth Woolf, Jeff Toom and John Haluch; (back row — left to
right) Erica Holvey, Shawn Kenney, Chris Fuss, and Erik Enright.

88 [email protected] Executive Summary

 Lightweight Applications

Applications in the Lightweight area typically have a deliberately limited feature set
and/ or may deal with a jocular, fun topic or task. The biggest challenge in creating a
lightweight application is not in deciding what to include, but in deciding what to
leave out.
These types of applications are often optional, that is, users can take them or leave
them. And users are all too likely to leave an app that doesn't instantly explain itself
and immediately enrich the user’s experience or make the users' activity easier.
Given these criteria, lightweight apps must be thoughtfully designed, responsive, and
easy to use, which is no short order, and these applications are not easy to design.
In fact, designers of applications in this arena need great discipline, clear goals, and
a deep understanding of what their users want and how they think. Designers must
also have the authority to strip away the superfluous and focus on making only the
necessary and important features. (It's not surprising that some of the best
lightweight apps come from small, nimble organizations, that don't have to worry
about too many cooks in the kitchen.)
Some UI elements that work well in this type of application include:
• Can't-miss affordances for primary actions
• Minimizing navigation elements or creating linear flows to focus on
current actions
• Ruthless pruning of clutter and unnecessary features

Winning applications in this category include:

• CycleStreets Mobile, helps cyclists plan safe, fast routes for cycle
journeys in the United Kingdom (and some other areas) using crowd-
sourced map data.
• VitreaTeach, an educational tool that radiology residents and practicing
radiologists can use to save, manage, and share medical images.
• Wakelicious, alarm clock that wakes you up by playing YouTube videos
sent by your friends.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 89

 CycleStreets

Organization: CycleStreets Ltd. and Anna Design team:

Powell-Smith • Martin Lucas-Smith, Director and Project
Application: Bicycle Route Planning Manager
• Anna Powell-Smith, Developer
Headquarters:
CycleStreets Ltd—Cambridge, United Kingdom
Anna Powell-Smith—London, United Kingdom
http://m.cyclestreets.net

WHY THEY WERE CHOSEN

CycleStreets Mobile is an excellent lightweight application because of its:
• Innovative solution to the widespread problem of small targets on touch
interfaces
• Elegant design, based on a deep understanding of the problem space
• Tight focus on core features, which makes the app flexible enough to work on
multiple platforms.

OVERVIEW
CycleStreets Mobile is designed to helps cyclists plan the safe, fast routes for cycle
journeys in the United Kingdom (and some other areas) using crowd-sourced map
data.
Why it’s unique: Many applications offer maps and route planning, but CycleStreets
Mobile is uniquely optimized for a specific task and audience. This lightweight
application is designed for cyclists on the move, and thoughtfully leverages the
characteristics of cycle journeys to create a fast and direct interface. CycleStreets
Mobile is delivered as an HTML5 website in order to make the tool accessible to all
mobile users, regardless of device or platform. Designers took advantage of the
constraints to create a simple and elegant interaction for the users’ core tasks.
The organization: CycleStreets is a not-for-profit organization created by cyclists
for cyclists. With just two employees, CycleStreets has built a sophisticated route-
mapping service and contracted with external developers and volunteers to build four
user interfaces: a full website, an iPhone application, an Android application, and a
mobile website application. Since its launch in September 2011, the CycleStreets
mobile application has been used to plan 322,000 kilometers of cycle journeys.
Design philosophy: Throughout the design of the mobile interfaces, CycleStreets
has been guided by the idea of quick and direct interaction: that a user should be
able to plan a route simply by marking the start and end points on a map. The
designers call this the “three-taps” interface: Start, Finish, Plan!

90 [email protected] Executive Summary

 The CycleStreets Mobile route planning page allows users to interact directly
with a map to set the start and end points for a journey.

WHO ARE THE USERS?

CycleStreets aims to serve people who are already active cyclists as well as those
thinking of taking up cycling, but the mobile site is primarily intended for current
cyclists who are about to begin or are in the midst of a journey. There are two main
types of mobile users:
• Beginner cyclists
• More experienced cyclists.
Martin Lucas-Smith, the Director of CycleStreets, explains, “From the feedback we
get, there are these really two distinct groups of people: those for whom we’ve
removed a barrier to starting cycling; and those who say, ‘Gosh, I’ve been cycling
around here for years and never found that cut-through and it saves me two minutes
of my journey away from this horrible road.’"
CycleStreets Mobile also addresses access gaps left by Android and iPhone
applications. While these native applications currently offer richer interaction, the
universal access offered by the HTML5 mobile web application is a significant benefit,
especially for a not-for-profit organization with limited development resources. At the
time of this writing, 322,000 kilometers of cycle journeys planned.

THE APPLICATION
CycleStreets Mobile is an HTML5 web application, optimized for mobile devices. The
routing engine helps cyclists plan the fastest, safest bicycle routes using map data
from the OpenStreetMap project.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 91

 Users can plan a journey by selecting start and finish points directly on a map, or by
searching for an address or place name.
The routing algorithm finds journeys that avoid highways and hills, and prefers cycle
paths and quiet roads, quite different from car-oriented routing. Cycle routing in the
UK (where CycleStreets is based) is particularly challenging, as there are few grid-
based city layouts.
In addition to distance traveled, the journey plan also displays estimates of the time
to complete the trip, calories burned, and CO2 saved by cycling rather than driving.
CycleStreets can display the route against three map backgrounds: OpenStreetMap,
OpenCycleMap, or Ordnance Survey. Ordnance Survey open data maps are official
government maps, while OpenStreetMap is a more data-rich, crowd-sourced map.
OpenCycleMap displays OpenStreetMap data, but highlights of official “cycle routes”
and other points of interest such as cycle parking, while de-emphasizing major car
roads.

CycleStreets Mobile displays route information such as estimated time, calories

burned, and CO2 saved; as well as giving users the option to switch between
faster or quieter routes, and to display the official government maps (OS) or
the crowd-sourced OpenStreetMap or OpenCycleMap versions. Turn-by-turn
directions are also available.

In addition to the journey planner, CycleStreets includes a “Photomap” area which

lets users upload geotagged photos. The Photomap is intended primarily to
document cycling-related infrastructure issues, which can then be brought to the
attention of authorities.

92 [email protected] Executive Summary

 INNOVATION
Most mapping applications, including previous versions of CycleStreets, let users
select any point on a map by tapping the desired location with a finger, or tapping
and holding to drop a pin. These methods lead to the “fat finger” problem: users
have trouble selecting small targets, because their fingers obscure the entire target
they are attempting to select, and they end up selecting the wrong item or not
selecting anything. This is bad enough when dealing with small buttons or text, but
for mapping applications the situation is even worse since the entire screen is
selectable and tapping just a few pixels away from the intended target can generate
a completely unintended route and set of directions. To make an accurate selection,
users are forced to either tap several times until they get the right point, or use
another gesture to zoom in on the map and make their target location larger.
The CycleStreets mobile application provides an innovative solution to this
interaction problem: The user is presented with a crosshairs indicator overlaid on top
of the map image. They can touch any part of the background map and drag it to
move their desired location into the center of the crosshairs, and then tap a large
button at the bottom of the screen to select that location. This interface neatly
avoids the necessity of blindly selecting a screen location that is hidden behind your
finger.
This unique selection method is especially appropriate for mapping cycle journeys
because they are generally shorter than car journeys. Users planning car journey
need to zoom in enough to see details such as street names, and then zoom out to a
larger scale to see the entire journey and endpoint. With shorter bicycle journeys
where both the full route and detailed street names can be seen in a single screen,
the only reason for zooming is to make it easier to select endpoints accurately. Thus,
the alternative selection method that does not require zooming significantly
simplifies the interaction for bicycle journey planning.
Although this interaction method is unfamiliar for many users, they quickly learn how
it works because as soon as the map loads the crosshair selector and button are
visible. These interface elements act as a cue to immediately let users know what
they should do. In contrast, applications which rely on a tap and hold gesture to
select a location usually just display a map with no controls, and assume that all
users already know how to make a selection. CycleStreets found that on their iPhone
and Android applications, which did not provide a crosshairs interface, some people
didn’t immediately understand how to use the map, as Lucas-Smith explains,
“Because there’s nothing much on the map that says ‘We’re waiting for you to
actually do something.’ People often would just accidentally tap on it, and then a
mark appears, and then they realize, ‘Oh I’m supposed to click on the map and set
my point.’ The crosshairs just seems to remove that problem.”

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 93

 Left: The beta version of CycleStreets Mobile required users to tap and hold a
point on a map in order to select that point. Users sometimes accidentally
selected the wrong location while they were trying to move the map.

Right: The final launched version of CycleStreets Mobile makes selecting an

exact location easy by letting users move the background map so that the
appropriate location is beneath a central crosshairs interface, then tap a button
to make the selection.

94 [email protected] Executive Summary

 Because cycle journeys are generally shorter than car journeys, both the route
overview and the map details can often be seen on a single screen, without
moving the map or zooming.

CONSISTENCY
While the map selection interface is different, the controls and labels for navigating
around the application itself are ones that most users will already be familiar with,
including the following:
• The Home button in the top left corner of all pages (except the homepage
itself)
• The settings screen, linked from the homepage with a gear icon and the label
Prefs
• A blue dot which appears on the map to indicate the current location, as
determined by GPS data.
These standard controls are instantly understandable, provide persistent access to
core features, and take up very little screen real estate.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 95

 The home button located in the top left, and the blue dot which indicates the
current location, are consistent with many other interfaces.

CycleStreets Mobile also provides excellent visual feedback to users throughout the
journey planning process. Once a start point is set, a distinct flag appears and the
interface labels change to indicate the next step, setting a journey end point.
These small details make a big difference in users’ ability to discover and take
advantage of application features. Ideally, all applications should be easy for new
users to understand and learn. This principle is even more important for web-based
applications such as this which are aimed at a broad audience, as they can expect
many first-time and casual users. Clear task labels rather than cryptic icons, and
persistent, understandable navigation controls are essential for such tools.

96 [email protected] Executive Summary

 Before (left) and after (right): Clear visual indicators let users know when they
have selected a location.

FLEXIBILITY
Because mobile devices are generally used by just one person who carries the device
around almost constantly, they lend themselves to more personal, customized
experiences. CycleStreets supports this dimension of the mobile experience by
letting users easily control and customize the journey planning process.
For example, some riders may be quite cautious and prefer to always take quiet
streets, even though a busier street might be quicker. CycleStreets mobile allows
users to set their default route type preference as well as their normal speed in the
Preferences area. They also have the option to turn off the GPS location feature, in
order to conserve battery life.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 97

 Users can set defaults for route type, map type, and estimated speed in the
Preferences area.

Preferences for route and map types can vary not just by individual but also by
situation. A single person may prefer busy, direct routes when she's in a hurry and
quieter trips when journey time is less important. CycleStreets mobile
accommodates the need to adjust these preferences on the fly by providing buttons
that let users quickly switch between quiet, fast, or balanced journey plans. The
controls for adjusting the route type appear directly on the journey map, and
because there are only three controls, it is obvious which option is currently selected.
The route switching controls on the CycleStreets Mobile application are a significant
improvement over the CycleStreets iPhone and Android native applications. On the
iPhone application, users can only select a route type before planning a journey; the
Android app does allow users to switch between route types, but they must use the
device's physical menu button (which users often overlook) in order to access the
route type menu.

98 [email protected] Executive Summary

 The CycleStreets Mobile website displays three tabs on the journey display
screen which let users quickly switch between quiet, fast, and balanced routes
to compare their options and select the best plan.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 99

 The iPhone application currently requires users to select a route type in the
application settings before planning a journey, which many users may never
discover as the (dialog box) modal explaining this quirk appears only once,
after planning your second journey. Comparing route types for the same start
and finish points is therefore not easy.

100 [email protected] Executive Summary

 On the CycleStreets native Android application, users must use the device's
physical Menu button in order to change the route type.

CycleStreets Mobile also gives users a choice of different input methods when
planning a trip. The crosshairs interface makes it easy to quickly set start and
endpoints on a map, but users can also search for a specific address or return to a
previously saved route.
Any route planned by a user is automatically saved without requiring any login or
explicit save command by the user. This model (dialog box) works well for an
application intended for personal mobile devices generally used by just one person.
Since the application is delivered through a website, it is subject to interruptions in
internet access; automatic saving helps users recover from network interruptions by
quickly recreating routes.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 101

 Users can choose between planning methods: by map selection, by address
search, or by using previously saved routes. Users may also select their current
location as a starting point.

INFORMATION DISPLAY
The CycleStreets Mobile application offers users a wealth of geographic information
in the form of three different maps:
• OpenStreetMap (OSM) is the “Wikipedia of maps” and includes crowdsourced
information about routes and nearby businesses
• OpenCycleMap (OCM) uses the same underlying data as OpenStreetMap, but
is optimized for cyclists including highlights of officially designated cycle
routes and hills
• Ordnance Survey (OS) includes open data maps, produced by the official
United Kingdom government mapping agency.

Each of these three options offers a unique value to the user. The crowd-sourced
geographic information on the OpenStreetMap and OpenCycleMap displays are
invaluable to cyclists interested in finding shortcuts that don't appear on official
maps. However these maps can sometimes be quite cluttered with information and
official routes, making it more difficult to see the basic geography. CycleStreets’
designers found it challenging to come up with a single route display that works well
against all three backgrounds. Many mapping applications display a map in the
background and highlight the route plan on top, but the OpenStreetMap and
OpenCycleMap displays already make extensive use of different colored highlights to
indicate major roadways and cycle paths.

102 [email protected] Executive Summary

 Instead of adding more highlighting to an already busy visual field, CycleStreets
Mobile marks the planned route with a thin but bright, saturated violet line. In
another concession to the complexity of the OpenCycleMap display, CycleStreets
shows only one route plan at a time, rather than attempting to display all three types
simultaneously. Lucas-Smith explains, "If we had drawn all three routes on the
screen, you have the two highlight colors as well as the general map base, and you
then have to have three very prominent colors that stick out, for balanced, quietest,
and fastest. It was very clear on the mobile that that really wasn't going to work
very well."

The OpenStreetMap (left), OpenCycleMap (center), and Ordnance Survey open

data (right) cartography designs offer different degrees of detail and cycling-
specific information.

DESIGN PROCESS
The CycleStreets Mobile designers conducted informal usability evaluations, simply
by asking friends to plan a journey using the mobile website. The underlying UI
design was based on collected user feedback about the other CycleStreets interfaces,
which provide the same functionality on different platforms. "… We’d already learned
a lot of the problems in the previous applications,” says Lucas-Smith. For example,
iPhone users find it difficult to switch between fast and quiet routes, since the iPhone
application requires you to start the journey planning process over at the beginning
in order to change the route type. “The HTML app is actually the easiest of all four in
terms of route switching. We made sure that this app dealt with all of that learning
over what’s been about three years of iterations of products," says Lucas-Smith.

Implementation
The need to make CycleStreets Mobile operate on many different platforms was one
of the biggest challenges of this project. Performance testing and bug fixing for
different platforms added complexity to the development process, especially since
CycleStreets as a non-profit did not have the resources to purchase different testing
devices. The need to deliver the application via the web also significantly influenced
the design.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 103

 For example, since additional features multiply the device compatibility issues, the
designers kept a tight focus on implementing just the core features essential to plan
a cycle journey. For example, advanced functionality such as the ability set the
starting point via map but use search to find the endpoint is not supported. Future
iterations may add this feature, but the time needed to support this complex feature
through the web (and make sure it works on all devices) put it out of reach for the
first version.
The addition of the crosshairs for map location selection actually simplified the
application code as well as the user experience. The beta version, which allowed
users to select any point on the screen as a starting point, required the application to
determine exactly where the user was tapping on the screen, but the crosshairs
interface only needs one preset screen location: the exact center, where the
crosshairs are focused. "Suddenly all the code to determine where exactly the person
is tapping on the screen disappears, because now when you click the button that
says Click to set the web browser knows exactly where it is because it's just the
centerpoint," says Lucas-Smith.

JQuery Mobile Framework

The use of HTML and JQuery Mobile rather than a native application toolkit influenced
the design in other ways as well. Although both the iPhone and Android CycleStreets
applications include a persistent menu bar, the JQuery Mobile framework was not
optimized to support this structure. Instead, the mobile website features a
homepage with five main navigation options, and all other pages on the mobile
website include a Home button rather than a full menu bar. This structure offers less
support for switching between tasks, but also leaves more screen space free to
display maps and routes.

Standalone Code
For the time being, the mobile website is a standalone site, and does not rely on any
of the main website code. “We took a very deliberate decision to do it as a
standalone thing because the source code for the main website was not in a state
where you could easily repurpose it for mobile. Also it was very clear from our
experience with the Android and the iPhone applications that there’s just too much
extraneous detail that you’d have to work quite hard to get rid of,” says Lucas-
Smith.

LESSONS LEARNED
The CycleStreets Mobile design team benefited from a few key practices during this
project. The resulting advice is descried here:

Limit your focus

The CycleStreets Mobile designers quickly realized that it was essential to focus on
core tasks and eliminate the nice-to-have type features. "We found that we had to
cut it down quite a bit compared to native apps: keep it as simple as possible and
make sure that some of the really irritating things with native apps were dealt with,
particularly the switching between routes and the problem with the thumb," explains
Lucas-Smith.

Track and learn from your mistakes

104 [email protected] Executive Summary

 Documentation of issues with previous applications proved to be an invaluable
resource during the design of this version. Lucas-Smith says, "I think we were very
lucky in this as it was the fourth time we’ve gone through making an interface to do
this job, so it wasn’t like traditional web application testing where you’ve got a
completely new thing and have people testing it out to see where the problems are.
We kind of knew what the problems were because we’d already had a lot of
feedback."

TEAM
The design team included:
• Project Management: Martin Lucas-Smith
• Developer: Anna Powell-Smith
CycleStreets Ltd. was created and is managed by Simon Nuttall, the lead developer,
and by Martin Lucas-Smith, who serves as the organization’s Director (in addition to
his full-time employment as a web developer at the University of Cambridge.) Both
are regular cyclists based in Cambridge, who have been involved with the Cambridge
Cycling Campaign, a local bicycle advocacy group, for many years. They plan to
release the CycleStreets code as an Open Source project and enable others to get
involved. The code of all the mobile apps has already been released.
Anna Powell-Smith is a Cambridge alumna and freelance front-end developer now
based in London.
With the team members based in different cities, most collaboration was done
remotely, except for a kick-off meeting which took place on a train between
Cambridge and London, and a curry dinner in Cambridge where the crosshairs
concept was conceived by Tom Steinberg, an acquaintance of the CycleStreets
designers.
So never turn down the opportunity to partake in a meal, especially a curry, with
great thinkers as wonderful things may come from the experience.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 105

 VitreaTeach

Organization: Vital Images, Inc. Design team:

• Navid Sadikali, Principal User Experience
Application: View and share medical images
Designer
Headquarters: Minneapolis, MN • Thomas Myers, Senior Manager, Strategic
Technical Marketing

OVERVIEW
Main function: VitreaTeach is an educational tool that radiology residents and
practicing radiologists can use to save, view, annotate, share, and present medical
images and case descriptions in order to support learning and teaching.
Why it’s unique: Normally, 'knowledge management” and “lightweight” are not
terms you hear in the same sentence. VitreaTeach breaks the mold by approaching
its complex subject matter with a simple interface that focuses on how users actually
behave, getting out of their way to make the content more visible and usable. The
application emphasizes the personal and social nature of information management,
especially in an educational context. It’s especially refreshing to find this approach in
the healthcare field, which is often dominated by confusing, inefficient interfaces.
The organization: Vital Images, now part of Toshiba, builds medical imaging
systems. The new in-house design team is focusing on driving product strategy
through great user research and design.
Design philosophy: Applications intended for physicians must offer significant
benefits and very low barriers to entry in order to be adopted. The healthcare field in
general, and Vital Images in particular are undergoing a transformative realization:
advances in medical practice have been limited not by technology but by poorly
designed tools that are ill-suited to the way medical practitioners understand and
share information.

WHO ARE THE USERS?

VitreaTeach has two target audiences with distinct goals and behaviors:
 Radiology residents: physicians in training who are highly focused on
learning from the dozens of cases they encounter each week
 Radiologists: experienced physicians who may play a significant role in
educating residents.
Two primary personas represent these user groups and helped guide the design of
VitreaTeach: Amanda, the resident; and Mack, the Chief of Radiology.
The Amanda persona was created specifically for this project, but Vital Images had
already been using the Mack persona for other products. Although VitreaTeach is
primarily intended for Amanda, it’s also a useful tool for Mack.
The design team initially thought that the learning process in radiology would follow
a student-teacher model, with supervisors assigning sets of cases to residents to
study. After spending time with users, they discovered that, in fact, most teaching is
informal and rapid, not explicit or managed.

106 [email protected] Executive Summary

 Amanda, Radiology resident

Wants “iTunes meets Google”

Goals:
• Learn what is important
• Avoid mistakes
• Survive

Residents work with real patients and cases, but their focus is on learning. They may
encounter dozens of cases each week, and participate in discussions or give
presentations about cases. Existing tools are optimized for maintaining patient case
files and interacting with different departments in the hospital, but offer little support
for managing that content for learning purposes. When Amanda sees an interesting
case, she needs a way to save it so she can access it again later, take some notes
about it, and share it with her colleagues or supervisors. Currently, this task of
saving interesting images relies on residents writing down the patient ID numbers of
interesting images, then later retrieving the images and perhaps burning them to a
CD.
Residents are also accustomed to highly-interactive, lightweight applications such as
instant messaging and Twitter. They want intuitive, flexible interfaces that give them
the freedom to use content in a variety of ways; for example, Amanda might like the
ability to quickly create groups and invite others to join, so she can easily share
cases with the other residents in her rotation.

Mack, Chief of Radiology

Goals:
• Be the vision holder
• Look good with peers
• Make sure his staff and residents are well served

The educator persona is very different from the resident; not only is he already an
expert in radiology, but he may not even be practicing much on a day-to-day basis.
Instead, he focuses on managing the department, which includes helping the
residents progress, and sometimes actually quizzing them. He may also be
interested in publishing an institutional library of interesting cases, such as “10
Common ER Trauma Cases.”
The organization is awaiting pending FDA approval.

THE APPLICATION
The early plans for VitreaTeach focused on the content—interesting radiology
images—as a valuable resource. The idea was to encourage users to upload such
images into a cloud-based application, and then later the team could determine how
to monetize that content.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 107

 This content-centric focus, however, led to the development of an application that
was too hierarchical and complex for physicians to easily incorporate into their
normal workflow. Vital Images works with medical professionals to test and evaluate
software. During these interactions the team learned invaluable lessons about their
users. For example, a senior physician who had a longstanding relationship with the
company and was one of the early users of VitreaTeach commented, "The way it is I
don’t really want to use it…It's too cumbersome. It's not what I need."
User-centered design methods such as observational research helped guide
VitreaTeach in a new direction: towards an interface focused on supporting social
interaction and organizing content in a way that is personally meaningful to users,
rather than by a universal hierarchy.
VitreaTeach integrates with existing hospital software where the target users do
most of their day-to-day work. When they come across an interesting image, they
can save it to VitreaTeach by right-clicking on the image.
Once in the VitreaTeach system, users can
• comment on images
• create unique tags for images
• share images with others
• present cases to others.

With this type of content privacy is a major concern. To protect patient privacy,
VitreaTeach changes patient names and blocks other identifying information,
rendering the content anonymous. In this way it’s possible for practitioners to view
images from other institutions without violating patient privacy rights.

108 [email protected] Executive Summary

 VitreaTeach lets users save, annotate, and share interesting medical images.

INNOVATION
Extremely sophisticated classification systems have developed over time to describe
medical information. For example, medical images can be categorized according to
the part of the body they depict or by the pathology they portray. While these are
both valid taxonomies, they introduce a level of complexity that is simply not
necessary for most of this application's target audience. Navid Sadikali, the principal
user experience designer of VitreaTeach, explains, "The existing process, with a huge
taxonomy was an important thing to get rid of. The medical dictionary is so big,
people don't even use it."
Instead of this top-heavy approach, VitreaTeach employs a flexible system that is
much more related to users' actual tasks. The Anatomy and Pathology hierarchies
are still present, but are deemphasized in order to focus on classifying images
according to social or personal dimensions.
For example, the dominant hierarchy is set up to match the radiology resident's
mental model of his own activities and the people he interacts with closely and
tangentially. Three levels display the distribution of images:
• Individual: My Cases includes only cases saved by the user
• Institution: Cases saved by everyone within the teaching hospital and
subgroups of the institution
• Public: Public Cases are those saved by anyone at any institution into
VitreaTeach.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 109

 Using observational research we realized that teaching meant tracking interesting
images from old cases for self-learning, more than formal top-down case libraries for
teaching.
This structure corresponds closely to the primary users' activities. The design team
observed radiology residents and found that rather than using formal top-down case
libraries, residents engaged in self-education by reviewing interesting images from
old cases. Thus, cases a user has personally saved for her own use are the most
important types of content. The next most frequent use would be to share images
with small groups of colleagues, such as the other residents in her current rotation.
Finally, an important but less common usage would be to review an institution wide
library of cases, or cases from another institution.
VitreaTeach also includes a system of tags and filters which provide a
complementary way of browsing through images. Users can create their own tags to
describe content according to its personal relevance. For example, a user could
create a Present tag to describe images he plans to use in a presentation. This
flexibility is a game-changer, because suddenly the application has become a tool
customized to serve the user's tasks, rather than a tedious, dense hierarchy he must
navigate.
So, for example, a user could browse to My Cases in order to see just the cases she
personally saved, then click on a tag she had previously created such as Discrepant
Findings to filter the content to only those bearing that tag.
Both the hierarchy and the user-generated tags remain visible in the left navigation
pane, making it easy to see which subset of content is currently visible. "Users loved
this screen and couldn't believe it was a web-based application," says Sadikali.

110 [email protected] Executive Summary

 Cases can be filtered by their associated groups or by custom user-generated
tags. Here, the images are filtered to display only My Cases which have been
tagged with Present.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 111

 An early prototype of VitreaTeach planned to combine anatomy and pathology
in a traditional hierarchy. This concept was discarded after users loudly rejected
it (although it is the same model of a few existing similar products.)

The first version of VitreaTeach used folders as a visual metaphor and devoted
a great deal of space to listing other institutions.

FLEXIBILITY
VitreaTeach offers users several different modes of use, each tailored to support a
specific type of task. The default mode includes persistent navigation that lets users
efficiently browse through cases, and a Case View screen that displays the full case
description and comments on a single page that can be consumed at a glance. This
mode works well for individual use, but another important task for residents is to
present a case to supervisors and colleagues in a collaborative learning setting.
Before VitreaTeach, users relied on PowerPoint for their presentations, cutting and
pasting images into a presentation file. The initial version of VitreaTeach maintained
this metaphor of creating a presentation, as Sadikali explains, "Initially, the
presentation was something you made, but one of the reasons for building this
application was that PowerPoint was too much work. Users didn't have time to copy
and paste images, they needed to be able to present right now. So we needed
something much quicker. That's why Present Mode was so important."

112 [email protected] Executive Summary

 The current version of VitreaTeach treats a presentation as a mode you enter, rather
than an object you create. From the list view screen you can click a button to enter
present mode. Present Mode displays a different format of the case information
which removes the application navigation controls and adjusts the content layout to
divide the case description into a series of screens with larger text, more suitable for
displaying to an audience. Users can configure the present mode to focus on the
most important elements.

The Present mode makes it easy to share cases with others in meetings. Each
segment of the case is presented on a separate screen, and text is enlarged for
easy reading.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 113

 Users can customize their presentations by selecting which elements to include

VitreaTeach also provides a Quiz Mode which temporarily hides the textual
description of a case to let users test their memory. Sadikali explains, "If you have a
case image and don't remember exactly what the diagnosis was, you can go into
Quiz Mode and the description and comments are hidden until you click to show
them."
The first version of VitreaTeach included specific areas for Presentations and Modules
in the main application navigation. Now, these areas have been entirely eliminated
and instead users can instantly switch any case into either Present Mode or Quiz
Mode. This shift helps keep the overall application experience lightweight and simple.

114 [email protected] Executive Summary

 Quiz Mode hides the case details until the user clicks to reveal them.

VitreaTeach also gives users the flexibility to share content with others in a variety of
ways. In addition to sending a message through the VitreaTeach system, users can
simply copy a URL and share it through any tool they choose, such as Twitter. This
option was important because visits with users showed that they already use
systems such as Twitter, instant messaging, and web conferencing to exchange
information. The VitreaTeach design team considered incorporating a live sharing
feature, but ultimately realized that the audience would most likely already have
established online collaboration channels, and it would be more effective to simply
make it easy to integrate VitreaTeach content into those existing channels by
providing URLs for each case.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 115

 Users can share cases saved in VitreaTeach either through the system, or by
copying a URL.

INFORMATION DISPLAY
Each screen of VitreaTeach is carefully optimized to make the most important
information easy find. The two main principles guiding the information display design
are simple, but overlooked by many application designers:

• Devote enough space to the most important content

• Don't show items that aren't necessary to the current task.

For example, the Case View screen displays one large medical image, thumbnails of
related images, and text describing the image contents. Initially, the majority of the
screen space was devoted to the primary image, but user research revealed that this
actually wasn't the ideal layout for users' tasks. Sadikali explains, "The image area
was originally larger, but through talking to users we realized that text is far more
important than we anticipated. Often they know the image but the meaning of it is
what the teaching system is about...in some cases the history and details are
actually just as important, or even more important than the text. That was part of
our learning, to realize that for this particular system, images are not necessarily the
focus."
Accordingly, the final Case View screen reprioritized the screen real estate, devoting
more space to the details and comments about the image.

116 [email protected] Executive Summary

 The VitreaTeach listing screen is another example of great information display—not
because of what it fits in, but because of what it leaves out. "We cleaned up a lot of
things...to emphasize the right information. When you're searching, the case
description and this history is really important, but everything else is not important.
The modality, date, and age columns are really critical, but people aren't really
interested in a whole lot of other information. We have a lot more information that
we could show, but it would just clutter the list," explains Sadikali.

This early version of the Case View screen devotes the majority of the screen to
images, leaving only a small area for the descriptions and annotations, which
wasn’t sufficient for users.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 117

 The final Case view screen layout reflects the importance of descriptions and
annotations for learning, and condenses the images into a smaller area.

118 [email protected] Executive Summary

 The case listing screen shows just a subset of the available information,
focusing in the key parameters: History, Tags, Modality, Date, and Age.

INTERFACE ELEMENTS
The VitreaTeach interface was built with both traditional GUI and touch device users
in mind, an important consideration because visits with radiology residents revealed
them to be extremely technology-savvy. One of the partner hospitals was even
considering providing iPads to all incoming residents, so tablets are a very realistic
platform for use of this application. VitreaTeach supports tablet users by providing
explicit, visible controls—rather than relying on context menus—for tasks such as
switching into Present Mode. The Case View screen also supports touch events, so
users can zoom in on an image using gestures.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 119

 The Present Mode and Quiz Mode buttons are easily accessible for users on
touch devices.

DESIGN PROCESS
Once users tried the first version of VitreaTeach, it became clear that it needed more
work. Vital Images has ongoing relationships with medical professionals who provide
user feedback on various products, and one influential user actually specifically
suggested involving a user experience designer to improve VitreaTeach.
While the initial feedback suggested that the visual design needed improvement, but
after a closer look at the structural issues and potential outcomes, Vital Images
realized the issues with VitreaTeach ran deeper than just the appearance, and
decided to pursue a full redesign. Sadikali convinced stakeholders of the value of a
full user-centered design process by using familiar terminology and concrete
examples, "I use visuals, like showing the wireframes and then the visual design that
came later, to express the difference between visual and interaction design. The
visual approach tends to work pretty well because people can see the difference. I
also used terms that people understand, like “workflow,” which has some cachet in
our industry.
Background Knowledge
Sadikali relied heavily on his past knowledge about radiologists (acquired over the
course of fifteen years working in field) to inform the design for VitreaTeach.
Because medical software can directly impact patient outcomes, its development in
the United States is supervised by the Food and Drug Administration (FDA.) Software
companies who build products to directly support the practice of medicine are
basically required by the FDA to do user centered design—that is, to have users
evaluate the products and ensure their effectiveness before the products can be sold.
Because of these requirements, Vital Images has ongoing relationships with several
teaching hospitals which provide product feedback. Design researchers regularly
interact with medical practitioners, as Sadikali explains, "We have a lot of
background user research that carries over. We actually film users in hospitals, and
have a lot of footage of radiologists. But we still do user research because things
change, and we need to understand what’s going on in the field, and what the new
procedures are."
The accumulated wealth of information helps create a rich picture of how radiologists
think, what their goals are, and how they are likely to behave. "We empathically
understand what they're trying to do...that's my role as the designer," explains
Sadikali. "We know they don't have to use the product and if it's not easy, they

120 [email protected] Executive Summary

 won't. We know our persona Alan and he doesn't tinker with his PACS 4, he tinkers
with his Porsche."
Targeted Research
To guide the design of VitreaTeach, Sadikali supplemented his general background
knowledge about radiology practice with specific research targeted at the learning
process for radiology residents. Activities included:

• Brief email surveys to assess the importance of competing goals and

design elements
• Telephone interviews to gather specific details about how residents study
and learn
• On-site observation of residents practicing and studying.
This additional work was crucial to the design of VitreaTeach because it led to some
surprising insights, such as the fact that most learning was informal and self-
directed, rather than official assignments created by instructors.
This target research also revealed that managing content was the most important
user need for VitreaTeach to address, even more than viewing the content. Free
competitor products support the viewing of images but the ability to easily save,
find, and share them is an area where VitreaTeach can provide unique value to
users.
Conversations with senior practitioners revealed key information about the business
model as well, as Sadikali explains, "We were thinking everyone would want to
share, but the reality is institutions aren't necessarily interested in cases from other
institutions. And some sites have told us that right now, they don't want to publish to
the public web, in part because they believe it is a lot of work to publish to the web,
and valuable teaching cases are seen as a proprietary work that serves the
institution."
"This design process reinforced the importance of observational research to find out
how people really behave. We broke through the communication that users typically
have which is, 'You need to include this feature and that feature,’ and got them back
to, 'What are you really trying to do, what is radiology?' I started by asking these
basic questions to work up to the more complex questions, to get to their mental
models. Even if you have been given a very explicit set of requirements, you don't
necessarily know what problem you really need to solve. They may not be the real
needs, or it may be a mapping of the real needs, but it's the wrong mapping, so you
need to do something different from what the requirement says to achieve the right
result," says Sadikali.

Expert Review
It's essential to find out whether users can understand and use an interface, but
difficult to get busy medical professionals to come into the lab for a usability test.
Sadikali addressed this by conducting “review” sessions with practitioners to get
feedback about his prototypes.
Via Webex online meetings, Sadikali showed practitioners screens built in Fireworks
and linked together. These sessions weren't traditional usability tests, but aimed to

4
The Picture Archiving and Communications System—PACS—are the basic systems
radiologists use for their core tasks.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 121

 evaluate users' understanding rather than simply explain the product, "In the expert
review, I don’t just do a demonstration. I’ll ask things like, 'If I clicked here what do
you think would happen?' or, ‘How do you think this would work?’ before explaining
the functionality. Sometimes I’ll show people a flow and ask them to comment on it,
people can comment on a workflow quite well. We can do a review like this before
we actually have working software, and at a stage where we can make larger
changes easily," says Sadikali.

Implementation
The design and development of VitreaTeach was accomplished in about nine months.
For the first six months it was a solo project, with one person working on both design
and development. A designer and team of developers took over for the redesign
during the last three months of the project.

PROJECT TIMELINE
2010 Investigation by marketing

Nov 2010 Version 0.1 pitched to focus group at RSNA, a large

radiology conference

March 2011 Phase 1: Alpha release to three teaching hospitals

April 2011 Feedback from some users that we need to work on

design, this filtered through organization as a “visual
design problem” or “graphics problem”

May 2011 Design team proposes a full redesign (rather than

visual tweaks)

June 2011 Directors approve redesign project

Early June 2011 User research with two sites, one radiologist, and
Chief Resident Radiologist

Late June 2011 Ideation and conceptualization (two weeks)

June 22, 2011 Expert review via WebEx using Fireworks mockups to
step through screens; subsequent reviews via email

July 7, 2011 Review work with original developer

August 2011 Phase 2:release to one teaching hospital

Although the redesign changed the front end, the developers were able to re-use
most of the back-end code written for the first version. VitreaTeach also makes use
of layouts developed for VitreaView, a related product; for example, the VitreaTeach
Case View layout is simply a modified version of the same screen in VitreaView.
Since the VitreaView design was already optimized for lower resolutions such as
1024x768, these layouts work well on the iPad. Sadikali explains, “The application is
web-based so you can access it on a tablet through the browser, and we also adjust
it specifically for different platforms, like Android. We’re planning some tweaks to
increase the button size and row sizes. Some of them were already large; but some
areas need improvement. We’re working now on refining it for tablet use.”

122 [email protected] Executive Summary

 LESSONS LEARNED
The process of creating VitreaTeach reinforced an important principle of user-
centered design: There is no substitute for working directly with users. Even if you
know a lot about the users in other contexts, if the activity type is new (such as
learning rather than practicing medicine,) it’s easy to make incorrect assumptions
about what’s important and how people behave. Sadikali says, “Start from the
beginning, and don’t try to skip the design process. It’s something you need to do
every time.”

TEAM
VitreaTeach was initially conceived and developed by Thomas Myers, a developer
working out of the marketing department. Navid Sadikali, a user experience
designer, led the redesign project.
Navid Sadikali is a user experience designer at Vital Images. He has an
undergraduate degree from Brown University and a Masters in Computer Science
from the University of Waterloo. His focus is on interaction design and he manages
visual designers to cohesively bring the entire user experience together.
”What motivates me is helping analytical-thinking businesses (medical, bio-tech)
grow using design. I believe leapfrogging the competition requires involving
professional designers in product strategy rather than having design be a
downstream activity you do once everything (markets, customers, requirements) is
figured out.”
Thomas Myers (Senior Manager, Strategic Technical Marketing, Member of the Office
of the CTO—Vital) is a technology entrepreneur who is passionate about building
products and services that make a lasting difference in the lives of patients.
“VitreaTeach will produce better trained radiologists who will be empowered to
deliver better care to us all and I am very proud to have been a part of the team
incubating it.”

Navid Sadikali

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 123

 Wakelicious

Organization: Utopian Army Design team:

Application: Alarm clock for friends to wake • Mike Holenderski, Developer
each other with YouTube videos • Rik Runge, Designer

Headquarters: Eindhoven, The Netherlands

www.wakelicious.com

OVERVIEW
Main function: Wakelicious is an alarm clock that wakes you up by playing YouTube
videos sent by your friends.
Why it’s unique: Wakelicious is a great example of lightweight design because it
makes a complicated process feel simple and intuitive. Instead of long explanations,
the application uses structure and action labels to help users understand how it
works. It makes hearing a wake-up alarm more fun by involving friends in the
experience.
The organization: Wakelicious is the first product released by Utopian Army, a
startup software company.

WHO ARE THE USERS?

Wakelicious is targeted at people who already use Facebook to interact with friends
and family, and are interested in sharing music with friends and discovering new
music. Although it was developed in the Netherlands, most Wakelicious users are in
the United States. The application is currently only available for iOS mobile devices,
and is used by more than 1,000 people.
THE APPLICATION
Wakelicious makes the moment of waking up a shared experience with a friend;
instead of a generic alarm sound, users start the day with a message or song sent to
them by a friend. As one Wakelicious user put it, "Using Wakelicious is like receiving
a present every day."
Users can set their Wakelicious alarm clock and when that time arrives, Wakelicious
automatically plays the first video in their wakelist, or video queue. Users can also
send videos to their friends' alarms, view their history of sent and received videos,
and share their favorite videos on Facebook.

Wakelicious in a nutshell: a social alarm clock.

124 [email protected] Executive Summary

 The queue indicator shows whether a user has any videos waiting, but Wakelicious is
designed to keep the experience of waking up a pleasant surprise by keeping the
titles and senders of videos hidden until the alarm goes off. (If no one has sent you
any videos, Wakelicious will play a video from a list built by the designers.)
Rather than trying to build an entire social network and content repository from
scratch, Wakelicious leverages two well-established entities, Facebook and YouTube.
When users start Wakelicious, they are prompted to link the application to their
Facebook account. From that point, the content from both Facebook and YouTube is
tightly integrated into the Wakelicious experience. From within the app, you can
browse your Facebook friends list and choose someone you'd like to send a video to.
Once you pick a friend, you are taken to a screen within Wakelicious where you can
search for a YouTube video to send.
Because Wakelicious integrates with these well-established sites, the learning
threshold for users to begin using Wakelicious is much lower. Their friends and a
huge number of videos to choose from are easy to access from right within the
application.
Integrating with Facebook also offers the potential for marketing the application to
potential new users by posting to existing users' Facebook pages, but Wakelicious
takes a cautious approach to this goal. An early prototype was set up to post a
comment to the user's Facebook wall each time Wakelicious played a video, but this
behavior alienated users and was quickly removed. Mike Holenderski, who developed
Wakelicious, explains, "We don't want to spam our users. We wanted to make an app
that we would personally want to use. We don't like those apps that spam your wall
or your email account, so we didn't want to do this to our users either. It's about
long term vs. short term benefits. Maybe from the marketing perspective it would be
better to be more aggressive, but it's better in the long run if the application doesn't
spam users."

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 125

 Wakelicious plays videos sent by your friends to wake you up.

126 [email protected] Executive Summary

 Users can choose videos from YouTube and send them to their Facebook
friends.

SIMPLICITY
One of the biggest hurdles for Wakelicious is simply getting people to understand
what the application is for and how it works. Holenderski and Runge took a few very
specific steps to simplify the Wakelicious experience. In fact, after spending three
months to build a working prototype, they spend the next nine months refining and
simplifying the design to make it easy to use.

Matched users' mental models

In order to minimize the development cost, the first version of the Wakelicious
mobile application could only receive videos, not send them. Users would receive
videos on their phone, but would only be able to send videos via a separate
Facebook application. Rik Runge, who led the design for Wakelicious, explains,
"When we started, it seemed like it would be technically difficult to build a mobile

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 127

 app that could send videos, therefore we adapted the Wakelicious concept to
eliminate sending from the mobile app. But later on we realized that was where we
got it wrong. For most users that’s not a logical way to use an app. People expect to
have two-way communication through the same app."
Holenderski adds, "This is a perfect example of where we started programming too
early, and the development influenced the design. Because it was technically
difficult, we took the feature out. But this was a mistake because the programming is
just there to support a great design: It should not be done the other way around. It’s
important to have a great design first."
After doing usability tests of the working prototype with the split functionality, it was
obvious that Wakelicious needed to be able to both receive and send videos from
within the same interface. Adding the send feature made the application much easier
for users to understand and engage with.

Top: The first iteration mobile app could only receive videos. Users had to go to
Facebook in order to send a video. Bottom: Now, users can both send and
receive videos from the iPhone app.

Less reliance on instructions

Early versions of Wakelicious relied on seven screens worth of copious instructions to
explain the application, but user testing revealed that people simply skipped over the
instructions, and just wanted to explore the application and start using it
immediately. Runge says, "When we made it we thought that we just needed a lot of
explanation to convey the concept. But we learned that people can't get the concept

128 [email protected] Executive Summary

 via text. They understand it by seeing the buttons and what they lead to. That helps
them build their mental model. They won't read instructions."

Because the prominent instructions weren't actually helping users, they were shifted
into an auxiliary role: still accessible for users who seek out the help icon, but no
longer a main navigation option.

Instead of instructions, the design now focuses on making sure the structure and
controls of the application are self-explanatory. Some contextual help is also built
into the interface; for example, a brief description of the backup alarm appears just
below the control for that feature on the Settings screen. An instructional screen also
appears once an alarm is set, to remind users to leave the application open.

Left: An explanation of the backup alarm feature is built into the Settings
screen. Right: Once an alarm is set, instructions for leaving the application
open appear.

Restructured navigation to prioritize key tasks

Making the interface self-explanatory required a big shift in how the functionality was
organized, presented, and labeled. One of the biggest changes was removing the
persistent menu bar at the bottom of the screen. Holenderski explains, "Initially we
had tabs on the bottom, but that's actually better for a different kind of application,
where you quickly switch between parts of the interface, independent of what you
were just doing."
Because the core Wakelicious activities (setting an alarm and sending a video) are a
linear series of steps, it's actually better for users to move sequentially through the

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 129

 necessary steps. Keeping the options for secondary tasks, such as adjusting the
settings, always visible just creates a distraction and more opportunities for users to
get confused and go down the wrong path.
Removing the persistent navigation bar helped de-emphasize secondary tasks, such
as adjusting the settings, accessing help content, and viewing the history. After
removing those distractors, Runge focused on making the remaining core tasks
obvious and understandable, "I came up with a two-button interface, instead of a tab
bar and a button. This made the app really focus on two key features. The two
buttons at the bottom are big and simple."
The naming of the two main task buttons on the home screen buttons was critical to
instantly conveying how Wakelicious works, "We talked a lot about the labels for the
buttons; when you push the bottom one, you’re not actually waking a friend. But if
you wrote select a video for a friend that would be less active. The two buttons on
this main screen help people understand what the app is about. It lets people know
that they can set an alarm and wake a friend. And they can start to understand the
application.”
Once these changes were in place, users were able to pick up the application and
start using it without reading the instructions. "The two main tasks, waking a friend
and setting an alarm, are more important than the history or settings. So those
buttons are smaller, the text is smaller. Seeing the time is important, so that's big.
So getting the size of elements right helps people understand the application. When
we added this step, people started to get it, and then the drill down interface guides
them through the sending process," says Runge.

130 [email protected] Executive Summary

 Left: Early versions of Wakelicious included a persistent navigation menu at the
bottom of the screen. Right: now, the home screen has just two simple calls to
action.

Wakelicious guides the user through the Wake a friend task with a linear
pathway which presents each step in sequence. Left: First, users see their list
of friends. Right: After choosing a friend, the user is taken to a screen search
for videos.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 131

 Left: After selecting a video, users have the option to play to preview the video
and make sure it’s the right one. Right: When watching the video preview, the
Done button returns you to the previous screen.

132 [email protected] Executive Summary

 Left: Once you tap Send Video, you see a confirmation message. Right: An
animated graphic is displayed while the video is being sent.

Once the send process is complete, the user is returned to the home screen and
a Sent confirmation is displayed.

Removing the navigation bar to simplify the task options was a vast improvement,
but there is such a thing as going too far with simplicity. Due to technical
constraints, Wakelicious at first did not have a snooze function. This turned out to be
another instance where user needs should be prioritized over development
constraints, however, because people really wanted to snooze. Wakelicious now
includes a snooze feature: when the alarm goes off, a video begins playing
automatically. If you tap I'm awake the video continues playing through to the end,
but if you tap Snooze the application returns to the clock screen and then launches
the video again nine minutes later.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 133

 The snooze feature will cancel the video, and then re-launch it 9 minutes later.

INTERFACE ELEMENTS
Wakelicious uses simple and familiar UI elements, and follows Apple’s Human
Interface Guidelines. Large buttons and fonts make it easy for users to navigate
between screens, and the dark background color is more comfortable for night
viewing. Runge explains, "We've tried different visual styles...We wanted something
remarkable, but being remarkable often gets in the way. We ended up with an
interface that doesn't get in the way. It's there, but it's not the main attraction of the
application."
In the History view, buttons at the bottom of the screen let you toggle between
Received and Sent videos. When an interface has buttons to switch between two
states, it is sometimes difficult to tell which state is currently active, because the
visual treatments of the two buttons don’t clearly look raised or depressed. The
shadow visual treatment on the Wakelicious Received and Sent buttons, however is
dark enough that the currently selected button does actually look depressed.

134 [email protected] Executive Summary

 The visual treatment of the Received and Sent buttons is strong enough to
clearly indicate which is currently depressed.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 135

 Left: Standard interface controls such as those for setting the alarm time are
familiar to iOS users. Right: The Slide to cancel control for canceling the alarm
resembles the standard iOS unlocking control.

DESIGN PROCESS
When they started the project, Holenderski estimated it would take about two weeks
to build Wakelicious. But after quickly putting together a working prototype and
testing it with users, the team realized that making this unusual functionality
understandable and easy to use would require a much more extensive investment in
design. "In order to save time for the customers, we had to invest this huge amount
of time. This is our third or fourth iteration of the design, because our initial designs
were just too complicated. We now have much more appreciation for simple
applications. Just because it looks simple doesn't mean that there's not a lot of work
behind it. Usually the designers put a lot into making it simple," says Holenderski.
The first version had actually been built in code before user testing revealed the
deep conceptual and structural problems. Though it was painful, Runge and
Holenderski went back to square one and started over with paper prototyping.
"With the first version, we had a working prototype very early. This was actually a
mistake because we hadn’t worked out the details of the interface yet, so we ended
up needing to do a lot of redesign. We started programming much too early. From
then on, we did more iterations of the interface in paper prototypes and wireframes
before programming," says Holenderski. Their subsequent prototyping efforts
focused on the processes unique to Wakelicious, such as sending and receiving
videos, to make sure these were understandable.

136 [email protected] Executive Summary

 Usability Testing
Usability testing was an invaluable part of the Wakelicious design process. "After a
while you get used to the flaws in the design. You lose the general overview and
tend to focus on details. It takes another set of eyes to pull you out. At this stage
you will get most insights by showing the work to other people through user testing,"
explains Runge.
The Wakelicious interface was tested with users at several different stages of
development, including
• interactive prototype (built in code)
• paper prototypes
• keynote presentations.
Because Wakelicious is intended for a broad audience—anyone with a mobile phone,
a Facebook account, and an interest in music—finding participants for usability
studies wasn't difficult, as Runge explains, "We tested it on our family and friends
and also with people who are really into technology. One guy worked at Google, and
we thought, ‘He'll understand it.’ But he didn't. At first he totally dismissed the
manual and started playing around. At that point we realized there was something
wrong with the way we designed it."
Simple, informal tests with friends and family interacting with paper prototypes
yielded important design guidance, "We learned that users dismiss anything that
looks like a lot of text, and go straight to playing with the buttons and interface.
Once we saw that, we realized that the application had to be easy enough for people
to get the idea gradually," says Runge.

To refine the user interface, designers conducted usability tests on paper

prototypes with friends and family.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 137

 Implementation
Wakelicious was designed using Adobe Illustrator to create visual designs, and Axure
and Keynote to build interactive prototypes. The actual application is a native iOS
application built with the iOS SDK. The native application platform was necessary to
minimize load times and include functionality that simply wasn't accessible through a
web platform. In order to make the application function quickly and simply for users,
the backend is set up to perform some actions concurrently, such as loading video
titles and thumbnail images.

iOS Constraints
The iOS provides a large potential audience, accessible documentation, and built-in
distribution method. But this platform created some complications and functionality
constraints. In order to automatically play a video at a certain time, Wakelicious
must be open and running in the foreground of the device. Although iOS does allow
applications to stay open in the background, the actions an app can take while
running in the background are limited to a few basic things like continuing to play a
song. They can't play a YouTube video, because background applications can't
connect to the internet. Only Apple’s own alarm clock built into the iOS system has
the ability to launch at a specified time even if it wasn't open in the foreground.
A few different features are built into Wakelicious to work around this constraint.
First, users are advised when they set an alarm that they need to leave the
application open. In order to conserve battery life overnight, Wakelicious also
instructs users to turn the device face down, and uses the proximity sensor to turn
off the screen.
Since waking up at the right time is potentially a really important task for users,
Wakelicious also includes a backup alarm, in case the application is accidentally
closed or the Internet connection is lost. If Wakelicious isn't open and running at the
time the alarm is set, the backup alarm generates a local notification and plays a
backup alarm sound. When the user acknowledges that notification, it triggers
Wakelicious to open and automatically start playing a video.

Facebook and YouTube Constraints

Integrating with Facebook and YouTube provided significant advantages in terms of
access to content, but also introduced difficulties and constraints.
When relying on external services, such as Facebook and YouTube, one becomes
dependent on changes to their APIs. For example, about two months after launch
Wakelicious experienced downtime due to a change in Facebook’s authentication API.
Holenderski has therefore subscribed to alerts about changes to the relevant APIs to
keep up-to-date. In addition, Wakelicious backend servers continuously monitor the
health of the system and generate alerts when something goes wrong.
The YouTube API is more stable, but because of the video format provided by the
API, Wakelicious cannot download a video in advance and then play it later. For now,
Wakelicious relies on the standard YouTube video format and advises users to make
sure their wireless or 3G connection is active.
YouTube also sometimes removes videos due to copyright violations. Wakelicious is
set up to automatically confirm that a video you've received is available, and if not,
remove it from your queue and play the next available video.

138 [email protected] Executive Summary

 PROJECT TIMELINE

September 2010 V0.1: Interface design

October 2010 V0.1: Working prototype

November 2010 V0.1: User testing (with working prototype)

January 2011 V0.2: Interface design (simple black, drill down, no

manual)

February 2011 V0.2: Wireframe prototype

V0.2: Working prototype

March 2011 V0.2: Paper prototype

V0.2: User testing (with paper prototype)

April 2011 V0.3: Keynote prototype

V0.3: Paper prototype

V0.3: User testing (with paper prototype)

May 2011 V0.3: Working prototype

V0.3: Beta testing (with working prototype)

June 2011 V1.0: Submitted to AppStore

June-August 2011 V1.1-3: Bug fixes

October 2011 V1.4: Added snoozing

LESSONS LEARNED
The Wakelicious team shared a few lessons learned from this design effort:
• Testing with users is essential. "You may think you have a good idea
about what will be easy to use, but you need to get your idea out, prototype
it and put it to the test. If the prototype fails to do what you have hoped for,
i.e. when you see users dismiss your whole interface or things you think are
important, you will learn a ton about how to improve the user experience. "
• Fail early. "The sooner you gain insights into how people perceive your
interface, the less effort it will cost to fix the design. Failing seems bad, but
those insights are so valuable. Working with paper prototypes allows you to
solve problems much more efficiently than during development.”

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 139

 • People don't read instructions. "People do not care for manuals—the
design must be self-explanatory. Our initial design required a lot of
explanation and we expected people to read the instructions; it just didn't
work."
• Consider building for multiple platforms. "If we had decided earlier that
we wanted to do it for multiple platforms, we would have used something like
PhoneGap, or created our own framework that would allow us to port the app
much more easily to another operating system. Those decisions have to be
made early in the implementation."
• Design first, code later. “Anxious to get started swiftly, we made some
decisions from a developer’s perspective. It was only until later, during user
testing, we saw how these decisions had influenced the user experience. So
take note that programming too early is likely to compromise your design
decisions.”

TEAM
Wakelicious was built by Utopian Army, which is a startup created by Rik Runge and
Mike Holenderski. Rik developed the concept, crafted the design and evaluated the
interface to ensure the app's ease of use. Rik also works as an interaction designer
at Mirabeau, a full service internet company in The Netherlands, focusing on mobile
development and user testing. Mike is responsible for the technical part of
Wakelicious, from app development to server maintenance. He is also employed at
the Eindhoven University of Technology, where he is currently finishing his PhD in
computer science in the domain of real-time systems.

Rik Runge (left) and Mike Holenderski (right) make up Utopian Army, creators
of Wakelicious.

140 [email protected] Executive Summary

 Workflow Applications

Sharing and approvals are necessary and overall improve content and keep
processes moving. But tracking and maintaining the integrity of content, and
communicating status throughout the editing or status-changing cycles has
presented a difficult design challenge for many years. Who has the baton now? What
does he need to do? What is he allowed to do? Where does it go next? Who should
be involved, and when? What if that person is not available? The question designers
must consider and answer seem to be endless, yet the great designers of the
applications in this section ferreted out all of the answers swimmingly.
Workflows may add levels of politics, security, and intricacy depending on who is
sharing the content, such as sharing between an organization’s workers, an
organization and its end users or customers, or B2B. Additionally, if there are
different user types in varying job roles, the interface needs to present different
views tailored to each context. And if diverse organizations or individuals actually
own the content, or nobody actually owns the content, these can bring in even more
complexities.
Designers working on these applications must thoroughly examine the workflow
structures for the tasks, users, and content in order to build great applications.
Some UI elements that may behoove people using this type of application include:
• Process steps (sometimes numbered) and progress indications to
define stages in the process
• Ability to comment and otherwise deviate from the process flow when
absolutely necessary, in the name of getting the tasks done and not
being beholden to a structured UI
• Wizards and linear information paths
• Pop-ups and tips within and throughout the process
• Previews
• Clear display of content properties such as creation time, person,
team; and what has been done thus far.
Winning applications in this category include:
• Climate Action Planning Tool, helps research campuses and
universities reduce carbon emissions.
• eReview, an intranet application that integrates complex business
functions into the organization’s intranet.
• Eventbrite, a self-service software tool used by event hosts to
organize, sell and manage tickets for events. The Create Event
workflow described in this report is the process used to set up a new
event within the Eventbrite system.
• Hobsons CRM product, product suggestion portal tool for the
educational institutions Hobsons serves: gives its customers an
opportunity to make comments/suggestions.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 141

 Climate Action Planning Tool (NREL)

Organization: National Renewable Energy Design team:

Laboratory • Ruby Nahan, Communications
Application: Carbon reduction calculator Representative/Project Manager
• Rachel Sullivan, Project Manager/Writer
Headquarters: Golden, Colorado
• Laurence Hoess, Developer
www.nrel.gov • Erica Augustine, Designer
• Shauna Fjeld, Usability Consultant
• Kristin Tomley, Web Content Producer

OVERVIEW
Main function: The Climate Action Planning Tool was built to help research
campuses and universities reduce carbon emissions. The application helps users
create realistic plans for reaching specific carbon reduction goals, by assessing the
total current emissions and calculating the impact of different possible reduction
strategies.
Why it’s unique: The Climate Action Planning Tool was chosen as an example of a
great design because it
• Helps users through a difficult task by breaking it up into steps that make
sense
• Uses interactive visualization to communicate complex information in an
understandable way.
The organization: The U.S. Department of Energy’s National Renewable Energy
Laboratory (NREL) is the only national laboratory solely dedicated to advancing
renewable energy and energy efficiency technologies from concept to commercial
application.
NREL develops renewable energy and energy efficiency technologies and practices,
advances related science and engineering, and transfers knowledge and innovations
to address the nation's energy and environmental goals.
Design philosophy: The design team had two main goals for the Climate Action
Planning Tool: they wanted the final design to be interactive, but simple enough to
be used without instructions.

WHO ARE THE USERS?

Management and facilities directors of research campuses who are tasked with
developing a climate action plan to reduce greenhouse gas emissions are the primary
target audience for this application. A typical usage scenario would be, “The user’s
boss says, ‘We need a climate action plan, figure it out.’ The user may not know
where to start in terms of taking a big picture approach to energy use across the
campus. The tool helps them figure out, if we want to get to zero greenhouse gas
emissions in 40 years, these are the areas we need to address. We can’t just look at
replacing light bulbs, we have to look at all of these other areas such as
transportation and building plans,” says Rachel Sullivan, the team’s project manager
and writer.

142 [email protected] Executive Summary

 This application is targeted specifically at facilities directors who manage campuses—
which could mean a university or college campus, but could also be a government or
private research facility. Essentially the tool is useful for anyone interested in
reducing the carbon footprint of a piece of property that is:
• Composed of multiple structures
• Owned or managed by a single organization
• Used by large numbers of people.

The target audience is technically oriented, but the team’s conversations with users
revealed a varying degree of knowledge about energy use and goals for greenhouse
gas emissions. Sullivan explains, “There was one person who knew exactly what the
goals were; another person was new in the position and really had no clue yet.”
Regardless of how long they’d been on the job, much of the information needed to
use the tool is so detailed that users simply won’t know it off hand. One of their
biggest needs is guidance about how to start the planning process.
As a federal government organization, NREL also complies with Section 508
accessibility requirements, and considers people with disabilities or those using
assistive devices as an important audience.
Potential users discover this application by visiting websites such as the National
Renewable Energy Laboratory, U.S. Department of Energy Federal Energy
Management Program, or Labs21 websites, or through outreach efforts such as
presentations by NREL technical leads and project leaders at industry conferences.
THE APPLICATION
A project nearly two years in the making, the application was originally an Excel
spreadsheet created by an industry expert at Cornell University. The NREL
Communications team was tasked with transforming that spreadsheet into an
interactive tool. The result is a web-based application that:
• Manages users’ expectations up-front with a downloadable form listing the
data points users need to collect in order to use the application.
• Allows users to enter their data via an accordion menu that automatically
calculates data.
• Produces a dynamically generated, printable graph that gives users the
overall picture of energy use on their campus.
• Quickly gives the user an overview of what their best options are for
developing a climate action plan by arranging the data so that the energy
sectors are in order of those that will have the most impact on reducing
emissions.
• Provides a dynamically generated summary report of all the data, graphs,
and calculation results that can be used to inform climate action plans.
One of the most important things this design team did was minimize explanatory
text, and instead let the functionality speak for itself. Sullivan explains, “Early on in
the process we had an instruction sheet … there was way too much content. It's
easier to write more content than it is to write a small amount … but people just
weren’t interested in reading instructions; they just wanted to put the data in and
see what happens.”

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 143

 The Climate Action Planning Tool provides a structure for calculating
greenhouse gas emissions and how to reduce them.

TASK FLOW
After testing the prototype with users, the team introduced several modifications
that dramatically improved the workflow, including the following:
• Revised the overview to articulate the specific steps
• Created a separate step for collecting data offline
• Numbered the steps and placed a progress indicator on each screen.

144 [email protected] Executive Summary

 The Climate Action Planning Tool homepage provides a basic overview of the
application, a call to action to launch the workflow, and a few links to related
information. Overall this page is not complex, but during usability testing the team
was able to identify some weak areas of the page and introduce improvements.
For example, while the first version of the home page included four bullet points that
basically describe the four steps of the tool, the descriptions were rather vague and
wordy, and the bullets actually make the copy look like a feature list rather than a
sequential set of instructions.
In the updated home page, the steps are numbered one through four to indicate the
linear sequence and the description of each step is shorter and describes very
specifically the exact action the user will be taking: for example,
• Adjust specific reduction goals by technology option to see how they
impact your overall goal
became:
3. Set and adjust emissions and technology options goals.

The updated page also moved the links to secondary information (such as Help and
details about the calculation formulas) from the left navigation menu to a footer
element at the bottom of the page. This is a vast improvement, because showing
related links in a left navigation bar violates user expectations: users who do look to
the left will be expecting to find global navigation options such as the main steps in
the application; users who are interested in the secondary related links will probably
look to the footer, as that is where such items are more commonly placed.
Adding icons to these links calls them out as meaningful content, but moving them to
the bottom and enclosing them with a box makes it clear that they are secondary to
the main focus of the page: launching the application.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 145

 The first version of the home screen was less specific about the steps and
included a left-side navigation menu.

The final home screen includes four brief numbered steps, a single strong call to
action, and a footer element for related links.

146 [email protected] Executive Summary

 In order to use the application, users must first gather specific information about
current energy usage at their facilities. These are the data used to calculate which
technology options will lead to the most significant reductions in consumption of
fossil fuels for each individual campus, and will therefore meet greenhouse gas
reduction goals.
User testing showed that gathering this information can be a fairly involved process,
“Often they didn’t know much about the different energy sources used by the
campus, so they’d have to go to other people and collect different sets of data. It
was crucial for them to go out and seek help from other people on the campus that
had more expertise in these different areas. That’s why we came up with a
downloadable worksheet, to help you gather the information,” explains Laurence
Hoess, the developer.

The worksheet was optimized to fit the way users perform the task and make it easy
to transfer the data into the application, as Erica Augustine, the designer on the
project, explains, “We did it in a PDF so that they can print it, and it looks exactly
like the input screen so you know it will translate directly. Also, there are some
people that want to just play around with the tool, so they can skip this step.”
Because collecting this information is so essential to getting value out of the
application, the entire Step 1—the first screen you see what you launch the
application—is devoted to the link to download the worksheet. Paradoxically, the
application is much improved by adding this step to the process, even though the
“step” is actually a wholly off-line activity, and the only thing for users to actually do
on this page is print out the PDF form.
In the first version of the design, the initial screen was called Before you start; it
described the need to gather information but it appeared to be instructional text and
many people skipped over it. Access to the form for collecting information was a text
link within a paragraph. This confused users, as Shauna Fjeld, the usability
consultant on the project, explains, “People just wanted to dive in and use the
application, but it doesn’t work that way; you need to collect data first. So we
changed the design to make it very apparent that you have to gather data before
you can use the tool.”
As soon as you start the application, a progress indicator appears showing which
step you are currently on, and how many steps remain. This indicator appears on all
subsequent pages, and is especially helpful for this type of process where you expect
people to move backward and forward through the steps, checking their work and
making changes.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 147

 Step 1 introduces a progress indicator and prompts the user to stop and gather
the energy use information specific to their institution that they will need to use
the application.

The first version of the first screen appeared to be just instructional text. Users
thought they could just skip it.

148 [email protected] Executive Summary

 After collecting all of this information, the user can input it into the form provided in
Step 2: Enter Baseline Energy Consumption Data. This form automatically calculates
the total carbon output from both indirect and direct emissions, based on subtotals
provided by the user.
During development, Hoess built a function to populate the application with sample
data in order to make it easier to test the application without having to make up a
whole data set repeatedly. Once the application was complete, he removed this
function but users asked to have it added back in. It turned out that people wanted
to try out the application and see how it worked before they invested time in
collecting all the up-front information, and found the pre-populate feature helpful for
‘test driving’ the application to get a sense of how it works and what they would be
able to do with it.
Step 2 also includes a bit more explanation for the last two input items (Growth
Rates) since testing revealed some confusion about that terminology. Now, people
can click on the What’s this? link next to those items to see contextual help
information.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 149

 The data input form in Step 2 looks almost exactly like the data collection
worksheet, with the addition of Total fields (which are automatically calculated.)

What’s this? link opens a pop-up window with contextual help.

FLEXIBILITY
The first few steps of the Climate Action Planning Tool are very straightforward, but
when users arrive at Step 3 they are confronted with a complex set of controls and
choices.
At this point in the process, users have entered the existing carbon emissions at
their facilities, and it’s now time to start figuring out how they can reduce the
amount of greenhouse gas emissions.
The calculations represented on this screen are complex, but several significant
design elements have been built in to simplify the process and give users the
flexibility to control the calculations.

150 [email protected] Executive Summary

 First, the nineteen different inputs are divided into six major categories, grouped by
topic such as Transportation and Renewable Electricity. This categorization allows the
interface to present the user with an organized list of understandable titles, rather
than an overwhelmingly long page. The sequence of categories is also intended to
subtly guide users towards more realistic and effective strategies for reducing carbon
emissions; goals related to energy efficiency such as optimizing the use of existing
space, conserving energy, and transportation come first in order to encourage users
to consider these as their primary tactics, before considering options that can be
more expensive (such as converting to renewable energy) or less direct (such as
carbon offsets.)
The categories are presented in an accordion with the first category expanded by
default — an absolute requirement with accordion or collapsible interfaces, since
people easily overlook such elements if they are all collapsed.
After making selections in the first category, a Next button allows the user to close
that category and automatically opens the next group.
This provides an easy and straightforward path for walking through the data entry
points. But the interface also makes allowances for users who want more freedom to
move back and forth between categories, or even more than one category at a time.
The colored title bar of each category acts as a button to collapse or expand that
module, and the small triangle on the far left rotates to indicate whether the module
is currently collapsed or expanded. An especially nice feature of the design is that
the entire colored title bar is clickable, rather than just the triangle or just the label.
This provides a much larger target area for the mouse, making it easier for users to
click on the control.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 151

 Step 3 lets users input their specific reduction goals; different types of
reductions are divided into separate categories such as Transportation and
Renewable Electricity.

152 [email protected] Executive Summary

 Users can open more than one module at a time; the entire colored bar is a
clickable button to toggle the state between open and closed, and the triangle
next to each module title rotates to indicate the current state.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 153

 INFORMATION DISPLAY
After entering the reduction goals and clicking the Calculate button at the bottom of
the screen, the graph in the upper right corner displays a visualization of the total
amount of greenhouse gas emission reductions over time.
The graph also divides up the total greenhouse gas emissions by category, to show
how much of the contribution comes from the different components of the reduction
strategy—extremely helpful information for policymakers who may be balancing
budgets and trying to decide what to spend their money on.
The colors were deliberately chosen to correspond with the colors used on the main
NREL website; in addition to color coding the graph categories, the legend is placed
right next to the graph in order provide a vertical label listing, which corresponds to
the order of elements on the graph — rather than relying exclusively on color to
distinguish which category is which.
Since users are likely to want to experiment with different reduction goals and
strategies, this screen retains all the inputs for selecting goals. The first version of
the application actually placed the graph at the bottom of the screen, rather than the
top; this presented huge discovery issues for users, who never noticed the graph
preview at the bottom of the screen. Hoess explains, “The first layout had the graph
at the bottom of the page and all the inputs at the top; we found that users weren’t
seeing that as they made changes they were reflected in the chart below. So we had
to redesign the layout to help make that connection between the inputs and the
graphic.”

Once the targets have been entered and calculated, the graph displays a
visualization of how much carbon emissions will be reduced over time.

154 [email protected] Executive Summary

 The first version of the application placed the graph at the bottom of the screen,
where it was easily overlooked.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 155

 After adjusting the goals to their liking, users can create a formal summary of the
plan. This step is important because the users are not actually the final consumers of
the application output, “Users also need a way to report back the information to their
management teams, a way to take the information that’s gathered in the tool and
incorporate it into their climate action plan or a presentation,” explains Sullivan.
The final summary report includes a recap of all the information entered into the
tool, the graphs, and a list of the numerical outputs, and links to additional
information about climate action plans on the NREL.gov website. In addition to a
print function, there are options to download each of the graph images for use in
reports or presentations.

The Generate Summary Report button appears after the first calculation graph
has been generated.

156 [email protected] Executive Summary

 The final report can be printed as a PDF document, and individual graphs can
also be printed our downloaded.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 157

 The bottom portion of the final report output, which includes the results in text
format to ensure accessibility.

DESIGN PROCESS
The application was built by members of the NREL communications team, which
operates at the direction of internal clients: technical experts who are also employed
by NREL. The team took a user-centered design approach to developing this
application, which included:

• User profiles
• Task analysis
• Sketching
• Usability testing of prototypes.
The project was kicked off with a brainstorming meeting that brought together
design, development, writing, usability experts as well as the internal client to
provide direct technical information. The ideas generated at this session provided a
great starting point for the project, but as the team discovered there is no substitute
for talking to actual users. Sullivan explains: “Typically on other projects we do
usability testing early on, but in this project we had a deadline to get a beta version

158 [email protected] Executive Summary

 ready for a conference that was happening in three months, so we didn’t have time
for user testing. So for the first version, our design prototypes and content were just
our best guess at what we should present to users, which reinforces the importance
of usability testing; we completely changed the tool after testing it. You don’t want
to create something that people won’t or can’t use.”

Usability Testing
After creating the first version of the application, the team evaluated it by conducting
usability tests with four different campus facilities managers. The internal project
client helped recruit participants through his industry contacts, and since they were
located in different cities around the country, all sessions were conducted remotely.
Users were asked to complete specific tasks with the interface and the design team
observed the sessions via LiveMeeting.

Each usability session was one hour long, and, overall, the usability testing phase
took about 20 hours for planning, testing, and analyzing findings. Even testing with
this very small sample of users yielded great design feedback. There was a lot of
repetition of themes between the different users, and the designers were able to turn
these into concrete interface improvements.

Implementation
The project was stretched over several different phases—and led by three different
project managers in total, due to shifts in staffing and resource allocation. This is far
from ideal but fortunately during the last phase, the writer who was already a team
member was able to step into the role of project manager, avoiding the need to
completely start over with a new person.
Aside from personnel changes, the team faced significant difficulties with
implementing the complex mathematical calculations needed for the application, as
Hoess explains, “From a development standpoint, it was pretty challenging to take all
the calculations from the spreadsheet and convert them into formulas that we could
use on the HTML page. I had to reverse engineer the spreadsheet … For each
formula I created unit tests … which are a way to testing small pieces of code, to
make sure it’s still performing the way you expect. Sometimes you have to go in and
make a change to a program, and it accidentally affects your output, but without the
unit test you might not realize that the output had changed.”

Transforming the calculation results into an appropriate visualization was another

implementation challenge, especially because most graphing programs are good at
straightforward data but this chart needed to show both a projection of current
emissions and subtract reduction goals from that to show the difference in emissions
if those goals are achieved. “The graphing tool is good at doing basic graphics, but
because this application is sort of working backwards, I had to think differently about
how to build the visual display for the end user,” explains Hoess.

The application is primarily built with ColdFusion and Javascript, with a jQuery plugin
to generate the graph. The first jQuery plugin used to generate the graph had
several limitations, including difficulty in lining the legend up with the graph
elements in order to comply with Section 508 accessibility requirements. Hoess
ended up switching to a different plugin. “I found another tool that added a lot more
animation and some nicer options, like the ability to export images.” The unit tests

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 159

 created to validate data calculations were especially helpful in switching between
graphing plugins, to verify that the formulas were still working appropriately.

PROJECT TIMELINE
February 2010 Project brainstorming to identify goal, target audience,
and user personas and scenarios

June 2010 Project kick-off

July-August 2010 Completed phase I content, design, and coding

September 2010 Presented beta version of application at Labs21

conference

November 2010 Conducted usability testing

December 2010– Client review of suggested changes based on usability

March 2011 test, budget re-negotiated, and phase II schedule
developed

March-June 2011 Completed all content, design, and functionality

changes based on usability test

June 2011 Launched the application

LESSONS LEARNED
The NREL team shared a few lessons learned from this design effort:
• Plan for change. “Include enough room in your budget to implement
changes identified in usability testing; we had to go back and negotiate more
money to implement changes from usability testing; It’s better to establish
metrics and a budget to track those metrics,” says Sullivan.
• Test early, test often. “We might not have had as many drastic changes
between phase one and phase two [had we tested earlier],” says Augustine.
• Test with actual users. “When you have your colleague sitting next to
you... they can point out some areas of confusion but it's not the same [as
watching a user,]” says Sullivan.

TEAM
The Climate Action Planning Tool design team included:
• Ruby Nahan, Communications Representative/Project Manager
• Rachel Sullivan, Project Manager/Writer
• Laurence Hoess, Developer
• Erica Augustine, Designer
• Shauna Fjeld, Usability Consultant
• Kristin Tromly, Web Content Producer

Erica Augustine is a senior graphic designer at NREL with more than ten years of
experience. She is the lead Web and user interface designer for NREL's
Communications Office, where she focuses on designing clear and consistent user

160 [email protected] Executive Summary

 experiences for websites and web applications developed for NREL and for the
Department of Energy Office of Energy Efficiency and Renewable Energy. Erica
enjoys refining user interfaces for web applications, studying best practices and
patterns for interface design, and observing user interactions with her work. She has
a bachelor’s degree in Graphic Design and a master’s degree in Marketing.

Shauna Fjeld is a senior web developer at NREL with more than 15 years of
experience in web application development. She works currently as a business
analyst and user centered design specialist for web applications, including audience
research, task analysis, and interaction design. She translates client needs to
technical requirements and works with the technical team to implement them.
Shauna’s passion is making web applications intuitive for end users. She has a
bachelor’s degree in Journalism and a master’s degree in Information Systems.

Laurence Hoess is a senior software developer with more than 10 years of

experience designing and building web applications. He has a special interest in
usability and user interface design. Before turning to software development,
Laurence was a technical writer focused on software documentation. In his free time
Laurence likes to fly fish the rivers and streams of Colorado.

Rachel Sullivan is a senior communicator at NREL with more than 12 years of

experience in communications, marketing, and public relations. Rachel has extensive
background in strategic communications, communications planning, project
management, and both print and electronic communication. At NREL Rachel is an
office-wide leader in user-centered design and usability testing. She is committed to
quality and uses her organizational and project management skills to help improve
every project she works on. Rachel has a bachelor’s degree in English.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 161

 TEAM ROLES

Communications • Interface with client, technical lead/expert, and

Representative manage client expectations
• Develop project goals, schedule, and budget
Project Manager
• Manage project team
• Provide updates to the communications
representative
• Lead user-centered design process
• Responsible for mapping out how users will
Designer
interact with content and functionality
• Responsible for validating design choices with
user testing
Developer • Implementing designs, testing, and bug fixing

Usability Consultant • Advises on user-centered design and technical

solutions

Writer • Writes all content per NREL’s style guide

Web Content Producer • Reviews content and application overall to

ensure consistency with NREL brand and web
content guidelines

The Climate Action Planning Tool design team included: Rachel Sullivan, Shauna
Fjeld, Laurence Hoess, and Erica Augustine. Photo by John De La Rosa, NREL.

162 [email protected] Executive Summary

 eReview (IMF)

Organization: Design team:

International Monetary Fund (IMF) eReview was originally developed in-house at the IMF.
The new interface (June 2011) was developed by an
Application: external agency, Eastbanc Technologies.
eReview, an intranet application that integrates
complex business functions into the Members:
organization’s intranet In-house: Martin Mühleisen, Assistant Director, SPR
(business sponsor;) Deb Reilly, Division Chief,
Headquarters: Washington, DC (USA) Information and Knowledge Management Division
(TGS;) Luzmaria Monasi, Senior Research Officer
(acted as business project manager;) Hari Maddineni,
Team Leader Desktop@IMF team (team responsible
for SharePoint integration, development and
support;) Helene Faures , IT Project Manager; Rajitha
Devineni, Business Analyst; Vera Rhoads, User
Experience Specialist; Bassel Kateeb, User
Experience Designer; Tushar Parekh, SharePoint
Developer; Vijay Challa, SharePoint Architect and
John Valter, QA Specialist.
Eastbanc Technologies: Wolf Ruzicka, CEO
Eastbanc Technologies; Slava Koltovich, Technical
Partner; Alexander Mandriko, Senior Developer and
Evgeny Popov, Head of Microsoft Department.

OVERVIEW
Main function: The International Monetary Fund (IMF) regularly generates reports
on individual countries' current economic situations, which go through several rounds
of internal review. eReview is designed to carry out this document review process on
an integrated collaborative web platform, rather than by exchanging drafts over
email. eReview has streamlined and modernized a core business process by giving
users one-click access to instructions, documents, and comments across an
extensive workflow.
Why it’s unique: While working through a very complex review process, users are
guided by a review status bar (located at the top of the UI) that provides a reference
so that they always know what has been done and what is left to do. Users can
perform their tasks easily using intuitive buttons. Buttons are customized for
different stages and user profiles, so a user only sees the actions available to him at
this specific point in the review.
The organization: The IMF is an organization of 187 countries created to foster
global monetary cooperation and help secure economic and financial stability around
the world.
Design philosophy: The team's design philosophy was: simplify, simplify, simplify.
Streamlining and simplifying the review process was the main goal of eReview when
it was first conceived and was also the driving force behind the most recent UI
redesign. The goal for the project was to put all the information required to conduct
a review in one single location, allowing access to instructions, documents, and
comments in one click, and allowing users to perform actions with ease. The new
interface accomplishes this and more. It is now so intuitive that the organization no
longer needs to provide user training.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 163

 USERS
The user base for eReview is broad and encompasses economists, legal staff, and
administrative personnel from all departments involved in the review process (a total
of thirteen departments plus IMF management.) The app is at the center of everyday
life for IMF staff. Each country report goes through the following stages: A team of
economists goes on a visit to assess a country, and produces a Staff report upon
their return. The country team sends the report to other departments (such as the
legal department) for review. These other departments post their comments, and
then the original authoring team incorporates those comments into the document.
The report is then finalized and sent to management for approval, then finally
presented and discussed at the Board.

So a report is touched by a wide range of people during the course of a review cycle.
The following list shows how the user groups handle the document as it goes through
the different workflow stages:

• Administrative staff:
o Initiates a review process
o Uploads the document(s) to eReview
• Economists:
o Posts comments: some via administrative staff and some
directly themselves
o Checks status: Economists often use eReview to check the
status of the document, find out where it’s at in the workflow,
monitor comments, etc.
o Uses previous reports for reference: Economists also use
previous reports and associated comments as reference
materials for their own comments on current reviews.
• Legal Staff:
o Reviews and comments on reports to ensure compliance

THE APPLICATION
eReview takes a myriad of steps between several people and teams, and whips them
into a clear, intuitive, manageable workflow process accessed from a central location.
This helps not only the reviewers, but also the IMF’s office of internal audit, which at
year-end counts how many reviews were produced and who was involved in each.
Without a central, consolidated application, this audit process was time-consuming
and potentially error-prone. In general, eReview is a one-stop-shop for all, where
users can read instructions, upload documents, post comments, and identify a
document's current status in a single place.
Users may access eReview by clicking an icon, but most users click on the link from
within the email notification automatically generated by eReview when a report
review workflow is initiated.
On the eReview landing page, users can see a list of reviews, filter the list by
country, and sort the list by Status, Area Department Due Date, and OMD due date.

164 [email protected] Executive Summary

 When looking at reviews, the right rail displays relevant calendars, contacts, and
email libraries. This is especially helpful as people may need to attend meetings
about the reviews, and they are all handily consolidated here.

From the landing page, users can see a list of current reviews and open a
review or filter the list of reviews by country (using a dropdown list.) Calendars
are handily placed in the right rail, for easy access to details about meetings
related to the review.

To begin a review process, the user clicks the blue Start Review button in the upper
right. Buttons may not be the most exciting UI elements, but let’s just state why this
one is good: 1) it looks like a clickable element due to the shape, size, edges, and
color treatment, 2) the words in the button, Start Review, imply action as well as
communicating the main point of the button, and 3) it is placed in a visible location,
with priority equal to the importance of the task the button supports.
eReview customizes the display of buttons based on two important criteria: 1) the
step in the process, as you can do certain things at certain points, and 2) the user's
identity, as based on her role a user may or may not see certain commands. So a
user who is not allowed to start a review wouldn’t see this button at all, clearing the
UI of unnecessary elements and feeding to the user the information and features she
does need.

Users may start a review easily by clicking the visible Start Review button in
the right rail.

If you are part of the department who is going to kick off the workflow process
(creating a review) then you begin with a form-like dialog box to assign various
metadata. All of the questions are on one screen, and most are drop-down lists. This
power of suggestion can make answering the form easier.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 165

 The dialog for creating a new review steps users through the process with
questions and answers.

The review process itself is quite complex, involving various steps and approvals,
from multiple departments and divisions. One way to help keep items in order in a
scenario such as this is to present a visual workflow. In this case, it is achieved by
the progress bar at the top of the eReview page. This is the anchor and shows each
step of the review process, including the following: Planned, In Interdepartmental
Review, Interdepartmental Review Completed, Awaiting OMD Clearance, OMD
Clearance Granted, and Country Review Completed. Steps in the workflow appear in
sequential order from left to right; as the steps are completed in the workflow the
corresponding boxes turn green. The current step is indicated in blue and the steps
that have not yet been addressed are grey. This helps users easily tell where an item
is in the process. A user who wants more details will find them in the Review
Information box in the right rail.
The Subscribe to Alerts button in the upper right lets users keep abreast of where
the document is in the review process by requesting to be automatically notified as
changes are made. Users also receive specific email notices that are automatically
generated by the system when someone else assigns them a task (for example, lists
them as a reviewer for a new report.)

166 [email protected] Executive Summary

 The status bar across the top of the review page shows the current step in blue,
and completed steps in green. Users can subscribe to alerts about a review
using the button in the top right corner.

eReview automatically sends users a notification when they are assigned a task,
such as responding to comments.

Once on the individual review page, the more pertinent information about this
particular review appears in the main content area, where the review name, due
date, review departments list, and comments are located. Most sections on this page
are collapsed by default to enable users to focus only on the expanded content they
are currently interested in, with the other content collapsed and out of the picture.
All comments are displayed under the Comments section of the individual review
page, and people may print or email those comments, giving users the flexibility to
keep and share comments in more ways than just in the eReview app itself. The Area
Dept. Review tab is related to the area or department which created the review, or
began the review process. The action buttons in the upper right present the next
steps once review comments have been received: Reply to Comments, Send Email,
or send to the next step in the workflow, Send to OMD (the Office of the Managing
Director.)

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 167

 As reviewing departments post final comments, their workflow tasks close
automatically, removing the need for a separate task window. Comments can
be printed or emailed from within the comments zone.

eReview has many features that streamline the review process, and take
unnecessary work out of it. For example, the document review process encompasses
two different workflows: a) cross-departmental: collecting feedback across multiple
different departments and b) intra-departmental: the process an individual reviewing
department goes through to draft and assemble comments it plans to provide to the
original authors.
The tabs at the very top of the individual review page clearly differentiate between
these two workflows. The first tab, labeled Area Dept. Review, is where the cross-
departmental interaction is managed, and the other tabs are where the reviewers
can conduct their internal departmental reviews. Thus, different actions are available
in the first tab than are available in the other tabs. In the first tab, the department
that initiated the review can respond to comments reviewers made, send an email
to Review Departments, and send the review to the next phase (after review) in the
process, which is to the Office of the Managing Director (OMD.) On the other tabs,
individual reviewers may do intra-departmental action such as reassign their task to
someone else in their department.

168 [email protected] Executive Summary

 A review document has different views accessible via tabs at the top of the
page; the left-most tab is one that shows the overall process, and another tab
is created for each department asked to review the report (to manage that
department's internal review tasks.)

The internal department review screen displays Assigned Reviewers, Reviewers'

Comments, and Internal Workflow History, and options such as Reassign Task;
departments can switch between this tab and the main Area Dept. Review tab
set up by the document’s originating department.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 169

 Within the department internal review workflow, individual reviewers use the Post
Comments dialog box to submit their work. Users are given the options to create rich
text with different typefaces and treatments such as colors and bolding.
Once these comments are submitted by individual reviewers, eReview automatically
notifies the main departmental reviewer (or Country Reviewer,) who aggregates the
comments and posts the official departmental comments on the main Area Dept.
Review tab.

The Post Comments dialog supports rich text formatting.

When individuals have posted comments in the internal department workflow,

eReview automatically notifies the main point of contact for a department and
describes the next step: post aggregated comments back to the original
authors on the Area Dept. Review tab.

After comments from each review department have been addressed, it’s time for the
department that originated the review to send it to the Office of the Managing
Director (OMD.) At this step of the process, they can upload the document they want
cleared, but all of the initial versions of the main document and supporting
documents as well as comments submitted by the Review Departments are available
for OMD to look at. The originating department can also send OMD an email with
additional information and instructions. The selection in the status bar in eReview
moves from Interdepartmental Review Completed to Awaiting OMD Clearance.

170 [email protected] Executive Summary

 When a document is finalized, it’s submitted to the Office of the Managing
Director as a link to the IMF’s document management system. The draft which
was circulated for initial review, as well as supporting documents, is also
present in the document zone. Note the status of Awaiting OMD Clearance,
which indicates that the document has been submitted to the Office of the
Managing Director, but not yet approved.

A pop-up message appears on hover and notes what the button does.

Once the document is cleared, OMD emails the initiating department that the
document has been granted clearance.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 171

 The message from the OMD states who approved the document and where
comments are posted.

It’s almost impossible to make a mistake using IMF’s eReview application. The
workflow, reviewers, departments, and associated steps are so ingrained in the
process that it is seamless. Creating, commenting on, and approving reviews is leaps
and bounds better than the organization’s old system, and holds a candle to any app
out there.

DESIGN PROCESS
IMF staff used to rely on email to share documents back and forth. Seeing the
limitations of this process, the IMF looked for a solution and SharePoint seemed to
do the trick. People in the user community liked the idea of eReview. They could see
why it would be useful to have everything in one location. But the process of getting
it in that location was the pain point.

Using the SharePoint 'Out of the box' UI

The original eReview user interface was set up using SharePoint out-of-the-box
collaboration features such as document libraries and discussions. Though the online
review platform and automated workflows were a step in the right direction, people
struggled because the user interface was difficult, confusing and lacking. This made
the whole review process cumbersome. People used the system since it was
mandated by IMF management, but they were not satisfied with it. The standard
SharePoint interface proved to be a functional technology solution, but it did not
provide a satisfactory user experience.

172 [email protected] Executive Summary

 Feeling the Pain
The first step toward creating a new interface was for the design team to map all the
places users were going to when they needed to retrieve information and figure out
all the steps they were taking to process their documents in SharePoint. This process
was eye opening for the design team. “When you really write down all they [the
users] needed to know to be able to use the interface, you realize the complexity of
the old UI,” says Helene Faures, IT Project Manager. “People who were using
eReview every day would suffer but at least would remember the steps to go
through. But those who were only using eReview occasionally would forget from one
time to the other.”
So, the users turned to the staff for help. “They would get confused and didn’t know
what to do to perform basic tasks such as uploading a document. They could find
some things, but not others,” says John Valter, QA Specialist. “It kept me busy.”
It had reached the pain threshold with support staff receiving 3-4 calls per day from
people trying to use the system.
“I was forever getting calls,” says Valter. “Sometimes people would even circumvent
the system. They would actually send [the document] through an email because they
got confused over how to use the old interface and therefore they thought it was
unreliable.”

Redesigning the Interface

It was clear that the UI needed work so the team conducted the following exercises
to create the new interface and make sure that it addressed the needs (and pain
points) of the users:
• Requirements gathering: Instead of a formal requirements
gathering process the team looked to a variety of existing data sources
to inform their design decisions. Through feedback received while
conducting user training, support calls, and desk support visits, the
team felt they had a very good understanding of the issues users were
experiencing.
• Wireframes: The design team went through several wireframe
iterations (with different layouts) and several designs (with different
color scheme and styles.)
• User reviews: Budget and time constraints meant the team had to
limit the type of task-based testing required by formal usability
studies. They chose to conduct a couple of working group sessions to
walk users through screen mock-ups and assess their ability to
determine a clear path to what they wanted to accomplish on the
page. “I think that we got valuable feedback from the sessions,” says
Luzmaria Monasi, a Senior Research Officer who acted as the Business
Project Manager.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 173

 • Working groups: Findings from the working groups helped the team
make design decisions, such as how to adjust the labeling of buttons
and headers to make them more explanatory. Users also suggested
including explanatory text when hovering the mouse over specific
buttons and headers, and provided feedback on other issues such as
the positioning of buttons, how to show certain tabs or headers, and
what not to show.

An analysis of some of the things needing improvement in the old UI.

174 [email protected] Executive Summary

 Pictured: The first design sketch the team scribbled out on paper to try to
improve the UI.

LESSONS LEARNED
The team members share a few lessons learned from the redesign effort:
• Involve the users. “It is very important to get user feedback while
going through the design process. Work together with them. We had
great partnership with our business sponsors and we involved actual
users, economists and assistants throughout the process. We ran the
design through them and asked specific questions such as how to
name a particular button so that it is most intuitive,” says Faures.
• Plan for delays. It may seem counterintuitive to plan for delays, but
even the most fine-tuned project plan can run into trouble at some
point and it’s better to leave a little wriggle room in the schedule for
the unexpected (but inevitable.) The project encountered several
unexpected complexities, “The development of the UI was more
complex than planned and introduced unexpected delays (particularly
the integration with our existing document management system.) The
QA testing and user acceptance testing took a lot more time than
anticipated due to the complexity of the application,” says Monasi.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 175

 • Building on top of SharePoint isn’t easy. What the team learned
about SharePoint was that it is complex to integrate a new UI into
standard platform. “We spent significant time trying to figure out how
to have the new UI integrate with our Document Management system.
In the old UI, we were using an out-of-the box third-party tool to
interface between SharePoint libraries and our document management
system. Now with this graphical layer on top of SharePoint, we could
no longer use that tool. We had to develop custom connections to the
document management system. It was a lot more difficult to
implement than we initially thought,” says Faures.
• Understand the other team’s perspective. “Being in the middle,
between the technical people and the business side, I realized that
what looked very straightforward and easy for the IT people, was not
straightforward for the users. So that was crucial: understanding what
users were seeing, and listening to them to try to understand what
they wanted,” says Monasi.

TEAM
The IMF hired Eastbanc Technologies to design, develop and integrate the UI into the
SharePoint environment. This project was a close collaboration between the two
teams involving both Eastbanc and IMF project managers, designers, usability
experts and developers.
The team roles are outlined below:

176 [email protected] Executive Summary

 APPLICATION DESIGN TEAM
Role Responsibilities
Project Manager
• Guide feature and functionality specifications
• Set project schedule
• Liaison with business stakeholders and users
• Facilitate user communication
Business Analyst
• Document functional and technical requirements
• Test design choices through user research
• QA testing
Designer
• Map out how users will interact with content and
functionality
• Create wireframes and graphical design
User Experience
• Provide input on usability of design
Specialist
Developer
• Implement designs
• Conduct testing and bug fixing
QA Specialist
• QA Testing

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 177

 The IMF eReview project team included two staff from the Strategy, Policy, and
Review (SPR) Department, which is heavily involved in the review of country papers
produced by the IMF, and several staff from the Technology and Services (TGS)
Department. A close partnership with the business users helped the technical team
understand the needs of the eReview user community as well as providing business
credibility to the effort.
Eastbanc Technologies was contracted by the IMF to design and develop the new
eReview interface. Using JQuery over standard SharePoint 2007, Eastbanc built a
user interface that has become the envy of other teams at the IMF.
The IMF’s Chief Information Officer, Jonathan Palmer, has long expressed the belief
that technology should be a “delightful experience” for the end users. This
philosophy was an underlying factor in the development of the new eReview
interface. In addition, all parties involved strongly believed that an easy, intuitive,
and pleasant interface could be the façade behind which complex business processes
are carried out, similar to well-known online shopping applications, where shoppers
can easily review the products they purchase. And this is indeed what the IMF has
managed to do with eReview.
In the months since implementation of the new eReview interface, user feedback has
been overwhelmingly positive. The team is often asked when eReview can be
expanded to include other types of papers produced by the IMF, proving that
eReview has indeed become a delightful experience.

Members of the IMF design team (left to right): John Valter, Evgeny Popov,
Helene Faures, Slava Koltovich, Tushar Parekh, Hari Maddineni, Vijay Challa,
Bassel Kateeb, Deb Reilly, Martin Mühleisen, Wolf Ruzicka, Rajitha Devineni and
Luzmaria Monasi. Missing from photo: Vera Rhoads and Alexander Mandriko.

178 [email protected] Executive Summary

 Eventbrite “Create” Workflow

Organization: Eventbrite Inc. Design team:

Application: Online Ticketing Platform • Melissa Cooper, Senior Interaction Designer
• Julia Thompson, Vice President of Product
Headquarters: San Francisco
• Laura Coltrin, Product Manager
www.eventbrite.com • Adit Vaidya, Group Product Manager
• Tom Censani, Product Design Manager
• Ryan Lottering, Visual Designer
• Dana Lamb, Senior Program Manager
• Galen Krumel, Engineering Manager
• Nathan Yergler, Senior Software Engineer
• Nesan Waran, Senior Software Engineer
• Michael Scheibe, Software Engineer
• John Terrenzio, Software Engineer
• Gordon Mei, Software Engineer
• Yinka Fatona, Senior Software QA Engineer
• Inga Gadimova, Software QA Engineer

OVERVIEW
Main function: Eventbrite is a self-service software tool used by event hosts to
organize, sell and manage tickets for events. The Create Event workflow described in
this report is the process used to set up a new event within the Eventbrite, including
entering details about the event and determining how to structure and price tickets.
Why it’s unique: The new Eventbrite Create Event workflow was chosen as an
example of a great design because it:
• Elegantly communicates a complex set of choices in simple terms, reducing
the number of steps from eight to just three.
• Makes excellent use of progressive disclosure to make complex settings
available, but not cluttering the main workflow
The organization: Eventbrite is used by hundreds of thousands of event hosts in
150 different countries to organize concerts, professional conferences, intimate
classes, private social events, nonprofit fundraisers, political rallies, competitive
athletic events, and everything in between. There were 21 million tickets issued in
2011. Eventbrite offers organizers an all-in-one, technologically sophisticated
solution with social media integration and seamless payment processing.
Design philosophy: Two major themes guided the redesign: the need to simplify
the design while retaining powerful features, and the desire to help users feel
confidence in the interface and in their own decisions. These principles informed
many different aspects of the final design, from the sequence of steps to the
importance of the Preview function.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 179

 WHO ARE THE USERS?
Eventbrite has two audiences: organizers and ticket buyers. The Create Event
process, however, is only used by people setting up events. Within this group there
are a few different types of people with distinct goals and behaviors.
Eventbrite started out as a do-it-yourself service for small events but has since
grown into a more robust resource for larger organizations. The range of people
using the Eventbrite Create Event flow now includes:
 Professional event planners view Eventbrite as their ticketing platform, are
likely to rely on the sales team for help in leveraging features, but may not
spend much time personally using the application
 Support staff for professional event planners who deal with day-to-day
management of event information, and tend to blame themselves if they can’t
find something
 Self-taught enthusiasts organize events related to their personal interests,
may have little knowledge about Eventbrite’s features, and are likely to
overlook or miss things and interact with the customer service team.
Both the larger professional clients and the “long tail” of individual users are
extremely valuable to Eventbrite, so the interface must meet the needs of both
groups. It must be simple enough for users to do tasks quickly without being
overwhelmed by detail, and powerful enough to handle more complex ticket and
event arrangements.
THE APPLICATION
The Create Event workflow is arguably the most important and difficult thing users
do with Eventbrite, and it affects the experience of everyone who uses the
application. The project to redesign this page, which began in 2011, had a few
specific goals:
• Make it easier to set up and publish an event page
• Make it easier to discover advanced features
• Help organizers feel more secure and confident about using Eventbrite
• Increase the number of events published
• Decrease the time to publish (lag time between starting the creation process
and actually publishing an event.)
The new design delivers on the above points with a simple and straightforward
process that is even more impressive when you consider the full scope of what the
page needs to accomplish. In addition to letting organizers input basic information
such as the event name, time, and location, this workflow includes the ability to:
• Create multiple ticket types and prices
• Set start and end dates of ticket sales
• Determine the event capacity
• Schedule recurring events
• Customize visual appearance (background color, text color, images)
• Select promotion preferences such as search categories

180 [email protected] Executive Summary

 In a single page, the Eventbrite Create workflow provides access to both basic
and advanced options for setting up an event and selling tickets.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 181

 TASK FLOW
The new Eventbrite Create Event flow reduces the number of steps in the event
creation process from eight to three while simultaneously exposing more settings
and options to users. This apparent contradiction was elegantly accomplished
through a combination of in-depth task analysis and layout optimization.
The previous Create Event flow had significant issues with feature discoverability, as
Melissa Cooper, Eventbrite’s Senior Interaction Designer, explains, “People would
talk to us because they loved our product and thought it was extremely intuitive, and
tell us a story about a function that we didn’t have and how it affected them, except
[I knew that] we actually had the function. So it was a case of the things that people
didn’t know that they didn’t know that was failing in our software.”
The design team considered several possible structures for the new Create Event
workflow:
• Edit the preview: Editable versions of each page the ticket buyer sees
• Wizard: Walk users through several steps
• All details, customize: A two-step process for entering information and
customizing
• Single page: All actions presented on one screen.

Early testing found that the Edit the preview workflow did not match users’ mental
model of the activity: They had come to the site thinking about creating an event
details page, not registration and order pages. The Wizard alternative was also
problematic, because separating the process out into several pages resulted in a
drop-off and fewer people actually completing the workflow.
The process of outlining and testing these alternative workflows was valuable
because it helped better define the problem space and constraints. Cooper explains,
“We ended up with the All details, customize structure, but it was the single-page
model that got us there. From testing the single page concept we learned that
people really wanted to be able to review everything in one place.”
While users liked the single-page structure, they were daunted by the length of the
page. “So then the challenge was to work out how to make the single page suit the
design goals of simplicity,” says Cooper.
In addition to current Eventbrite customers, the Create Event flow serves an
important function for people who are just considering the service, and want to find
out what features and functionality it offers. The single-page design has a clear
advantage for this type of user, since they can immediately see all the functions.

182 [email protected] Executive Summary

 Early planning explored several different possible workflow structures, such as a
linear wizard or an editable preview.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 183

 Users liked this early prototype of the single-page structure, but found the
length of the page daunting.

Improved Layout
One of the most obvious improvements in the new design is the change from a two-
column to a single-column layout. The previous version was intended to mimic the

184 [email protected] Executive Summary

 end product, the Event Details page, which displays the event location and time in a
right-hand sidebar.
In general, using consistent layouts is helpful, but in this case there was not a close
enough relationship between the Details page and the Create Event page to make
the consistency meaningful. When people first arrive at the Create Event page, they
may have never seen a Details page before, or may not have looked at one recently.
Users aren’t likely to recognize the layout or have a strong association with that
placement, so there isn’t much value to be gained by making them the same.
There is, however, a significant drawback to the two-column layout for a data entry
page: It doesn’t present a clear path to guide users through the process.
Observation of people actually using the page showed them jumping around
haphazardly between different fields and elements. In part this was because the
placement was optimized for people consuming or reading details about an event,
rather than for people writing or engaged in inputting event details.
The new layout arranges the steps in a single column down the page, creating a
straightforward progression for users to follow.
The new visual design also calls out each of the four steps with a bright number that
jumps out from the page. This page element is a favorite with the Eventbrite
customer support team because the clear numbers act as labels for the steps, which
makes phone support conversations much easier.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 185

 Left: The previous Create Event workflow placed some elements on the sidebar,
creating a scattered visual appearance. Right: The new design presents steps in
a single vertical column, creating a clear visual path.

186 [email protected] Executive Summary

 Eventbrite’s standard event details page displays the ticket information and
event description on the left and the time, location, and host information in the
right sidebar.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 187

 The previous Create Event page relegated core elements such as the event time
to the right sidebar in order to use the same layout as the end product (the
Details page). This design was not entirely successful.

188 [email protected] Executive Summary

 Task analysis of the previous Create Event flow shows how the layout forced
users to jump back and forth around the page repeatedly. Each colored line
represents a different user who was observed using this page.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 189

 Improved Task Sequence
In addition to changing the layout, the redesign significantly altered the order and
grouping of input fields. For example, the new step 1 includes the input fields for
title, details, time and location; which are items that used to be divided into steps 1,
3, 4, and 5.
Most users coming to this page already know exactly what they will put in the fields,
so the form is easy to complete. Grouping fields in a single task doesn’t feel
overwhelming to users, and in fact gives people a sense of accomplishment and
progress when they make it through the first step so easily.
This grouping specifically omits step 2 from the old design: Add Ticket Information.
This omission was a deliberate choice, because of the complexity of setting up ticket
types and prices, says Cooper, “We were asking people to do the hardest thing on
the page as the second of eight steps… But in testing we found that people just
wanted to do the easy stuff first: name, date, time, location. And that resonates in
the design that we have today… With step one you’re halfway through the page, and
you’ve done it in just five minutes.”

Step 1 now combines the easiest input fields (which used to be divided into four
separate steps) into a single task, letting users quickly make significant
progress in setting up an event.

The previous sequence was also problematic because it encouraged users to overlook
some important options, contributing to the feature discovery issues experienced by
many users. The last three steps on the page were:
• Step 7: Add Colors
• Step 8: Set Privacy

190 [email protected] Executive Summary

 • Additional Options.
The privacy options relate to whether the event will appear in Eventbrite search
listings, and which keywords and categories should be used to describe the event.
These settings are important for people interested in promoting an event, but the
placement after the Add Colors step acted as a signal to some users that the rest of
the page was not important. “Personalization is adored by some people, and others
don't care about it at all, they think it's just colors and don't find the important
features underneath,” says Cooper.
The new design has moved the Additional Options items to different locations where
they are more relevant (e.g., Show Number of Tickets Remaining now appears in the
ticket settings step.) Set Privacy has been renamed Promote Your Event to better
communicate the purpose of these options, and promoted to be the third step of the
Create Event process.

The previous Create Event page listed the Set Privacy step after Add Colors,
which made it seem less important to user.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 191

 Settings relating to how an event will appear in Eventbrite’s search listings now
appear as step 3, instead of step 8.

FLEXIBILITY
The new Create Event workflow relies on progressive disclosure to meet the twin
goals of simplifying the interface and making more features visible. This technique
consists of displaying information and elements only at the point of need, which
keeps the main interface uncluttered for the majority of users, while giving those
who need it the flexibility of more advanced options.
For example, not all users need to see the full set of options for creating a recurring
event, as these settings are only relevant for the minority of people who actually do
need to set up a repeat event. Instead of including these input fields on the main
interface, they are located in a lightbox that appears only when a user changes the
value for the Event Repeats dropdown.

Since most events don’t repeat, the default setting of the Event Repeat
dropdown is Never, and most users don’t see or interact with the full settings
for repeating events.

192 [email protected] Executive Summary

 Event Repeat settings are located in a lightbox, which appears only after users
interact with the Event Repeats dropdown.

The procedure for adding an event’s location was significantly streamlined by shifting
the bulk of the process from front-end user input fields to a back-end map search.
Now, instead of five fields to specify the location, users see just one field which is
actually a search box. As the organizer begins to type in the field, he is shown search
suggestions. And after selecting a suggestion or entering the full address, a map of
the location appears.
“Using the Google map was important for simplifying the page: It’s one field now,
not five. Google is kind of doing the work for you there, its assisted search for a
location. We still remember people’s locations, which is important. And if people
want to put in information that wouldn’t traditionally come up in a Google map they
can still do that,” explains Cooper.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 193

 The previous Add Where step had 5 fields.

The new interface has just one field for entering the location.

194 [email protected] Executive Summary

 As you type a location you see search suggestions from Google Maps.

When you enter a location, a map is displayed to help verify that you’ve
entered the correct address.

INFORMATION DISPLAY
Step Two in the Create Event workflow, setting up the ticket structure, is the most
complex part of the process. There are all sorts of choices for organizers to make
here including:

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 195

 • What kinds of tickets to offer (e.g., general admission vs. VIP)
• How much to charge for tickets
• Whether to absorb Eventbrite's service fee or pass it on to the customer
• When tickets sales will begin and end
• How many tickets are available, and whether this information should be
displayed on the event listing page
• How many tickets can be purchased in a single transaction
Not only are there a lot of decisions to make, but the set of decisions can have a
huge impact on the success of the event. And the answers are not necessarily easy
to go back and correct later. For example if a user accidentally selects to pass on the
fee but later decides to absorb it, he is faced with the prospect of have to figure out
how to refund everyone who already bought a ticket.
The previous workflow attempted to simplify this process by using the progressive
disclosure techniques that worked so well for other aspects of the design. The
ticketing structure, however, is actually a case where it's better to show users more
information up front so they can understand their options and make an informed
decision. The old design led to mistakes and missed opportunities, because users
didn't dig deep enough to understand their choices.
For example, organizers need to know the venue capacity when determining how
many tickets they can sell. Cooper explains, "One thing that was really failing in the
old design was, when you create the ticket quantity and save your settings, you
couldn’t see them anymore. And the event capacity was off to the side and hard to
discover. Now we’ve lined them up, and you can go in and create your tickets and
make 100 general admission, and 100 VIP. That adds up to 200, but you can have a
capacity of 150 and you just don’t care how many of each sell. You can lock that
capacity and keep adding different ticket types but your capacity remains at 150.
Things like that are really important in terms of managing capacity, and those are
things that the long tail weren’t discovering. That’s where the interface really helps
make your event more successful, because you’re no longer making those mistakes
or overlooking those details."
The need to help people set up complex business rules while maintaining a simple,
self-explanatory interface creates a huge design challenge. As you can imagine, a
great deal of design exploration, prototyping, and testing went into making the new
ticket creation process.
As Cooper explains, the key goal for this step was to create confidence by providing
straightforward information, “This ticket module is [designed] around transparency
and instilling trust. It’s very much like what the event details page is going to look
like once they get done with it, and it's easy for that self-taught event planner to sit
here and play with the numbers."
A few specific elements contribute to making the entire ticket and pricing structure
much easier to understand and control. These are the following:
• The end price that will be paid by the buyer is automatically calculated and
dynamically updated
• Different service fee options are visible without expanding
• The number of each type of ticket and the total event capacity are visible
• Individual ticket settings are visible on hover
• Two rows of tickets are shown by default, to introduce the idea of creating
more than one type of ticket.

196 [email protected] Executive Summary

 The new Create Tickets module lets organizers see the total price and service
fee options by default, without having to access advanced settings.

Clicking on the Settings link within the ticket row expands a window where
users can customize individual settings (such as the sales start and end time)
for just that ticket.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 197

 Organizers can review the individual settings for each ticket type by hovering
over the Settings link.

The previous design for the ticket creation modal did not display any
information about service fees or sales timeframe in the basic settings.

198 [email protected] Executive Summary

 In the previous version, after creating a ticket you could not see the quantity of
tickets available or the total price paid by the ticket purchaser.

INTERFACE ELEMENTS
Sometimes a picture is worth a thousand words. This often the case when
customization apps, where a real-time preview of changes is much more effective
than trying to describe them.
The Eventbrite Create Event workflow allows users to edit the appearance of
different background and text elements, and makes this process much easier by
providing a preview version that shows the actual page on the same screen as the
customization controls, and updates dynamically as you make changes.
This full preview was difficult to build but really helped meet the goal of building the
organizers' trust and confidence. Cooper explains, "You want to see all your work
and feel confident that your settings are correct, and you understand what’s going to
happen with your page when you publish it.”
The previous Create Event workflow did offer a preview function, but it launched in a
new window and did not incorporate the actual customization controls into the
preview screen. So organizers couldn't actually see the changes they were making in
real time, they had to switch back and forth between the controls and the result.
The preview function is accessible through a tab rather than being a step in the main
workflow because, as Cooper explains, "The [preview] is of really high value to many
of our customers, but it’s also something that some people don’t even want to look
at. So they can easily skip it.”
For those who do care about the preview function, the design provides constant
access by keeping the top tab control bar visible even if you scroll down the page.
“As you scroll down the page, the bar at the top is persistent to the top of the
browser. So we’ve kept the Save and Live buttons visible at all times, and you can
toggle between [Event Details and Preview & Appearance] without having to save,"
says Cooper.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 199

 The Preview and Appearance displays the event information as it will appear on
the final event listing page, along with customization controls. Changes made in
the top control panel are immediately reflected in the page below.

200 [email protected] Executive Summary

 The preview is accessible via a tab at the top of the page; as you scroll down the
page, the tab moves with you so that it is always visible.

DESIGN PROCESS
Since Cooper had just joined Eventbrite as the user experience lead, she conducted
extensive user research to familiarize herself with the product and customers. This
research was particularly focused on the Create Event workflow (since it is such a
critical piece of the Eventbrite experience,) but she also took time to gather general
background information about event organizers and their goals and behaviors, “This
was a huge project but it’s really core to our business, it’s absolutely crucial for
people to complete this process. …The product managers are always out there
talking to customers, but they tend to have very specific needs when they go out. It
had been a significant amount of time since we’d done some dedicated, deep-dive
research with our customers.”
Her user research activities included:
• Observation of customer support calls
• Workshops with customer support representatives
• Exploratory interviews with customers
• Usability testing of the existing design

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 201

 • Behavioral Analysis of web analytics data
• Usability testing of multiple prototype iterations.

An outline of the user research and design plan for the Create Event redesign
project.

Workshops with Customer Support

In order to take advantage of existing internal knowledge about users’ needs and
problems, Cooper listened to recorded support calls and conducted a series of
workshops with Eventbrite’s Customer Support team members. “I threw key pages
from the website up onto the wall and had them hurl problems at me for an hour per
workshop. So I knew from that that there was really poor discoverability in our
Create Event flow.”

202 [email protected] Executive Summary

 During a series of workshops, customer support staff identified recurring issues
with the current design.

Exploratory Interviews
In addition to talking with support staff, Cooper interviewed Eventbrite users directly
to find out about their usage patterns. A range of users from both large and small
organization participated.
These sessions were targeted more towards gathering general knowledge, “I’d start
broadly with the general event planning process and see where they go, then shift
into more focused questions about software. I learned how people view the product
as a whole, what it means to them, how they position it in their mental model, and

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 203

 what needs were and weren’t being met, before going anywhere near event
creation… When we got to event creation…because I’d conducted the support
interviews I knew there were things about the page that didn’t work. And when
people said, ‘It’s easy,’ I knew there were things they hadn’t found, because of
things they’d said earlier…I found a lot of cases where people didn't know what they
didn't know.”

Usability Testing
Behavioral research in the form of usability testing was another important piece of
the design process. The first phase included usability tests of the existing design,
with users from different kinds of organizations who had different levels of
experience with Eventbrite.
After synthesizing all the findings from the interviews, support team insights, and
user testing of the old design, Cooper identified several different possible workflow
structures, and developed visual representations of each. These were then tested
with users by simply opening the image file in a browser. Cooper used Silverback to
recording the sessions and GoToMeeting to broadcast them to others in the
organization.
The next design phase continued to rely on user testing but evolved into a more
focused, detailed prototype, “Once we chose our direction we moved into html
prototyping, and focused on different parts of the page every week. As the weeks
went on it became more and more interactive,” says Cooper.

204 [email protected] Executive Summary

 Rather than a long formal report, the usability testing results were recording on a
“live wall.”

Overall, the extensive user research performed for this project has helped build “a
much better sense of what’s of high value to different verticals, which is really
important,” says Cooper.
It’s easy and quick to throw a whole list of features and controls up on a page, but
the end result is often a confusing jumble that leaves users feeling lost.
Paradoxically, the designs that seem the simplest and most intuitive are often those
that someone spent the most time thinking through and creating. This was definitely
the case with the Eventbrite Create Event workflow, as Cooper explains, “Sometimes

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 205

 I feel like I just facilitated this design. We did so much user testing, then it was like,
‘Oh! This is what you were meant to be.’ That’s great design: When it pops out at
the end it feels like it should have been obvious.”

LESSONS LEARNED
The Eventbrite team shared some lessons learned from this design effort, including
the following:
• Leave time for planning and synthesis of research. “At the end it felt
huge… I did not leave enough time for the synthesis and reporting of my
research. That part’s always going to take a lot longer than you expect. Also,
there's nothing more valuable than using a transcription service, and
recruiting also takes so much time.”
• Iterative design with users is invaluable. “It only takes about two
seconds… for you to actually be able to see something with fresh eyes.”
• Be strategic with your research efforts. “These techniques are highly
valuable and really work, but they are also a significant investment. For us it
was applicable to all of Eventbrite, not just this project. But moving forward
we have to be strategic about where we really want to invest in that level of
research-driven design, and where we want to trust in the experience of the
team.”

TEAM
The design team included:
• Melissa Cooper, Senior Interaction Designer
• Julia Thompson, Vice President of Product
• Laura Coltrin, Product Manager
• Adit Vaidya, Group Product Manager
• Tom Censani, Product Design Manager
• Ryan Lottering, Visual Designer
• Dana Lamb, Senior Program Manager
• Galen Krumel, Engineering Manager
• Nathan Yergler, Senior Software Engineer
• Nesan Waran, Senior Software Engineer
• Michael Scheibe, Software Engineer
• John Terrenzio, Software Engineer
• Gordon Mei, Software Engineer
• Yinka Fatona, Senior Software QA Engineer
• Inga Gadimova, Software QA Engineer

206 [email protected] Executive Summary

 The Eventbrite design team. From left: Back row: Melissa Cooper, Gordon Mei,
Adit Vaidya, Ryan Lottering, Tom Censani, Nathan Yergler. Front row: Inga
Gadimova, Yinka Fatona, Laura Coltrin, Dana Lamb, John Torrenzio, and Nesan
Waran.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 207

 Product Suggestion Portal (Hobsons)

Organization: Design team:

Hobsons, Inc., a (SaaS) provider of CRM Members of the Hobsons, Inc. team did all design and
technology and marketing solutions for education development for the Product Suggestion portal in-
professionals house.
Application: Members:
Product Suggestion portal Abram Greene, Quality Assurance Engineer; Vanessa
Keeton, UI Designer; Jaswinder Johal, Software
Headquarters: Cincinnati, Ohio (USA) Engineer; Sidharth Malhotra, Business Analyst and
www.hobsons.com Vishal Saboo, Product Manager.

OVERVIEW
Main function: As a customer-centric organization, one of Hobsons’ guiding
principles is to seek and incorporate market feedback into its products. The Product
Suggestion portal provides end users of the company’s CRM suite a simple tool so
they can communicate directly with the company’s Product Management team.
Hobsons’ CRM product is a mission-critical tool for the educational institutions
Hobsons serves. With the product suggestion portal the company gives its customers
an opportunity to make suggestions on the spot, while using the CRM suite as part of
their day-to-day workflow. So if they’re in a specific section in the CRM and they feel
like a certain feature or a certain function can be enhanced, or if something’s
completely missing, or if they’re not able to figure out something, they can instantly
make a suggestion, vote and comment on others’ suggestions and track the status of
those proposed improvements as they make their way through the pipeline. The tool
ensures that Hobsons customers have a voice in a product suite that affects their
day-to-day job function.
This tool was a major leap forward for the organization, replacing a previous system
that created too much inbox clutter and prohibited any meaningful analysis of the
data it collected.
The organization: Hobsons is a ‘Software as a Service’ (SaaS) provider of CRM
technology and marketing solutions that empower education professionals to
manage the entire student lifecycle, including recruitment, application, enrollment
and retention. The CRM suite (which this application is a part of) provides clients
with the ability to communicate with prospective and current students using multiple
communication channels: email, text, web portals, social media, telephone, letters,
etc. The application suite provides clients with the ability to create customized
college applications, manage application data, communicate admissions decisions,
and provide a customized experience to each student that utilizes the system.
Design philosophy: “We try to balance the product management principles,
innovation and usefulness in thinking about new products and features. The Product
Suggestions portal is a good example of executing these principles primarily because
of the speed at which we were able to bring a simple, elegant, and integrated portal
to our customers. It continues to be one of primary tools for communication of new
features and ideas between end users and Product Managers,” explains Vishal Saboo,
Product Manager.

208 [email protected] Executive Summary

 WHO ARE THE USERS?
Users of the product suggestion tool are the same users who rely on the CRM
application suite in general. These users fall into two main groups:
• Admissions staff at higher education institutions: Nearly 2,000
end users at 500 client institutions utilize the portal each month and
submit new suggestions to improve the functionality of the CRM and
Application suite of products. They vote and comment on existing
suggestions and follow the status of suggestions.
• Product Managers: Ten users in the Product Manager role are
responsible for reviewing, moderating, and publishing incoming
suggestions, merging duplicate suggestions, commenting on
suggestions, and editing suggestion statuses.

THE APPLICATION
Every page of the CRM suite includes a footer link to the Product Suggestions portal.
When new customers start using the CRM and go through the professional
implementation and training, they see that there is a channel to communicate with
the product management team while they’re using the product. Existing customers
have become accustomed to this link. Before the Product Suggestions portal was
created, users could click a footer link at the bottom of any page to send feedback,
but the link simply generated an individual ticket and sent it to Hobsons’ Product
Managers. There was no system to provide visibility to users about what happened to
the suggestions after they were submitted, and dealing with and analyzing the
suggestions was a near impossible task. The Product Suggestions portal is a savior in
helping to manage and address product suggestions.
The top two tabs in the app, View Suggestions and Submit Suggestion, make the
most important tasks readily visible and achievable, and the names on the tabs
clearly denote the more passive activity, View, and the more active one, Submit. In
the View Suggestions tab users may sort by the sub-tabs: All Suggestions,
Suggested by Me, and Suggested by Co-Workers. This enables people to easily
review and compare suggestions depending on the source. Still, dealing with these
categorized lists of suggestions could be taxing, so designers included filtering
options in the Refine Results section in the left rail. In the top box, a user may refine
results in a few different ways: he can type in a keyword to search, or select a
checkbox to filter by status (Suggested, On Roadmap, In Development, or
Launched.) Finally, he could also sort the list by other criteria such as Recently
Updated. Building a variety of sorting and filtering mechanisms into a UI is an
excellent way to accommodate multiple usage scenarios.
The Suggestions by Product section (above the refining section) offers a list of
popular topics, accompanied by the number of items in each topic. Each topic is an
underlined link for easy affordance and access. UI’s such as this take advantage of
the power of suggestion: a flat list with all items visible, as long as the list is short
and understandable, exposes the user to possibilities she may not have thought of.
Making smart suggestions to the user is not only helpful, but breeds engagement
and loyalty.
Since a finite number of items will fit in the allowable screen real estate, designers
enable users to choose the number of items that appear on the page. This, as well as
the pagination option, appears in the grey menu just above the list of suggestions.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 209

 In a relatively small space designers are able to describe a lot about each
suggestion, namely: suggestion title, who created the item, date it was created,
status of the suggestion, a longer description of the suggestion, a category for the
type of product being suggested, a short description, a link for comments, a link to
choose to subscribe, and finally thumb up and thumb down voting buttons along with
the current vote score. The horizontal section for this area is full but not crammed.
Designers use a combination of larger and bolder fonts to direct the eye to the
different content.

On the View Suggestions, All Suggestions main page, users see the most recent
suggestions activity at a glance. From here, the user may vote and comment on
a suggestion, search for suggestions, or submit a new suggestion.

210 [email protected] Executive Summary

 Sort options on the left include Score (High to Low,) Score (Low to High,)
Newest, Recently Updated, and Oldest.

When a user makes any kind of input—such as an order, comment, or suggestion—

he wants some feedback and the ability to seek out status when he is ready to do so.
The Suggested By Me tab makes it easier on the user to track the progress of his
own suggestions. This view resembles the All tab, but the voting icon colors are less
vibrant, to indicate that the vote has already happened. Additionally, a note at the
top of this page states that even items which are assigned Pending status, and not
posted for voting and comments by others yet, are visible here. But it further
explains Hobsons’ procedure for managing suggestions: to review and then post
them for votes and comments once a month.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 211

 The View Suggestions> Suggested by Me section provides a listing of
suggestions submitted by the user and includes the status of the suggestion.

Allowing people to vote on suggestions serves a few purposes, not the least of which
are: 1) more customers feel heard, and 2) Hobsons can get a better sense of how
important each suggestion may be to the customer base. The voting process itself is
expected and easy, with just two icons placed near the suggestion title. The standard
thumbs up with a green background, and thumbs down with a red background are
understandable. While the icons are small, they are distinguishable based on both
color and the thumb graphic. This is positive for people who are colorblind because
the graphic suffices. For people who are not colorblind but who may still have lower
vision, the different background color may make it easier for them to pick out the
right icon.
Once the user votes, he is presented with a message, for instance, Your Thumbs Up
Vote has been successfully saved. This type of feedback gives users confidence and
enables them to feel in control.

212 [email protected] Executive Summary

 Voting provides a confirmation message in a dialog to ensure that the user is
aware that their action was recorded.

In addition to voting on suggestions, users may also comment on them. Again, this
feature can help the customer base feel heard; it also allows people to offer a
different perspective, and can help Hobsons get a better understanding of users’
desires and views. Users may also read other customers’ comments for more
perspective. The Comments link displays the number of comments already made,
disclosing how much activity there is surrounding a particular suggestion. These
niceties give more information with little effort on the design and development side.
Clicking the Comments link expands an area to expose all comments which have
already been made, leading with the most recent first. Users may, of course, scroll to
read the less recent comments. Above the list of comments appears the Leave a
Comment field. The label accompanied by the open box and Post Comment button
are inviting elements and summon people to type their thoughts. An effective status
label below the field counts down the number of characters the person has left to
play with, a la Twitter’s UI.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 213

 Users may read and add their own comments, elaborating on a suggestion or
providing a different perspective or use scenario.

If a user is particularly interested in a suggestion he may choose to subscribe to it by

clicking the Subscribe link. The message the user is presented with communicates
that the user has subscribed, the name of the suggestion he has subscribed to, and
choices to be notified when status changes or comments are posted. All of this is
pertinent, helpful information for a confirmation message.

214 [email protected] Executive Summary

 After subscribing to a suggestion users are presented with a confirmation
message with the suggestion name as well as subscription options. Once the
user confirms, he will begin receiving email notifications when the suggestion is
updated.

In addition to viewing, commenting on, and subscribing to suggestions, users may of

course make new suggestions. They do this under the Submit Suggestion tab. Here
they are presented with a basic form with fields for Suggestion Title, Product, and
Description. The Product field has a drop-down list, so users don’t have to type in the
product name.
Once a suggestion is submitted, users are again presented with an important
feedback message, which explains what will happen to their suggestion next: Thank
you for your suggestion. All product suggestions are reviewed by our Products Team
on a monthly basis. We publish suggestions based on similarity to the development
work already in progress, the level of effort, and market relevance. Once a
suggestion is published, all clients have the opportunity to vote for and comment on
the suggestion…

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 215

 Submitting a suggestion provides a detailed confirmation message explaining
the process and setting the user’s expectations.

Once the suggestions are submitted, the product managers must review and publish
them before they are available to the rest of the user community for votes and
comments. The Manage tab in the Product Suggestions Portal is designed for these
back-end users who need to analyzing and publish suggestions.
To review suggestions product managers may search and filter by various criteria,
such as: keywords, name of the suggestion, status (such as inactive, archived,
pending, suggested, on roadmap, in development, launched, and merged,) and date
or range of dates the suggestion was created on.

The user may also click the Download Suggestions button to copy the suggestions to
a spreadsheet file where they can manipulate the information. Allowing people to
download or export the data and use a different, powerful program such as Excel is
usually far better than trying to build some of the basic spreadsheet features directly
into the app. Users may also view the Manage> Reports area for a summary of
suggestions.

216 [email protected] Executive Summary

 The Manage Suggestions tab provides Product Managers with a simple way to
search by different criteria for suggestions. Suggestions can be modified in bulk
and downloaded into a spreadsheet for additional analysis.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 217

 The two suggestions reports provide a snapshot of the Product Suggestions for
all products in the various statuses; the top report summarizes all portal
activity, and the bottom report shows only activity from the last month.

As product managers address the suggestions they are able to change each
suggestion’s status to reflect its current stage in the workflow. They may do this
directly on the page of suggestions by selecting the item’s related checkbox on the
left, then choosing from the Mark Suggestion as drop-down list of choices.

218 [email protected] Executive Summary

 Product managers may select suggestions in the Manage tab and then assign a
new status.

Once the status is updated, the product manager sees a confirmation, and is asked if
he would like to send an email to subscribers notifying them of the status update.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 219

 Product Managers see a confirmation when they update a suggestion’s status,
and can choose to email subscribed users about the new status.

The email notice sent to subscribers presents some automatic information about the
updates, such as the name of the suggestion and the new status category assigned
to it. It also has a link to the suggestion for review.

220 [email protected] Executive Summary

 Product Managers can easily send an automatic message to subscribers,
informing them of a status change.

Additional maintenance is required to keep the suggestions in order, namely

eliminating duplicate suggestions. The team’s research shows that 20% of
suggestions each month are duplicates, and users do not search for suggestions to
see if their idea has already been entered before creating a new suggestion. In the
future the team plans to offer a smart, dynamic area on the page where people
make suggestions. As the user types, the app will query the current body of
suggestions for similar or same suggestions and present them to the user to choose
from and comment on.
In the meantime, until that feature is created, product managers may merge a
duplicate suggestion and notify the user who made the suggestion. They find like
suggestion entries and then use a merge feature to consolidate them. After opening
a suggestion they would like to merge, the product manager can choose the Merge
To button in the upper left of the page.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 221

 Users select the Merge To button in the upper left to merge the currently
selected suggestion with another suggestion.

At that point the suggestion is ready to be merged, and the next page says as much,
repeating the suggestion again, and asking for the ID for the suggestion it should be
merged with. Once the ID is entered and the page submitted, the system displays a
page with both suggestions visible. This preview helps users understand the process
and feel confident about what they are about to do.

222 [email protected] Executive Summary

 The Merge Suggestion page displays both suggestions to help product
managers understand exactly what it about to happen, before the system
actually merges the documents.

Once the suggestions are merged, the merged suggestion shows up at, basically, the
first comment in the comments area. The merged suggestion appears in front of a
yellow background, so it is differentiated from the comments, and the heading is
titled Merged Suggestion.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 223

 Merged suggestions become special “comments” to the main suggestion. Also,
any user votes for a merged suggestion are automatically added to the count
for the parent (merged with) suggestion.

Since Product Managers are in contact with many different types of customers and
people at the organization who interface with customers, they have their finger on
the pulse of what customers want. Thus, the Suggestion Portal app makes it possible
for Product Managers to vote about a suggestion on behalf of someone. This way the
importance of the suggestions they hear are not lost, and instead are voted upon in
the system as though the third party had access, interest, and time to do it.
To use in this feature, the Product Manager clicks the Product Manager Update
button in the Edit Suggestion view. After clicking it they are presented with a dialog
box and a drop-down list of sources to choose from, including: professional services,
current client, prospective client, sales, pre-sales, product engineering, product
development/ QA, or other.

224 [email protected] Executive Summary

 Product Managers can vote on behalf of an internal or external customer using
the Product Manager Update feature when editing a suggestion.

Product Managers are asked to choose a source for the vote.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 225

 Product Managers may vote on behalf of the following sources: Professional
Services, Current Client, Prospective Client, Sales, Pre-Sales, Product
Engineering, Product Development/QA, or Other.

The Suggestion Portal at Hobsons supports a core business edict as well as workflow.
It’s imperative that the organization collects feedback about their apps. On the
flipside, customers expect to be communicated with and responded to when they
make a suggestion. Ordinarily this makes for a lot of work for an organization, but
this suggestions portal gives users much more visibility into the process, with far less
effort on the part of product managers. The app takes much of the burden off the
users and deals with it efficiently and transparently. The creation, review,
commenting, and sorting of suggestions is an inclusive, easy process that is
seamless to work with.

DESIGN PROCESS

Usability Methods
The design team knew they had to address the inefficiencies of the old system, but it
wasn’t clear at the start if they could or should build their own tool, or piggyback off
an existing platform. So, the first step in addressing their challenge was to
exhaustively research the landscape of ideas and tools while conducting a thorough
cost analysis.
Learning from similar tools. One of the first research efforts the team embarked
upon was to look at other, similar tools, across their own organization for guidance
and best practices. Though these tools are admittedly different products with a
different user base, the concept was similar: give users the ability to submit
suggestions. The design team felt they could learn from their colleagues’ hits and
misses by getting first-hand knowledge from those teams about what worked and
what didn’t during their development efforts. The design team also analyzed public-

226 [email protected] Executive Summary

 facing customer suggestion tools, such as Starbucks’ Ideas Portal and
GetSatisfaction.
The research results made it clear they would have to build their own tool, from
scratch. So, the first step was to recruit a development team to build a prototype as
a proof of concept. This effort cemented their mission: to build a custom tool.
Site visits. Hobsons Product Managers occasionally visit customers, so during the
design phase of the project, they used some of the time spent during those visits to
gather data about how customers use the Application Suite. This data influenced how
they designed the suggestion tool. Observing users in the field was a big opportunity
for the design team. During a routine customer visit, if there was a little extra time
at the end, the team would make the most of it. They shadowed customers as they
used the application, probed when necessary to understand a workflow, and basically
just observed how customers were using the product. These observations, along with
feedback in response to ideas generated by team members informed the feature set
and design of the tool.
Using customer calls to inform design. The business analysts who create the
wireframes do not attend site visits, but when work begins on a new feature they are
there, right from the start. During the initial design phase for this project they were
also able to listen in on customer calls. Customer calls are an opportunity to give
customers a glimpse of what to expect in the product over the next six to nine
months and for the team to hear the feedback. There may be up to 40 customer calls
per year so the design process is closely in touch with the opinions of the end users.
The feedback gathered during these sessions gets rolled into the wireframes the
analysts create.
All of these activities—competitive research, site visits, and customer calls—informed
the development of early wireframes and prototypes. The team relied on user
feedback to guide some key changes as the design evolved, as Product Manager
Vishal Saboo explains:
• Combined search box: “The search box was originally placed in the
top right because that’s the most common place for a Search box.
However, as we iterated through the design, we realized that
searching, filtering and product breakdowns made sense to all be
together. That’s what we did in the final design.”
• Additional statuses added: “We started out with just four status
options: In Review, On Roadmap, In Development, and Launched. But
as we iterated through the design, we realized these were not
sufficient. We added five additional status options: Pending,
Suggested, Archived, Merged, and Inactive. The Pending status was
added to help incorporate a significant process change suggested by
one of our senior executives: that Product Managers should review all
suggestions before making them visible to other customers for voting
and commenting.”

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 227

 Pictured: Originally, the search box was placed on the top right, and there were
only 4 status options.

Final Design: Searching, filtering and sorting options are grouped on the left.

Designating a stakeholder team. Another way to keep the end users close to the
design process is by creating a proxy group that represents the needs of the end
users. The design team at Hobsons handpicked a group of internal and external
stakeholders, drawn from design partner organizations and clients, to participate in a

228 [email protected] Executive Summary

 steering committee. The design team returns to this group every six months or so
and presents whatever’s on deck on the roadmap in terms of product development
and features. This group, with its interest and expertise, can act as a sounding board
for potential initiatives — both providing feedback on new ideas and providing ideas
themselves. The Hobsons team did this early on in the design process for the
suggestions tool.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 229

 PROJECT TIMELINE
Late 2007
Launched a “Product Suggestions” link in the CRM tool, but
beyond simply collecting the feedback, there was no follow-on
process and no good way to parse the responses. It was
untenable as a feedback collection mechanism. The team was
bombarded with emails, clients complained about getting no
response to their feedback.

January 2010
Commissioned a project to investigate better options. Some
existing ideas portals were mined for ideas and best practice
examples, including: Salesforce.com, Starbucks Idea,
GetSatisfaction, etc. Buying and integrating an existing tool was
deemed too expensive so the decision was made to build one in-
house

February 2010
Development team built a prototype proof of concept (took one
month to build)

April 2010
Official project launch: the company gathered a team of
business analysts, UI designers, developers and started
wireframes and design mock-ups

August 2010
Application launched

Post-Launch
The application has gone through four additional releases since
its August 2010 launch. These releases have added incremental
functionality including:

• Ability to download suggestions (for product managers to

parse the suggestions)
• Reports
• Added a merge feature to reduce the amount of
duplicate submissions. This sends and automatic
notification to client that their suggestion has already
been submitted.
• Notes: product managers can make notes on suggestions
• “Proxy vote": feature that allows a product manager to
vote on behalf of someone else
• Subscribe and Notifications features added
Future Phase
The team is considering integrating the suggestion application
with internal roadmap tool (an Excel spreadsheet)

LESSONS LEARNED

230 [email protected] Executive Summary

 The Hobsons design team shares a few lessons learned from designing the
suggestions tool:
• Start small. Sometimes even a good idea can be too big a project to tackle.
Start with a small, achievable piece and work from there.
• Choose the right people. When choosing members of a stakeholder group,
choose wisely. Choose people who are close to the customer, who do training,
who visit clients regularly, or who are early adopters.
• Stay in touch. Once you have a solid steering committee in place, make
good use of their services. Reach out to them every six months, at least, to
keep their interest and participation alive.

TEAM
“Our customers, who are the users of the Product Suggestions portal, are the ones
that seeded the idea for the Product Suggestions portal and deserve ultimate credit.
Once the need was evident, a small team got together to do a rough prototype,
somewhat of an anomaly from our typical Product Management process. The team
was influenced by Marty Cagan’s book Inspired: How To Create Products Customers
Love,” explains Saboo. The design team for this project is located within in the
product management department; their roles and responsibilities break down as
follows:
Application Design Team Responsibilities
Role Responsibilities
Product Managers
• Ownership of features and functionality;
Determining when features are released and what
is going into each release
Business Analysts
• Brainstorming with Product Managers and end-
users to create wireframes and functional specs
User Interaction
• Realize the vision of the wireframes in HTML and
Designers
mockups with a special attention to usability

QA Engineers
• Writing Test Cases, integration/performance
testing, reporting defects, regression

Back-end developers
• Implementing designs, unit-testing, and bug fixing

Vishal Saboo is a Product Manager at Hobsons. He is passionate about design and

loves simplicity. He is responsible for all the functionality in the product and is the
driver behind new features and improvements. He graduated with a Masters in
Business Administration from The Ohio State University, and a Bachelor’s degree in
Instrumentation Engineering from Mumbai University. His interests include tennis,
photography and music.
Sidharth Malhotra is a Senior Business Analyst at Hobsons. He is a creative
problem solver. He is responsible for requirements gathering and building solutions
for the product roadmap. He has over 10 years of IT experience and has worked in
different roles in the software development lifecycle. He graduated with a Master’s in

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 231

 Business Administration from Chenoi University in Mumbai, and a Bachelor’s degree
in Business Administration from Mumbai University. His interests include squash,
cooking, travel and music.
Abram Greene has been with Hobsons for over three years. He is a graduate of
Spencerian College, with a degree on Computer Engineering Technology. As a key
member of the Quality Assurance Team, he is responsible for making sure our
products are effective, functional, consistent, intuitive, efficient, usable and reliable.
Outside work, Abram is passionate about his UK basketball team, and loves to play
golf.
Vanessa Keeton is an Interaction Designer at Hobsons. Vanessa is very passionate
about clean and reusable solutions for design and coding. Before she became a
professional designer, Vanessa played and taught saxophone professionally. She has
a Bachelor’s in Jazz Studies from Morehead State University and an Associate’s in
Interactive Media Design from the Art Institute of Ohio — Cincinnati.
Jaswinder Singh Johal is a Software Engineer at Hobsons. He is a developer with
wide array of skills and a broad knowledge base. He is responsible for writing and
maintaining the code behind the product. He graduated with a Master’s of Science in
Information Technology from Punjab Technical University, and a Bachelor’s degree in
Computers Application from Punjab University. His interests include music and
learning about new technologies.

Pictured: (left to right) Vishal Saboo, Sidharth Malhotra, Abram Greene,

Vanessa Keeton, and Jaswinder Johal.

232 [email protected] Executive Summary

 Multi-Platform Applications

Applications’ in the Multiplatform area have transmedia strategies and goals which
are well defined such that the applications actually remain true to their brand across
multiple environments. Designers in the multiplatform area ensure that the core
functions are supported appropriately across diverse platforms while taking full
advantage of each medium’s finest elements, and avoiding the feeble areas. For
example, designs in this group exploit the power and screen real estate a desktop
environment affords, adding more features, such as analysis and manipulation,
commenting, and sharing. These app designers focus on the importance of mobility
and always-available environments, but they know the limitations of small screen
displays and restricted input devices, and design within those confines. Designers
also know the nuances between similar media, such as the fact that designing for a
smartphone display is only similar and is not the same as designing for a tablet.
The winning design teams in this category were able to keep straight not just one
operating system or hardware platform but several. They handled multiple
strategies, schedules, and the tasks of sharing design sometimes across teams, all
while keeping true to the wonder and limitations of the application’s central goals
and to each medium.
Winning applications in this category include:
• Dixio, multiplatform lookup technology
• Highlight, lets users with iOS devices make audio recordings, mark
interesting points, and share recordings with others.
• Mobile Unified Communication (UC) client, allows business customers
to link two phone numbers, their mobile and business lines, with their
mobile phone.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 233

 Dixio Multiplatform Lookup Technology

Organization: Semantix Group SL Design team:

In-house: Ferran Gatius, Product Manager; Jesús
Application: Dixio multiplatform lookup Liarte, Developer; Francisco Javier Hidalgo,
technology Developer; Laura Vilar, UI Designer and Joaquim
Headquarters: Barcelona, Spain Triginer, Quality Testing.
External Agencies: Blit Software and Vinco Mobile
www.semantix.com

OVERVIEW
Main function: Dixio is a multiplatform lookup technology that, together with
dictionaries and encyclopedias, provides readers with the most relevant reference
information possible without leaving the document that he is reading. Dixio works on
several platforms and in many applications, by accessing the Dixio API that connects
to the Semantix Cloud Knowledge Engine (CKE) or the Local Knowledge Engine (LKE)
to provide the best possible result.
• Dixio Desktop can be used on just about any desktop application on
a user’s computer system: documents, emails, PDFs, spreadsheets,
etc. The desktop tool is used by approximately 82,164 users.
• Dixio eReader Solution can be used on a variety of devices from
smartphones to touch screen tablet devices and is used by
approximately 10,000 users.
Why it’s unique: The knockout feature of this technology comes from the semantic
engine that drives the lookup results. This “capture & search” function reads the
entire context of a given word, analyzes the content and returns a context-sensitive
definition in a just a few seconds. This ensures that Dixio can provide the best, most
appropriate definition possible, each and every time. It is non-intrusive software that
remains silent until the user calls it and it works on the fly, rather than requiring the
user to launch it as a separate task. Dixio is able to identify the language of the text
in question, in order to provide definitions and/or translations.
The organization: Semantix, the company behind Dixio, develops tools to help
people communicate better. Its tools bring together people who speak different
languages, making their communication easier, either in their own language or
across different languages. Semantix expertise is drawn from its diverse staff
including: linguists, computer engineers, philologists, translators, lexicographers,
communication specialists and marketers, all working to improve the power of
communication amongst people of any language.
Design philosophy: While most dictionary apps rely on the age-old paradigm of a
paper dictionary where you “go” to look up a word, the Dixio technology is trying to
turn that idea on its head. Dixio gives the dictionary a whole new look and feel,
where the dictionary becomes more of a gadget than a book. It’s there when you
need it but it won’t get in your way when you don’t.

USERS
Dixio users are drawn from a broad range of people who use an array of electronic
devices (such as e-Readers, tablet, and smartphones.) The users are an eclectic

234 [email protected] Executive Summary

 group of individuals from educational centers, e-learning platforms and publishing
companies.

THE APPLICATIONS
On all platforms, Dixio acts like a very good restaurant waiter: there in full force
when you need him, but not interrupting your dining and conversation when you
don’t. Dixio is an ambient power, falling into the background in any one of these
three devices: the desktop, phone, or tablet. Hiding the app until needed enables
people to read, work, and engage with content without having to tune out the noise
of an overzealous dictionary app. Designers wholeheartedly took this approach, as
they believe, according to Laura Vilar, “[Users] are interested in reading the book,
not working with an application.” Astute designers are comfortable with this concept,
and have little ego about needing their app to be center stage.
Once the user is stumped on a word he just read and summons its definition, Dixio
snaps to attention. Calling the app is simple on all platforms: on the web, double
click; on desktop, control + right click; and on smartphones and tablets, long tap
and double tap. The features are advertised and described to the customer when he
begins using the service, and there is a small banner in the right rail of webpages
also announcing the service and features.
Upon completion of the small summon gesture, which works when directed at any
word on a page, a result dialog box appears containing a contextual definition
gleaned from more than 95 dictionaries, including Oxford, Vox, and Semantix
Bilingual.
Multiple Media
In each environment, the Dixio UI is simple, and consistent with other platforms to a
degree, but also exploits each platform’s offerings. According to Vilar, “As long as
each application is used for a different platform, we had to adapt it to each device.
Obviously if you're reading in a tablet it has to be something light that appears and
disappears easily.”
And Ferran Gatius, Product Manager, agrees, stating that designers needed to make
some tradeoffs and change navigation concepts which are understandable on the
desktop but not on other devices. For mobile, he notes a known design guideline, “It
doesn't make sense to have so many features on the tablet, like settings; we need to
make the navigation simple.”
Knowing each platform’s strengths and weaknesses and designing for those is part of
what makes a great transmedia strategy, as does defining which elements should
remain consistent no matter what the medium. In the case of Dixio the following
three items are those the team determined should be implemented consistently
across all platforms:
• Content: The word definitions are the same.
• Access: The “crown” area which appears on the top of the tabs, as well
as the tabs themselves.
• Visibility: The Dixio window will never appear over the word a user has
selected to define.
The semantic engine which is the power behind these apps’ success is not exactly
the same for all platforms, though this is on the product roadmap. The current plan

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 235

 is to move towards a server-based lookup function, as cloud computing becomes
more popular and stable. In the meantime, some features which are available online
are not available in the local versions, for example, only on the EReader Solution can
you click a tab to get the online definitions.
Dixio Desktop
The Dixio Desktop is more than your average dictionary, with easy design and
powerful features. In the search results, users may select their preferred dictionary
via the top horizontal tabs, which display not only the source name but also the
number of items on the respective results pages. With this, users get a sense of the
results before they click or read anything. They may keep the smart default selection
tab, All. In many apps like this where data is coming from multiple sources and the
user may actually care to know the source, the results are sorted in tabs by source.
But since it is quite possible the users won’t see, or may just not be inclined to click
the tabs, then having the All tab—and having it as the default—is essential. While
there is more information to scroll through in the All tab, there is a higher chance
that people will find the information there than on the other tabs. The drawback is
that, it's more work for users to look through the information categorized by source
on one long page. But, that’s what the companion UI tabs are for, and the two
interfaces provide an effective safety net for one another, and cover a wider range of
user scenarios.
The app also helps people to recall and learn definitions, as users may view the
previous word they looked up by clicking the arrow to the left of the search field.

236 [email protected] Executive Summary

 When using Dixio Desktop, users can look up a word's definition by pressing
control + click (or a customized key + click combination of their choosing.)

Dixio Desktop not only defines words, it also translates them. The semantic analysis
gives the most appropriate meaning according to the context. And because the
engine reads the whole article and has context for words, the translation can easily
help people find out what a random word in another language is. So if the user is
reading an article in English but there is a Spanish word in the text which he chooses
to translate, Dixio presents the definition of that word in English, the language most
of the article is written in, rather than in Spanish. This helpful feature makes content
more seamless, engaging, and understandable. Users may access translation
features via the Define and Translate button in the upper right.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 237

 A submenu by the search in Dixio Desktop allows people to define and translate
words. In this example content comes from Wikipedia, and can be visual and
thematic, for example regarding: medicine, law, sport, computing, etc.

For even more knowledge about the lookup in question, users may obtain their
results as images instead of text. Images are retrieved from Google, Flickr, or the
Semantix visuals gallery by selecting from the search sub menu in the upper right.

238 [email protected] Executive Summary

 On Dixio Desktop you may switch from text results to images.

The desktop application not only works on websites, but also with MS Word, PDF,
email, and spreadsheets, offering the same commands and features.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 239

 Dixio Desktop work on MS Word and PDF documents, as well as websites.

To make Dixio work best for each individual, the app allows people to see and make
changes to their account, change the language, and change preferences about
internet connectivity, as well as the dictionaries. Users may also add their own
customized glossaries. (An administration panel allows IT staff to keep this in check.)
These types of customization features increase ease of use and build loyalty.

240 [email protected] Executive Summary

 Users may edit preferences such as which dictionaries to search and when to
connect to online dictionaries over the web.

The Contents tab displays the Dixio Library of dictionaries that users may include in
their definitions. Users can choose which dictionaries to enable or disable, and
rearrange dictionaries so that definitions from their preferred sources appear first.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 241

 The Dixio Library, under the main menu, allows users to change their dictionary
preferences.

Web
Dixio Technology may also be installed on a website by the site’s owner so that
everybody who visits the site may use it, even if they don't have Dixio Desktop
installed locally.

242 [email protected] Executive Summary

 When installed on a website, users may look up any words on the page using
Dixio for Web by double clicking on any word.

Dixio eReader Solution

To start the eReader Solution application, users double tap or tap and hold the
device screen on the word they want to define. A simple dialog box appears, which
displays the definition and two tabs: the default Dixio tab, which uses only local
content, and the iDixio tab, which accesses online definitions on the cloud if an
internet connection is available. Local search is the default, which is important
because device storage is always an issue, and this helps keep the app quick and
responsive. But at any time, the user can switch to a wider range of definition
content through second tab.
The dialog also displays icons in the upper right to access secondary functions. For
example, the books convey which dictionaries the user is working with, the
magnifying glass opens a new dialog to allow manual search, and the X dismisses
the dialog box.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 243

 When a user looks up a word, the definition appears, along with tabs that allow
users to switch between local and online definitions. The dialog also displays
controls for secondary functions: the books convey which dictionaries the user
is working with, and the magnifying glass opens a search box.

When users tap the magnifying glass icon, a new dialog opens to perform a
manual search.

244 [email protected] Executive Summary

 The definition dialog box also displays the term a user searched for, the dictionary
searched, and controls at the bottom to view definitions of similar words and
phrases. Taking advantage of the swiping gesture available on tablets and smart
phones, users may swipe in a definition to display another definition from a different
source for the particular word.

Users may swipe within a definition to display more definitions for the particular
word from different dictionaries.

In addition to working with tablets and smartphones, eReader also works well with
electronic readers.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 245

 Dixio eReader also works seamlessly with electronic book readers.

Dixio vanishes as quickly as it appeared, when users click outside the dialog box, or
on the X in the upper right. But with an app this good and helpful, we are sure users
will call it back often.

DESIGN PROCESS
The Dixio technology originated as a desktop application called Gotitclear, which was
basically a dictionary app. In order to speed up the lookup process, the company
decided to move the technology to the Cloud.
The team has employed a number of research methods to inform their journey from
Gotitclear to Dixio and to improve Dixio over time. These include:
• Competitive analysis: The team started their research by analyzing
similar products available in the market and studying design guides,
prior to developing their own. This snapshot provided a good starting
point for the Dixio project. This competitive research included a
thorough look at the tablet space writ large, to see where tablets were
succeeding and where they were failing to stack up to eReader
Solutions. What they found was a clear demand based on how the
tablet apps performed as eReader Solutions.

246 [email protected] Executive Summary

 • Reading blogs: An important font of knowledge for the team was to
read blogs where users were expressing their opinions (and
complaints) about existing dictionary apps. In this forum the team was
able to learn a lot about the user experience flaws people were
encountering in their reading.
• Studying user patterns: One of the key metrics the company used
to guide their re-design from desktop to Web was to take a deep dive
into the user stats and let the user metrics and usage patterns guide
their decisions.
“We started measuring what the user was doing with our product,”
says Gatius. “And looking at that [data] we are constantly changing
small things to make it easier [to use] and eliminating functions that
are not being used.”
A good example of this type of change cycle was when the team
noticed that the data showed users were frequently accessing the
library within the application. The design team decided to move the
library to the main menu to give users quicker access.
• Expert review: The Dixio team commissioned an expert review of the
functionality, tools, and content from an outside agency.
• User trials: Users at a school were given a trial version of the Dixio
app and allowed to use it for a limited time. The students were
surveyed after using the tool, and results from their feedback were
rolled into the product.
• Observing users: An effective method the team has employed is to
take a group of users who are unfamiliar with the product and just sit
them down in front of it and say, “use it” and see what happens. And
this has produced some of the most surprising results. “This is
normally a process that always surprise you,” says Gatius, “because
when you are working on the same thing for a long time, you know all
the details behind the screen, and you don't see what users do, which
may be staring right at you on the screen. So by observing people that
have no experience with [Dixio], we get important information to
improve the product.”
One of the most surprising findings from these observation sessions is
that no matter how hard the Dixio team tries to push the envelope and
provide users with a robust contextual lookup via the “capture”
function (and no matter how much they try to educate the user about
the same), most of the users (approximately 60%) are still using the
tool for a very simple and traditional type-a-word-in-the-search-box
definition search.

Implementation
Designing the Application Engine
The application engine is written in C++ and there are cross-compiling tools that are
used for each platform. "We implement an upper adaptation layer that interfaces
with the devices," says UI Designer Laura Vilar. “For example, the Android

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 247

 application has a C++ core and is compiled for that processor, but we need to put an
adaptation layer on top that is able to communicate within all the Android systems."
This same approach will be applied across other devices and platforms such as
tablets and iOS. At the same time, the team is working on the online engine. This
has created some differences between the online engine and the local device engine,
which the team is working to resolve.
“Some of the features of the online engine are not available on the local engine,”
says Vilar. “So now we are starting the conversion process to have exactly the same
on both platforms. We are working to transfer the benefits of the C++ technology on
the devices over to the server engine.”
Moving from a Device Model to a Server Model
Next on the company’s roadmap is an effort toward moving the application
processing from a device model to a server model; for example, by allowing the local
engine that is supplied on the eReader Solution to also call to the Dixio server for
supplemental content.
One of the problems of this model is of course the space limitation on mobile
devices. The application accesses large dictionaries and that creates a storage space
problem on mobile devices. So the solution is to offer complementary information
that sits on the server.
“So [mobile devices] would have the ability to look for online resources,” says
Gatius. “You would get the local immediately, but if you want the online ones, you
would click the Online tab, and have access to more dictionaries online than you
have locally."
Moving toward an online model means using different technology. The online engine
was developed using standard Web languages and standard server languages.
“We’ve determined that we could improve the server engine a lot by using low-level
languages like C++, and we would benefit as well from having common code,
common functions, and common behavior for local and online engines,” says Gatius.
“So our plan is to, as much as possible, synchronize the modules running on our
servers with the local device modules—while taking into account the fact that some
language processors can be too resource intensive for local devices. Maintenance will
be easier if the engines are fully synchronized.”

248 [email protected] Executive Summary

 PRODUCT TIMELINE
2006
• Gotitclear (predecessor to Dixio) released. This is a
desktop dictionary/look-up app. The goal of this app
was strictly just to sell local dictionaries.
• Gotitclear application and dictionaries are able to
detect inflected forms and idioms.
Q3 2007
• Gotitclear product renamed Dixio and moved online
to make app lighter and faster and remove
processing/space limitations of desktop
environment.
• In the new tool (www.dixio.com) users can look up
words, create accounts and manage their own
dictionaries.
• Programming begins on Dixio for Web (DFW), a web
app for use on any website.
• Both products work thanks to the linguistic server,
the CKE (Cloud Knowledge Engine) that is released
in 2007.
Q4 2007
• Dixio online beta (www.dixio.com /DFW) released.
• Development begins on Dixio Lite, a thin client for
Windows.
Q1 2008
• Dixio Lite released (thin client Windows PCs.) This
product only allowed capturing and displaying
definitions on the desktop.
• Preference management and dictionary management
are now handled online.
Q2 2008
• Development begins on Dixio Desktop for Win, Mac
OS and Linux.
• Dixio desktop is a multiplatform online desktop
application with library management and Cloud
storage.
Q4 2008
• Dixio Desktop v.1.0. for Windows released.
Q1 2009
• Dixio Desktop v.1.0. for Mac OS and Linux released.
Q1 2010
• Redesign of Dixio Desktop (Windows) v2.0 begins.
• Throughout 2010, the Cloud intelligence of the CKE
is improved:
o The app identifies the language of the text
the user is reading.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 249

 o The app makes a linguistic analysis of the
captured text.
o The app indexes online encyclopedic content
(like Wikipedia), which is continuously
updated.
Q2 2010
• Dixio Desktop v2.0. released
• Dixio.com is no longer available; instead a new
desktop version is released with new design and
usability rules. The main functionality is the same
but the GUI and engine changes significantly.
• Team defines metrics to analyze user behavior (such
as consulting the library, capturing a word, or using
the searchbox for manual lookup.)
Q2 2011
• Redesign of the Dixio Desktop v.3.0. for Windows.
• User experience testing and modifications to
improve usability.
• Develop and implement Dixio for e-readers (Dixio
eReader Solution.)
• Dixio is installed in devices such as Movistar ebook
bq, Fnacbook and Tagus.
Q3 2011
• Development begins on Dixio for tablets, beginning
with the Android OS.
Q1 2012
• Launch of new Dixio Desktop 3.0.
• Launch of Dixio for Android.
Plans for 2012 and
• Launch of Dixio for iOS.
beyond
• Changes to improve the CKE (Cloud Knowledge
Engine) intelligence.
• Expanded language options, including French and
German.

LESSONS LEARNED
Listen to users. The primary lesson the Dixio team has learned through their design
efforts is the importance of getting actual information from users, whether it be
analytics about their behavior or direct conversations. “Listen to the users and try to
give them what they want and what they need,” says Vilar.

250 [email protected] Executive Summary

 TEAM

DIXIO PROJECT TEAM ROLES

Role Responsibilities
Product Manager Determine features and functionality
Ferran Gatius

Developers Implement technology solution and perform bug

Tomas Couso (agency) fixing
Jesús Liarte
Francisco Javier Hidalgo

Blit Software
(Agency for Desktop app)
UI Designer Design interface and graphical elements
Laura Vilar

Quality testing QA testing

Joaquim Triginer

Pictured (left to right): Jesús Liarte (Developer), Laura Vilar (User Interface
Designer), Ferran Gatius (Product Manager), Xavier Hidalgo (Technical Product
Manager), Joaquim Triginer (QA Testing.)

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 251

 Pictured (left to right): David Llort, Carlos Martin, Daniel Ricart, Maria
Monterde, Montserrat Civit, Laura Vilar, Francisco Javier Raya, Ferran Gatius,
Christian Cicceri, Laia Pacheco, Javier Miguel Sastre, José Córdoba, Joaquim
Triginer, Sergi Fernàndez, Marc Centelles, Xavier Hidalgo, Jesus Liarte, Antonio
Riu and Fernando Ferri.

252 [email protected] Executive Summary

 Highlight

Organization: Cohdoo, LLC Design team:

Application: Record and Annotate Audio • Ryan Russell, Designer
• Eric Wuehler, Developer
Headquarters: Portland, OR
• David Stewart, Design Consultant
www.cohdoo.com

WHY THEY WERE CHOSEN

Highlight was selected as an outstanding example of a multiplatform experience
because of several elements of its user experience:
• Simplicity: By maintaining a tight focus on the core functionality, the
application omits clutter and confusion, and delivers an intuitive experience.
• Flexibility: Optimized for phone, tablet and personal computer, Highlight
offers a feature set and interface tailored to fit each device and context of
use.
• Purpose: As our gadgets proliferate, we are overwhelmed with tools for
creating content but far too few of these help us manage and consume
content in a meaningful way. Highlight addresses this need with an elegant
system for extracting meaning from content and sharing it with others.

OVERVIEW
Main function: Highlight lets users with iOS devices make audio recordings, mark
interesting points, and share recordings with others.
Why it’s unique: Quite a few audio recorders are available for the iPhone -
including the standard recorder provided by Apple; Highlight distinguishes itself with
an incredibly intuitive design. The interface presents both basic and advanced
features in just a few self-explanatory screens.
The organization: Cohdoo, LLC is a startup created by Russell and Wuehler. Since
Cohdoo is a side project for both men, they constrain their project scope by aiming
to design, code, and release each product within 90 days. This timeframe limitation
helps ensure that the project stays manageable and the end product isn't rendered
obsolete by changes in the platform or market.
Design philosophy: A simple principle guided the design of Highlight: a smaller
focus leads to a bigger impact. Russell explains, "By focusing on one thing you can
have a much bigger impact. There’s a quote I really like, that goes something like,
‘Your ability to have an impact is directly related to your willingness to narrow focus.’
We wanted to have an impact with our application and delight people, so we knew
we’re going to have to put all our effort on one thing, because if we try to do a lot
it’s just going to be mediocre."

WHO ARE THE USERS?

Highlight's core user group consists of people who conduct in-person interviews for
ethnographic or design research. Russell explains that he knew from personal
experience how time-consuming it is to analyze audio recordings and extract key
insights.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 253

 "Synthesis and processing of recordings takes so much time. If you’re in the field an
hour, that could mean two, three or more hours of processing. [People] who work in
design firms say, the client pays for us to do all this research and they want us to
move instantly into design work. It’s like, synthesis is just a box to be checked; they
want it done [instantly]… What good is 100 hours of fieldwork if it never gets used?"
says Russell.
The Highlight designers aimed to meet the specific needs of user experience
researchers, calculating that doing so would also lead to a product that was good (if
not perfect) for other types of users, such as:
• Journalists
• Lawyers
• Students and Professors.
Russell says, "If we can delight our primary user — designers — we hit a whole
secondary ring around them that will be pretty happy too: anyone who does audio
recordings. It may not fulfill exactly their need, but it will be pretty close."

After Highlight was launched, its creators found that it was especially valuable to a
particular group that they had not previously considered: students taking courses in
their second language. "Highlight allows them to focus more on what’s going on in
the classroom and not have to struggle so much with what the professor is saying,
because they can go back to it later,” says Russell.
THE APPLICATION
Highlight lets users create audio recordings using an iPhone, iPod Touch or iPad, and
offers
• a simple, intuitive interface
• the ability to easily share the recording with others
• the ability to mark particular points of interest within a recording.
Making an audio recording seems like a straightforward task, but many of the
applications targeted at this audience are surprisingly complex, offering little visibility
into how files are created, stored and managed. Highlight stands out from the pack
because it features incredibly simple controls for starting, stopping, editing and
sharing recordings. Russell explains, "What makes Highlight special is its simplicity.
We worked very hard to remove every unnecessary element and feature from the
application until we got Highlight down to its pure essence. There are only five
screens in the entire application, and our users consistently comment on how easy
and intuitive Highlight is to operate.”

Highlight’s users need to share their recordings with others, but just sending an
audio file to colleagues is not enough. Many research interviews are one to two
hours, but few designers have time to listen to hours and hours of audio recordings.
Highlight lets the field researcher mark particular points of interest within a
recording; listeners can then skip straight to these points, vastly reducing the
amount of time needed to consume the content.

In addition to marking highlight points while making a recording, users can make a
few modifications to the recording file after it has been completed:
• Name the recording
• Add a picture

254 [email protected] Executive Summary

 • Name the highlights
• Add more highlights
• Add a note
• Rate the recording.

Highlight lets users create audio recordings, mark points of interest within the
recordings, and share them with others.

INNOVATION
As technology advances and application users work on multiple devices, it makes
sense to expand an application experience to reach them on new platforms. But it is
essential to approach new platforms in a thoughtful way, to ensure that the features
and offerings are appropriate to the platform.
The ability to share recordings with others is essential for Highlight’s core user
group. The smartphone is an excellent tool for field researchers who need to create
recordings without hauling around large devices or accessories. But the personal
nature of such mobile devices makes them less than ideal when it comes to making
content available to more than multiple users.
Since the standard audio players can’t display the unique metadata (such as point-
of-interest markers) in Highlight recording files, Highlight’s designers created
Highlight Player, a companion application for consuming Highlight recording files on a
desktop or laptop computer.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 255

 The feature set for both Highlight, the mobile application, and Highlight Player, the
desktop application are tailored to address the users’ tasks while keeping the device
capabilities and limitations in mind.
For example, early plans for Highlight included features such as the ability to tag
recordings with different labels. Eventually the designers realized that while the
phone was great for capturing recordings, the ultimate purpose of doing research —
to influence the team’s design decisions — was better served by encouraging users
to get the recordings off their phones and onto another device or platform where
more people could access the content. As Russell explains, “We’re not against people
storing recordings on the phone but we really see the value when you get it off your
phone. When you start to proliferate it through your organization, your team, the
community — then it becomes much more powerful. …So instead of promoting the
ability to hoard on your device, we want to promote the ability to push it off your
device.”
Likewise, the Highlight Player could have included many additional features, such as
the ability to edit and manipulate the content; however in the interest of keeping a
consistent experience and releasing this essential tool in a timely manner, the
designers ended up going with a very simple tool. As Wuehler explains, “We needed
something to let people give their teammates the benefit of the highlights. Because
of that we had to build a player. …We kept the functionality focused…so people can
play back what someone else created. This let us limit the scope enough to quickly
get it to market. We thought about shipping with some basic editing capability, but
we ended up pulling it back in order to get the playing capability out the door.”

256 [email protected] Executive Summary

 Highlight encourages users to transfer recordings out of the application and
onto another platform via email or using Dropbox (a popular third-party file
sharing service.)

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 257

 The Highlight Player is a simple tool that lets people listen to highlight
recordings and skip to specific highlight points on a laptop or desktop computer.

Highlight’s also takes a proactive approach to saving content. Russell explains,

“There’s really no concept of save anymore. If I write on a piece of paper, it doesn’t
magically get up and vanish. If you do work it’s always there. You can get rid of it
later if you want to, but you don’t have to explicitly tell me to save it. It’s valuable
because you spent time making it.”
The first version of Highlight gave users who paused a recording the option to
resume, save, or cancel the recording. To minimize accidental deletions, the ‘cancel’
button was followed by a confirmation dialog. “It just felt awkward; I hate having to
ask people, ‘Do you really want to cancel?’” says Russell. When working on the
second version, Highlight’s designers revisited the ‘cancel’ scenario and realized that
there really wasn’t a need for a cancel button at all — because the only time it saw
significant use was while testing the application or demonstrating its functions. The
few cases when real users needed to discard a recording were easily addressed by
the ‘delete’ functionality available elsewhere in the application.
Acting on this principle of ‘everything the user creates is valuable,’ Highlight Version
2 now offers a single button labeled “Done,” rather than a choice to either save or
cancel a recording.

258 [email protected] Executive Summary

 Left: Highlight saves every recording; the secondary action of deleting a
recording can be accomplished through the home screen. Right: A previous
version of Highlight gave users the option to either save or cancel a recording.

CONSISTENCY
The Highlight Version 2 recording application is designed to work on both the iPhone,
iPod Touch, and the iPad. While the layout changes to take advantage of the iPad’s
larger screen size, the content and functionality remains the same, creating a
consistent user experience regardless of the device type.
On the iPhone, the home screen displays all the recordings in a vertical list, with
thumbnail images showing the picture associated with each recording.
The iPad layout uses the same picture as an icon for the recording, in a horizontal
array. The delete control is also placed differently in the iPad layout. On the iPhone it
appears on the top left corner, but on the iPad layout many users would be likely to
overlook a control placed so far away from everything else; instead, it is moved to
the top right corner to form a sort of ‘utility controls’ area of the screen. Eyetracking
research shows that people tend to look in the top right corner for such controls, so
this placement conforms to user expectations. (The placement works on the iPhone
version because the screen is so much smaller that everything is closer together and
easily noticeable no matter where you are looking.)

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 259

 The Highlight is designed to work on both iPhone and iPad. The iPhone home
screen layout displays recent recordings in a vertical list.

260 [email protected] Executive Summary

 The iPad home screen layout differs slightly from the iPhone version, but
maintains the same content and features for a consistent experience.

Similarly, the Highlight Player application expands on the Highlight Recorder

interface but maintains a consistent experience. While the player interface adds
visual indicators of highlighted points on the timeline and a details panel on the right
side, both interfaces use the same simple playback controls and list of highlighted
points to let users jump to important moments in the recording.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 261

 The Highlight iPhone application allows users to play back recordings from the
Details screen, using simple controls and a list of highlighted points in the
recording.

262 [email protected] Executive Summary

 The Highlight Player maintains the same simple interface as the mobile and
tablet application: simple playback controls with just a bit of additional detail,
such as the timeline markers indicating where highlights appear.

FLEXIBILITY
Applications should strive to accommodate the user’s task, rather than forcing people
to change their behavior in order to use the application. Highlight includes several
features specifically tailored to the interview and recording process to make the
researcher’s task easier.
If you’re interviewing someone it’s probably because you need information they
have, but you don’t. That is, you don’t already know what they’re going to say. This
obvious circumstance has important implications for the process of marking points in
a recording, because many people won’t recognize the points they want to mark until
after they have occurred.
Highlight makes it easy to mark these moments that have just passed with an offset
feature, which automatically adjusts the marker time to be 2, 5 or 10 seconds in the
past, depending on your settings. (You can also choose ‘no offset’ to have markers
applied in real time.)
Instead of a stop button, Highlight uses pause as the primary control for ending a
recording. Users love this simple accommodation, absent from many other recording
applications, which makes it easy to omit dead time (such as bathroom breaks) from
the final recording without stopping and creating several different files.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 263

 The offset setting lets users place markers a few seconds before they actually
tap the highlight button, more accurately identifying the beginning of important
points.

Researchers can take a picture of their subject and associate it with the Highlight
recording, which helps to ‘put a face to a name’ for people consuming the research
later on. It also allows users to browse recordings by either name or thumbnail
image.

264 [email protected] Executive Summary

 The Pause feature allows users easily to omit inactive periods from a recording,
the quickly resume when the interview begins again.

INFORMATION DISPLAY
Although Highlight deals with a relatively small volume of information, the screen
size limitations of mobile devices make it essential to convey information clearly in a
very small amount of space.
The Highlight home screen has evolved through several iterations, each one
increasing the utility of the screen by providing access to relevant data. In early
sketches, the home screen was simply a menu listing the main tasks, such as
starting a new recording. Existing recordings were listed on a separate screen. As the
design progressed, the home screen evolved from a simple task hub into a more
information-rich screen displaying recent recordings, as Russell explains, “Initially we
thought, ‘We’ll have a button that goes to all of them and you can pick.’ Well really,
people want to get to the ones that they just did, so how do we make that easy and
straightforward? …We’re trying to help the user get to the information that they want
quickly, without having to dig through the interface.”
The process of optimizing Highlight for the iPad actually led to further revisions of
the iPhone home screen. Showing only the three most recent recordings left a great
deal of empty space on the iPad screen, which sparked the idea to show all
recordings rather than just the most recent. This idea was then applied to the iPhone
design: Now the iPhone home screen also shows a full list of all recordings as a
vertical list. “Now instead of the last three recordings, I can show all of them, and

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 265

 we’re able to get rid of the all recordings screen and provide an easier to use
program, but it was really created because of that iPad work,” explains Russell.

Early sketches for the home screen included just a menu of pages and did not
display any actual recording titles.

266 [email protected] Executive Summary

 Left: The first version of Highlight for the iPhone had six screens, including one
screen to display all recordings. Right: Version 2 of the application displays all
the recordings on the home screen, rather than on a separate page.

In an early prototype of Highlight, ratings were indicated with blue stars. If a user
hadn’t yet rated the recording, it displayed five white stars. So the star was both the
method for assigning a rating and the method for displaying a previously assigned
rating, with only the color difference to distinguish between the two states.
Because stars are so frequently used to display ratings, many users guessed at the
display function and assumed that the five white indicated a five star rating, never
guessing that they were in fact an input method.
To eliminate this misunderstanding, the interface was changed to use different
elements to signal the input and output of the rating: before a rating has been
assigned, five small dots appear in that row, and these dots change into stars when
tapped, to display the assigned rating. Separating the two functions into two
different visual indicators is much more effective at communicating their purpose
than relying on just one symbol to fulfill two purposes.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 267

 Left: In an early prototype, users did not realize that the white color indicated
that there was no rating. Right: When tapped, the stars turned blue to indicate
the rating.

268 [email protected] Executive Summary

 Left: The ratings area now initially displays five small dots. Right: When
tapped, the dots change into stars to display the assigned rating.

INTERFACE ELEMENTS
The most noticeable elements of the Highlight interface are the enormous Start and
Mark Highlight buttons; each of these takes up more than half the screen. This
approach is unusual among mobile applications, where screen real estate is at a
premium; but the emphasis is extremely effective at optimizing the screen for the
single key task users need to accomplish.
Because Highlight is intended to assist users while they are primarily engaged with
another activity, conducting an interview, the design acknowledges that the users’
attention is not primarily focused on the screen, especially during the actual
recording process. Russell explains, “The highlight button that allows users to
gracefully touch it while keeping eye contact with their participants, an important
part of building rapport."
Early sketches for the New Recording screen devoted more space to other
information and elements, such as the elapsed time indicator and a list of recently
created highlights. But eventually the designers realized that those items weren’t
really needed while making a recording, and that the user’s activity would be better
served by placing even more emphasis on the Mark Highlight and Pause buttons, the
two primary actions that every user will need to perform quickly.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 269

 The Mark Highlight and Start buttons dominate their respective screens,
occupying more than half of the screen real estate. The buttons are easy for
interviewers to click.

270 [email protected] Executive Summary

 Early sketches for the recording screen devoted less space to the Highlight
button and more space to other items such as the elapsed time and data about
recently created highlights.

The playback area of the Details screen includes controls to play the recording, skip
back ten seconds, and create a new highlight. Highlight users report that skipping
backward is one of their favorite features, and both that and the triangle control to
begin play are easy to recognize. The third element, the light bulb symbol, however,
presented some difficulties. It is used to represent the action of marking a highlight,
but some users did not notice or understand this element. The light bulb symbol is
used in several areas of the Highlight application, sometimes to represent an action,
and sometimes just as a branding element. The dual meaning of this element and its
placement on the far right of the Details screen, away from the timeline, combined
to make this element appear to be purely decorative rather than functional. To
address this issue, Highlight Version 2 has adjusted the placement of the three
playback controls so that they appear close together, all directly over the timeline, to
make it obvious that they are all related.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 271

 Left: In Highlight Version 1.0, the control for creating a new highlight appeared
on the far right of the Details screen, directly over the end time display. Right:
Now in Version 2.0, the control has been moved in closer to the other playback
controls, to make its function more obvious.

Although all the elements on the Details screen can be edited by tapping and holding
them, this screen also includes an explicit Edit button in the top right corner. Users
will probably learn about the advanced selection options after they gain some
experience with the application, but the visible signal of a button is extremely helpful
for new users. This is especially true for touch interfaces: with a cursor, users are
already interacting with the interface and moving the mouse around, often
automatically pointing at the element they’re interested in, which creates the
opportunity to communicate with hover effects. With a touch interface, however,
there is generally no contact between the user and the interface until the user has
already decided what to tap, so the functionality must be clearly communicated by
the basic interface.
Russell and Wuehler chose to label the button Edit to help keep the interface simple
enough to use immediately, without any instructions or help files. As Wuehler
explains, they felt that “if it was complicated enough to need a help file, it was
probably too complicated. …That was definitely one of our goals: to make it obvious
enough that somebody can just pick it up and use it? …That’s why we chose the word
‘edit’ instead of a pencil or something: because that’s obvious. If we’re not going to
include a help file, we have to make it self-explanatory.”

272 [email protected] Executive Summary

 Left: Users can tap and hold each element (such as the recording title) to edit
it individually; Right: tapping the Edit button changes the entire screen into an
editing mode.

A final interesting aspect of the Highlight interface is an element found on many

mobile applications, but not in Highlight: a bottom navigation bar. During the early
planning stages of the design, Russell and Wuehler considered including persistent
menu bar as well as more navigation controls between different screens of the
application. Ultimately, they decided against these options in order to keep a strong
focus on the primary task of each screen. For example, when creating recording, the
entire screen is devoted to that task and the user is expected to be immersed in the
process of managing the interview and marking highlights. Because the task is
resource-intensive, it’s highly unlikely that users would want to multi-task and do
other activities within the app (such as edit a different recording.) So a persistent
menu bar giving access to other parts of the application would complicate the
recording screen and add the potential for errors, while offering very little benefit.
Accordingly, Highlight’s designers went with an immersive navigation experience
where users drill up and down from a home screen rather than a more flexible
persistent menu bar which would have given users access from any page, to any
page.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 273

 DESIGN PROCESS

Usability Testing
Highlight’s creators subjected each stage of the design to quick, guerilla-style
usability testing. Friends and family served as test subjects, both for convenience
and to make sure that the application would be easy enough for anyone to pick up
and use immediately, regardless of technical ability. Wuehler explains, “I would
make a build and hand it to my twelve-year-old and say, start recording and mark
two highlights. Once she had seen it a few times she could do it so I had to switch to
my wife and my other kids...it was kind of a poor man’s usability test.”
Findings from these tests were immediately incorporated into the design. This
lightweight methodology allowed the designers to explore a variety of design options
in a short amount of time. “We could experiment, try stuff, get feedback, and iterate
in these really tiny loops. So we got to try tons of things in a short period of time to
get the winner, versus doing stuff that was very heavyweight and only getting to try
two or three options because the testing was expensive, or done at the end but
nothing at the start,” says Russell.

Beta Testing
With any application, and especially with mobile applications, actually using the
program reveals issues that don’t emerge in simulated settings. For example, the
possibility of getting a call that interrupts your recording, or having files that are too
large for email. Local designers helped beta test Highlight by actually using it for field
research, which yielded valuable design insights. Russell explains, "One of the
biggest things they said is they really want to get [the recording] back to their
broader team... initially the only way you could share the recording was to email it to
someone. People said, ‘We want to share it in other ways, we can put big files in
Dropbox but the email server can’t handle something too big.’ That was something
we learned in research, and were able to add."

Implementation
While Russell sketched iterations of the front end design, Wuehler began developing
the back end functionality. This simultaneous approach required significant up-front
communication and ongoing check-ins between the two, but ended up being far
more efficient and productive that traditional linear approaches. “It worked well
because as we progressed we had both sketches we could play with and stuff we
could try that was actually in code. The two together were really powerful, because
we learned a lot in development that informed the sketching,” explains Russell.
Another benefit was that because the development was initially focused on just the
back end functionality, the front end user interface code was left unfinished until
later in the process, after sketching and testing had refined the design into a better
user experience. From the development perspective, Wuehler felt that this approach
improved both the design and the code, "Really tight collaboration between the
design and the development allowed design to understand what the development
constraints are, and the flip side, having design push me to think a little bit
differently about how I might be able to do things."
Wuehler’s early development efforts focused on created a ‘black and white’ functional
prototype, built mostly with the standard widgets in the iPhone toolkit. This

274 [email protected] Executive Summary

 prototype actually helped define the structure of the user experience, because it
illustrated the basic platform behaviors shared across many applications that would
be familiar to platform users and easy to implement. While the final application
doesn’t consist solely of these standard elements, it’s helpful to know the default
way of doing something that will be easiest to implement and easiest for users to
understand. “Then you can decide, is the tradeoff worth it, having to teach your
users something new as opposed to trying to find something that they’re more
comfortable with,” explains Wuehler.
The process of creating a simple functional prototype and then integrating it with the
polished user interface design makes for an efficient and graceful process, as
Wuehler explains, “With design, you can imagine anything… With development
you’re a little bit more restricted with what you can do easily or quickly. By agreeing
on the basic functionality up front, then as a developer I know I can [build it] in
time. So this black and white design and feature agreement… gets us 80% of the
way there. As we start to merge the black and white with the design we can finish off
that last 20% of marrying the features to the design.”

Left: The ‘black and white’ functional prototype of the Highlight Home screen.
Right: The functional prototype of the New Recording screen.

The close collaboration between design and developed saved time, both by helping to
quickly determine which design alternatives would be most practical to build, and by
solving the major usability issues while the interface design was still in the form of
sketches rather than code. Russell explains, “We really had very little rework. In
many projects...a bunch of stuff gets coded, you do a big usability test halfway
through the project and [find problems.] Now the cost [to fix it] is huge because
we’re almost done. Our rework… was pushed to the lowest-cost step of the project.
By collaborating with sketches intensely, there was more rework on sketching than

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 275

 there was on Photoshop, and more rework on Photoshop than there was in the code.
We went through quite a few iterations but it was very inexpensive.”

Long-term planning
Highlight’s creators also strive to keep future product plans in mind during the design
process. ‘We try to design out a couple of features ahead…For example we know
there’s a couple of things that we won’t have time to build into Highlight Version 2.
But we’re designing out what those should be so we can make sure that as we’re
developing it now we don’t box ourselves into a corner,” explains Russell.
This approach applies to both specific features and to entire applications, and this
philosophy of long-range planning and reuse has worked well. For example, a
product called the Highlight Converter used to be offered as a way of transforming
Highlight files into several different audio formats. The Converter was a less than
ideal solution as it did not preserve any of the highlight metadata, but Wuehler was
able to re-use the core basis of the Converter when he built Highlight Player. Russell
and Wuehler are now considering creating a web service for sharing Highlight
recordings, and plan to re-use the code from the Highlight Player. “Because of
limited development resources, we have to figure out how to achieve multiple things
with one product. We think about not just features but technology use one or two
products ahead, and how, in what we’re building today, we can re-use some of those
guts tomorrow,” explains Russell.

LESSONS LEARNED
Russell and Wuehler shared a few lessons learned from this design effort:
• Get your design onto the device ASAP. Interface elements may look big
enough on a monitor, but it’s not until you actually see them on a small
device that you understand how easy or difficulty they will be to manipulate.
"Regardless of how you’re rendering the design, you’ve got to get it on the
device really quickly. Even just saving it as an image file and putting it as a
photo on your phone. On a monitor…everything looks big enough to touch,
but as soon as you put it on a device, things are too packed together, they’re
too tight and not enough room for my fingers in there. So now I render in
Photoshop and copy it over to my iPhone or iPad to quickly see what it looks
like and if the spacing’s ok,” says Russell.
• Don't try to be clever. "If we’re trying to be clever it usually means we
don’t have the best solution, and we need to keep working. Because clever
usually equals ‘hard to figure out.’ It’s only clever to the person who came up
with it,” says Russell.
• Fight Feature creep. “With Highlight, we’ve been able to prove that when
you focus on figuring out what the heart of the application is… you can create
something that’s really magical and that delights people. There’s always a
push for more features, so it was fun to build something focused and spend
our time removing, not adding… We can be better than our competition by
doing one thing well, even though they do 10 more things than us. There’s a
creative director at a local design firm who uses Highlight as an example for
his clients. When their client wants a million features, they show them
Highlight and say, ‘Look how easy it is to use, look how delightful it is. We can
make your app like this.’ It’s exciting for us to have someone using it as an
example of that philosophy,” says Russell.

276 [email protected] Executive Summary

 TEAM
Highlight was created by Cohdoo, a small Portland, Oregon startup composed of:
• Ryan Russell, Designer
• Eric Wuehler, Developer

"We are focused on helping individuals bring their ideas to life. As UX professionals
we had a need for some additional tools to help us get our job done. We made a few
tools for ourselves, others liked them too, so we decided to go into business," says
Russell.
“The right gear in your pack makes all the difference. It allows you to be agile and
ready for what’s ahead. At Cohdoo we believe great gear does what you need it to
without getting in your way. It’s not about the gear, but what the gear allows you to
do."
The company name is taken from “Coh doo,” words that rang throughout Russell's
home as his little boy. This small phrase was the beginning of a more complex
declaration, “I can do that.” In those first words an experience was built, an idea
expressed: Cohdoo’s goal is to create products that help grow others' ideas from
infancy to maturity.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 277

 Mobile Unified Communications (UC) Mobile Client
(Verizon)

Organization: Verizon Wireless Design team:

Application: The application was designed in-house at Verizon
The Mobile Unified Communication (UC) client Wireless in Waltham, MA. It was designed internally
allows business customers to use two different by a cross-functional team drawn from members of
phone numbers, their mobile and business lines, the company’s Device Marketing User Interface
on a single mobile device. Design group, Product Management and Product
Services Integration teams.
Headquarters: Basking Ridge, NJ (USA)
Members: Linda Borghesani, Manager, Device
Marketing; Samantha Kalita, Consultant, Device
Marketing; Graham Sysko, Consultant, Device
Marketing; Kathleen Murphy, Manager, Product
Development; Victor Alba, Manager, Product
Development; Yeharn Hwang, Distinguished Technical
Staff, Technology and James Elter, Manager,
Technology.

OVERVIEW
Main Function: The Mobile Unified Communication (UC) client allows business
customers to link two phone numbers, their mobile and business lines, with their
mobile phone. It runs on Verizon Wireless Android Smartphones and enables
customers to use one mobile phone both at their desk and on-the-go to make and
receive personal and business calls.
Why It’s Unique: Simply put, this application is unique because it lets the user be
connected anywhere. No matter what kind of call a user needs to take - business or
personal - he can receive it on a single mobile device, on his mobile phone,
seamlessly. Additionally, when on-the-go, users can place business calls and select
whether their caller ID appears from their business or mobile phone number. Plus,
this app affords them business capabilities they would have on a desk phone such as
merging and switching between calls, transferring calls, and creating up to a four-
way conference call while mobile.
The Organization: Verizon Wireless is the largest mobile communications network
operator in the United States. It had 108.7 million subscribers as of the end of 2011,
the largest number of any wireless service provider in the United States.
Design Philosophy: The genius behind this application is how effectively the design
team has put a simplified face on top of complex technical architecture. Though
there are a lot of complicated things going on behind the scenes, the user is exposed
only to a coherent, smooth interaction with quick access to what he needs, when he
needs it.

USERS
The primary audience for this application casts a wide net. It includes any business
customer who has an Avaya, Cisco, or BroadSoft PBX in their office and has the need
for employees to be mobile. This means users are primarily (though not exclusively)
employees of larger organizations, which are more likely to have such PBX systems.

278 [email protected] Executive Summary

 THE APPLICATION
The Mobile UC client balances the traditional office phone features with the
convenience of a mobile phone. The UI is intuitive, simple, and covers a variety of
tasks as well as user scenarios, all making a great user experience.
The first interaction with the Mobile UC client, the Device Setup page, helps users get
their app set up in the way they need it. A wizard-like process leads people through
the set-up. And while a wizard process can be overly constraining at times, designers
in this case avoided this by planning for various user scenarios to make this process
helpful and productive for all. These scenarios include users who:
• Have a hardware dock to use with the app and their mobile phone: A
message on the Device Setup page suggests that users with a dock
should dock the device to assist with set up. The docking station
enables users to automatically configure their service after they have
been provisioned for the service by their PBX administrator. (A
separate PC-accessible administrator configuration tool is provided to
PBX administrators for provisioning users.)
• Have an email from their system administrator with configuration
information: A message on the Device Setup page suggests that users
with an email from a system administrator select the Import button.
• Desire or need to set up the app for himself: The user may choose the
Manual setup button on the Device Setup page.
• Don’t have time or cannot set up the system now: The Set up later
button defers the configuration steps.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 279

 A wizard walks the user through the setup process, during which he may import
a configuration file, manually set up, or dock the phone in order to easily
configure the service.

The docking station was designed to replace existing desk phones. It is specifically
designed to have a universal cradle, so that different devices can fit into the same
type of dock, even though each device may have its micro-USB connection in a
different place. Additionally, it goes the extra mile to connect to mobile phone with
cases on them: No need to make your mobile naked to use this dock.
When the mobile phone is docked, Mobile UC turns a regular mobile phone into a
powerhouse desk phone, as it supports the following functionality: corporate
directory lookup, multiple-line and shared-line access, transfer, conference, “do not
disturb,” mute, hold, voicemail, speed dial, and call history of incoming and outgoing
calls. The docking station also charges the phone battery, and the phone can be used
for all other smartphone functionality even while it is docked.
Mobile UC also allows users to keep the current call active while they either dock or
undock the mobile phone, so users don’t need to pause or be disconnected when
they need to leave the office or arrive and want to dock.

280 [email protected] Executive Summary

 An accessory, the universal docking station supports the Mobile UC application
while a user is sitting at his desk. There is a universal connector to support
many devices in any desired screen orientation.

When the device is docked, the Mobile UC home screen allows users to accomplish
many tasks, such as: call, check voicemail and call logs, and retrieve contacts and
favorites, including any smartphone functionality. Designers chose to display each
command as its own button in a flat display rather than consolidating in menus,
drop-downs, or panels. Since there is only a small number of frequently used
commands, the display accommodates them well and is a smart choice. The screen
shows the following large buttons for easy access: Phone, Voicemail, Call Logs,
Favorites, Contacts, Applications, Settings, Add, and Undock. The user may
customize and reorganize these buttons. Buttons show recognizable and recallable
icons along with clear text.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 281

 The Home Screen for the Mobile UC application when docked displays the
following default buttons: Phone, Voicemail, Call Logs, Favorites, Contacts,
Applications, Settings, and Add. The user may customize and reorganize these
eight shortcut buttons.

But not all important commands are presented at that flat top-level. The designers
made decisions at each step of the way about the most expected and helpful places
to put each command in the call workflows. For example, in the case of making a
call, designers chose to use progressive disclosure—hiding some commands until the
user is ready for them. Once the user selects to make a call he is then presented
with the options to make a Business call or a Mobile one. Designers could have
placed these commands at the home level but it would have made that page too
busy, and potentially confusing to offer types of calls at that level.
These buttons are color coded so users may easily pick them out, and the colors are
carried through the UI. If you are on a business call it is indicated in blue and a
mobile call is indicated in green. The call type also displays in the top status area
above the buttons.
Users may also add customized widgets their device’s OS home screen, so that when
the phone is not docked they have a shortcut to make calls of a specific type. The
widgets can be customized to Business, Mobile, or Prompt (which prompts users to
choose a call type every time they place a call.) A second widget, Quick Call,
changes appropriately depending on the default settings so that a user can make a
call quickly.

282 [email protected] Executive Summary

 Mobile UC Widgets allows users to easily select and change their calling
preferences while undocked. There are two Mobile UC widgets: the default call
type and Quick Call which can be installed. The widgets are color coded to
match the application where business calls are shown in blue and mobile calls
are colored green.

If the widget is set to Prompt, then the user is presented with a screen that allows
him to then choose Business or Mobile. This UI helps people who make roughly an
equal number of calls from their business and mobile lines. So regardless of her
usage scenario—mostly make business calls, mostly make mobile calls, or make
roughly an equal number of mobile and business calls—the UI lets the user be in
control and happy.
To set the default call type selection users may tap the Always place calls this way
checkbox.
The buttons are quite large, easy targets to hit, which comes in handy when the user
is walking or otherwise moving while using the phone. Designers strove to make the
buttons throughout the UI as large as possible so people who are moving, and yes,
even driving, can make selections quickly and error freely, having no-miss
experiences.

When placing calls while undocked, the user has the option to select whether to
place the call using their business or mobile line. Business calls are shown in
blue and Mobile calls are colored green.

Incoming calls are as easy to deal with as outgoing ones are in Mobile UC. When the
phone rings the user is prompted with a screen with the most important elements
displaying, including:
 Action items: large buttons to Answer or Decline the call, and
 Status items: the caller’s name, telephone number, and image.
The affordance on these items is appropriate, as the status items are legible and
recognizable, and the action items (buttons) appear to be interactive elements in all
important areas, like: size, visual treatment, and verbal terms used as commands.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 283

 When receiving a call, users are prompted with caller ID information and the
ability to answer or decline a call. If they have multiple lines, the incoming line
information is also displayed.

Once on a business call, the user may create a conference call or transfer the call.
(On a Mobile call the phone will be as it was before the app.) Users may choose who
to transfer the call to by tapping in their contacts, call log, or favorites.

If a user decides to transfer a call he may enter the number or select from the
call log, their contacts, or their favorites. He may also have the choice of talking
first (consultation), transferring without talking, or sending the call to his
voicemail.

The panel displays the length of time of the call in the lower left and displays the
time of day in the lower right. The name of the person on the call is clearly displayed
in the upper left. The commands to add more people to the call, transfer, and end
the call are all visible and are easy touch targets. The user may also conference in
another person, or make another call by tapping the New Line button.

284 [email protected] Executive Summary

 When a user is on a conference call, the other callers are displayed, and the UI
indicates that all parties are on a conference call.

The user may end the entire conference call, or even disconnect the last user who
was added to the call.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 285

 All callers are shown during a conference call. Users can disconnect from the
last person they added to the conference call by selecting the End Last button.

Sometimes a user may want to say a few words to a caller before adding him to the
conference call. Imagine you are talking with a manager or client and she needs
another person to answer a few questions. You may want to call the second person
and say just a few words to brief him before conferencing him in with the manager
or client. In these cases the user may choose the Talk First button.
After talking with that third party, the user has the option to add him to the
conference or cancel adding the person to the conference call.
When docked, users may have up to three concurrent calls and may start a new call
by tapping the New Line button. This button is only visible adding a new line is an
available option. While docked, all available features are shown on the screen,
including the ability to make another call, transfer, conference, or look up a contact
in the corporate directory. Each person the user is speaking with is displayed—name,
picture, and phone number—at the top of the call.

286 [email protected] Executive Summary

 When docked, users may have up to three concurrent calls and may start a new
call by tapping the New Line button.

The New Line button is only present when users can actually add another line
(that is, they have fewer than three currently active calls.)

When the user has three lines in use, the three callers are displayed at the top and
the New Line button is removed since a user can only have a maximum of three
concurrent calls.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 287

 Docked users can have up to three concurrent calls. Caller ID information for
the calls are provided along with indicators to communicate when a call is
active (green) or on hold (red.) A user can switch between calls by tapping the
call they wish to make active.

The Call Log feature is one that is often not available on traditional office phones, but
it is helpful and available in Mobile UC. Selecting Call Log displays a list of all
incoming and outgoing calls, sorted by most recent. Whether the call was outgoing
or incoming is displayed with text and icons in the list. Received calls show a blue
down arrow, placed calls show a green up arrow, and missed calls show a twisted red
arrow. Designers made it very easy to take action from this page too, rather than
making it status only. Helping users to move forward easily is one sign of a great
app design. The caller’s name and phone number are visible, as are the possible
commands to Talk First or Transfer.

288 [email protected] Executive Summary

 The Call Log displays all incoming and outgoing calls, plus the ability to select
calls that were recently dialed or received, for transferring or other available
functions.

Users may access their settings via the options menu. The following setting options
are available: show dial pad when docked, authentication credentials, notify low
signal, screen timeout, doc configuration. These follow Android conventions.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 289

 Settings are easily accessible from the options menu and follow android
conventions. In addition, there is a setup wizard allowing customization the first
time the user launches the application.

While most users probably won’t need help in this app, designers still provide a
thorough description of features and how to’s. A few of the topics described in help
include: Mobile UC features, placing calls while undocked, receiving calls while
undocked, placing and receiving calls while undocked, and transferring calls. Each
topic is presented in the list with an arrow to expand and collapse the write-up, so
users may read the instructions in place.

290 [email protected] Executive Summary

 An on-device Help system is available allowing the user to understand how to
use application features. Each section is expandable.

Mobile UC marries the power of the traditional office phone with the flexibility of the
mobile phone to make for a convenient, reliable, and simple experience for business
users who also have a personal life.

DESIGN PROCESS

From Idea to Execution

The journey from idea to execution followed an unusual path for this application. It
started with concept work and design, and then was marketed within the company to
convince stakeholders to say, “Yes, this is something we’re interested in pursuing
and developing!” The designs the team developed were instrumental in achieving
that outcome.
“Usually,” says Linda Borghesani, Manager, Device Marketing, “people come to us
[the design group] saying, ‘okay now [much further downstream] we need your
support.’ But in this case, it actually started with the graphics and the idea of doing
this dual persona application work on a cell phone.
The project originated from a kernel of an idea that was being circulated through
Verizon Wireless and the process of evangelizing the idea to key IT decision-makers
at various organizations drove the need to create design comps as a way to better
illustrate the concept.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 291

 “Humphrey, our director, was giving a lot of PowerPoint presentations, mock-ups of
what it might look like, to IT decision-makers at various organizations. So there was
a need for graphics just to imagine the product and make the product a little bit
more real,” says Borghesani.
That need was a lucky coincidence for the design group, which jumped on the
opportunity to flesh out the concept. This gave the in-house team a rare chance to
lead the charge and see to it that the product development eventually followed the
vision articulated through the designs.
“It was just a happy coincidence that we were on the same team that would
ultimately be responsible for developing the design so it just got started in our
department,” says Borghesani. “The early technical architecture discussions weren’t
done in isolation from the design work.”
Designing the application in-house put the design team in the driver’s seat and
ensured that they remained central to the process along the way.
“This [project] was much better because we had control and we did a lot of things in-
house,” she says. “I worked on another project where we had one company do the
wireframes and another company do the comps. We were just third party to
everything and the comps didn’t match the graphics nor match what we thought the
product should be like. So it was the fact that we were part of the core design team
[for this project] that has made the difference.”
Having the entire design team in-house allowed much more control of the product,
and helped avoid delays or costs associated with making design changes that were
due to technical constraints or other issues.

Usability Methods
The design process involved a number of iterative steps that were required to both
refine the user interaction layer of the application and at the same time ensure that
the application could deliver functionality despite (and constrained by) the challenges
of the multiplicity of PBX environments. The team used several methods to refine the
application design:
• Iterating wireframes: The team began with very detailed
wireframes that demonstrated task flows exactly, as in, “When you
press this button this should happen.” The wireframes even included
all the error messages that the user would receive when completing
the task. While the application functionality was effectively illustrated
by these detailed wireframe designs, the wireframes had to continue
to be iterated over time to accommodate changes that were required
for technical constraints as they were discovered. And this was a
constant flow of changes, across the designs. So, adding a design
layer to this meant keeping precise track of which wireframe aligned
with which comp.
“Then what we did was we had comps to go along with all the
wireframes,” says Borghesani. “So we labeled everything by screen
number. So [for example] this is undocked screen 621. Then we would
create a comp that would be marked ‘U621’ so that they’d know
exactly which comp matched up with which part of the wireframes.
That was a nice syncing process.”

292 [email protected] Executive Summary

 • Heuristic Review: True task-based usability testing was not possible
because the design was a moving target. The design had to be refined
to adapt as the team uncovered new technology constraints. To add to
the challenge, the team was working with a portable PBX system to
simulate the PBX environment of a company phone system. To work
around these limitations, the team chose to conduct mini-heuristic
reviews against the proposed functionality as it developed.
• Usability Testing: Once the design was stable, but before it was
implemented, the team conducted a usability test with eight people
who represented potential customers (Android users.) It was a
traditional talk out loud session where participants were asked to
interact with a paper prototype with screen variations. The product
idea scored very high with participants and some improvements to the
UI resulted from the usability testing. The feedback helped the team
remove screens and simplify some of the interactions. This gives the
user more feedback when he needs it, especially in using the transfer
and conference functions. The testing also contributed to the following:
o Improving the nomenclature for buttons and features.
Borghesani explains: “For several features, we subjectively
asked which wording users preferred and in other cases we
tested several concepts. For instance, we probed about the
wording to add a second call. One option was Add Call which
confused users [too similar to adding a conference call to the
current conversation]. In this case New Line was preferred and
is the wording we used.”
“Participants preferred seeing the names and pictures,
especially during conference calls to easily see who they are
talking with and for call control. In some cases, more feedback
was requested, such as more indication when a line was on
hold and adding a confirmation of the transfer before being
brought to the call log, which is the native approach in Android
today after completing a call. We added a message to inform
the user that the call was transferred successfully.”
Participants preferred to return to the phone’s home screen
when calls ended rather than the call log (which is how the
native Android O/S works.)
o Reducing taps: Borghesani explains, “We worked creating a
design that uses the least number of user actions whenever
possible, and we also tried to come up with as many defaults as
we could. We really tried to keep the application as simple as
we could because it is a very complex application to begin
with.”

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 293

 • Flattening hierarchy wherever possible: Sysko explains, “We
tested five concepts using an interactive prototype on an Android
device which varied in how we presented options to the users. The
designs we tested included a Full Carousel, Split Carousel, Up/Down
menu, Full Menu, and Partial menu, where less frequently-used options
were hidden rather than being presented on the screen all the time.
Users preferred to have all options visible to reduce extra steps rather
than a step to open and close a tab to see additional options.” Another
example is the Transfer function. “We could have put the Talk,
Transfer, or Send Voicemail on secondary screen after dialing; or we
could have users choose the action first and then choose the
extension. We decided let’s put it all out there for the user. Let’s let
them dial and then figure out what they want do with the call, but let’s
put it on one screen rather than displaying another pop-up to tell them
what to do.”

Designing to the Least Common Denominator (in PBX systems)

The team had to design for a variety of PBX systems, each with its own set of
technical constraints. While some of the PBX systems that the application works on
offer functionality that is above and beyond what’s included in the application, it was
a simple reality that the team had to draw a line in the sand somewhere, and decide
what to include and what to leave out.
The team settled on a manageable set of functions that were common across all the
PBX back-ends. This meant that there were some edge cases where this one PBX
could do something but it was decided to not include it in the app because it would
have added to customer confusion and wouldn’t have worked all the time, across
platforms.
This pantheon of PBX systems, even if designing to the least common denominator,
presented challenges for the team, especially during testing. They had to
accommodate multiple vendors on the PBX side, some with multiple versions of each
PBX. So there were many variables and everything needed to be tested with each of
those variations.
“I don’t think anyone realized there were so many difference between them or the
versions of the PBXs,“ says Borghesani. “Even with just Cisco there are so many
differences based on the version the company is using.”
Examples:
Sysko explains some examples of technical considerations that had to be resolved
along the way:
• Multi-person conference calls: “We could have provided
functionality to accommodate conference calls with more than four
people in mobile mode but it was determined that this would be
overwhelming and not used by the vast majority of users.”

294 [email protected] Executive Summary

 • Shared lines: The idea of shared lines is another example of
something that is supported in certain PBX systems. “There’s the
concept where an administrative assistant can access somebody else’s
line. That functionality is supported when the phone is docked and the
plan was to support that when undocked as well, but the problem is
not every PBX implements it in the same way or even has that
capability. So that is still being worked on and may be included in a
future version.”
• Mobile dial pad: “There were originally designs intended to replace
the mobile dial pad. So in dock mode we would be presenting a new
design for the dialer. In mobile mode, we originally had designs for
that. It was decided as a combination between usability simplification
and technical simplification to make use of the existing stock dialing
experiences that the phones had.”
• Reducing lag time: “We worked on how long should it take for a call
to be placed because behind-the-scenes what it’s really doing is two-
stage dialing. So we set a limit of six seconds to be able to do that
second stage dialing - that the call should not take any longer
(between the time you dial and the time it rings) than six seconds.
That required working with the PBX for the two-stage dialing.
• Adapting to switching behavior: “There are differences with the
switching behavior of mobile calls that are regional within the United
States. So, as with any system you have to integrate with other
systems and that requires technical adaptations in the application. This
was especially true with the conferencing feature, for example.”

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 295

 Project Timeline

PROJECT TIMELINE

2010 May
• Product ideation and internal championing
June
• Marketing Service Requirements (MSR) created
to describe the product
• Series of discussions with Product Marketing to
determine the scope of the features in addition,
we also had specialists: one team member was
a PBX expert, one an Android development
expert
July
• Draft graphics and design concepts to be
applied to use cases. These helped the team
evangelize the product and what the app was
going to “do.”
Summer
• Vendor RFPs issued. The team needed two
vendors to build the product: software and
hardware
August /September
• The usability team developed draft comps to
support use cases and product functionality.
October
• Usability testing on specific parts of the product
that the team thought would be most confusing
to users
November
• Began detailed wireframing to determine what
functions to include in the application and to
match the user cases and design specifications
December
• Baseline wireframes completed (docked flows,
undocked flows, docking, undocking, settings,
web portal and the setup wizard)

January 2011 • Official product announcement at the

Enterprise Connect Conference where the
application was named as a semi-finalist. An
alpha version of the application was
demonstrated on a working device with a
(portable) test PBX. The demo was tightly
scripted so the team could control what they
wanted to show and how it was demonstrated.
We received great feedback and enthusiasm
from Enterprise Connect attendees.
March
• Designed final graphics, layout guides and

296 [email protected] Executive Summary

 animation guides
• Tested alpha software to ensure all interface
elements displayed properly
• Tested the alpha software to be sure that
actions and results matched what was specified
in the wireframes (when you press this button,
this should happen)
Conducted a pixel perfect UI review. This review
happened in parallel with the wireframe
revisions and consisted of overlaying the
wireframes over the design comps to ensure
that everything was laid out properly and
would display well in the final product
April 2011
• Showed wireframes of the provisioning tool to
IT decision makers at corporations that were
interested in beta testing this product to get
their feedback on the system administrator tool
for setting up and configuring the service and
revised the wireframes as needed.
• Signed up these IT stakeholders at outside
corporations as beta testers

April-August 2011 • Testing and refinement of the Beta software for

user interface and technical issues

August 2011 • Designed new layout guide for portrait mode

devices
• Designed User Guide and Online Help System

September 29, 2011 • Official Product Launch

October 2011 • Chosen as Best Enterprise Product at CTIA

LESSONS LEARNED
Verizon team members share a few lessons learned from the redesign effort:
• Create good wireframes and design against them. “The wireframes
helped so much,” says Borghesani, “and we had comps to go along with all
the wireframes, and each box in the wireframe had a comp to go with it that
was numbered and labeled.”
• Create a generous timeline. The team learned the hard way that
application design is very time-consuming and without a generous timeline
things like testing get eliminated. “It takes a lot more time than you think and
I wish there had been time in the schedule for usability testing. We used

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 297

 ourselves internally and would walk around to each other so we could ask
somebody, ‘What do you think about this?’” says Borghesani.
• Budget for usability testing and budget time for the inevitable
changes. This rule is especially true, depending on the complexity of the
wireframes and the technical architecture. The more complex (and detailed)
the wireframes, the more time required to adequately test and revise them.
• It’s hard to test a moving target. The constant flow of iterative changes to
the wireframes meant there was no time for formal testing. “If only I had
known how much time it was going to take, especially to do all the revisions
and the testing. It was a very aggressive schedule to begin with and I think
that was part of the problem. This was the first time something like this was
done [in our group] so there were a lot of things that we had to change
midstream, either adding or taking away functionality that needed to be
redesigned into the application. And we needed to figure out where else that
might affect things, which is why the wireframes were so pivotal,” says
Borghesani.

TEAM
The team responsible for the application is in a user experience design group that’s
part of device marketing at Verizon Wireless. This placement of the team within the
company shifted as the project proceeded, but despite this shift, the team remained
with the project to its conclusion. Both a PBX expert and an Android expert were
added to the team to address technical considerations of the project. And as
mentioned before, the team hired both a hardware and software vendor to realize
the vision of the Mobile UC application and docking station.
The design team roles were as follows:

298 [email protected] Executive Summary

 DESIGN TEAM RESPONSIBILITIES
Role Responsibilities
Graphic Designers
• Design graphics, animation and layout (with a good
understanding of technical issues, PBX, and Android)
UI designers
• Develop settings for: set-up, provisioning, docking
and undocked state, and docking/undocking
• Create application logic and wireframes
• Test against wireframes
• Write screen verbiage
• Coordinate with vendors
Third Party testing
• Conduct product testing based on use cases
company
Product Managers
• Guide feature and functionality specifications
• Set project schedule
Third Party
• Implement designs
Developers
• Test and bug fix
• Design docking station

Pictured: Members of the application design team manning the company booth
at Enterprise Connect 2011 where the product was first announced: Stanimir
Dinov, Samantha Kalita, Kathleen Murphy, Linda Borghesani, Yeharn Hwang.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 299

 Ten Application Design Mistakes

Among the many submissions we judged that did not make our short winners’ list
were some truly wonderful applications. Also in this group were some applications
with some severe design and usability issues. Without mentioning any specific
applications, organizations, or features, the list below summarizes some of the top
reasons why some applications did not make the winners list.

1. STEPS IN A TASK OR PROCESS PRESENTED OUT OF ORDER ON

A PAGE OR SCREEN
Good screen or page layout is difficult for many reasons, namely it can be thorny to
prioritize tasks, and there is rarely enough space to present everything you want to
the users. Of course practical constraints can make layout particularly testing, but
what should not be a contributor to poor layout is ignoring the chronology of events.
If the order of tasks can match the user’s workflow or a known, logical, or obvious
process, then tasks should be presented in that order. Arrange forms chronologically
as you would a food buffet: plates to put the food on, the food, the condiments and
accoutrements, and lastly the utensils (if they are not on the table already.)

2. MISSING SIMPLE PATHS

When people finish a step they should not only have some sense of what happened,
but also where they might need to go next. Applications are getting better and better
at giving users feedback, but they are still sometimes missing that extra step of
leading the user to the next most likely choices and giving guidance for those
options. Some organizations know what a user's most likely choice is, but purposely
make another button more attractive--usually one that leads to more conversions. In
these cases sales or clicks may look positive in the short term, but that is not way to
forge a long-lasting relationship with customers.
Focusing only on what you (or the business) want people to do, not what they want
to do is a major issue. At a conceptual level some apps have major issue: designers
or business owners clearly just haven't thought about why people would actually use
it. At best these app designs turn out to be something nobody wants to use. At
worst, some apps actually mislead users and hide critical information in order to get
people to take some desired action. These short term gains end up hurting the
brand.

Finally, another element of poor paths for the user comes in the form of lacking
affordances, or not communicating when items are clickable, moveable, or in some
way possible to take action on. This is especially true on touch interfaces where
controls are very subtle, too small, don't look clickable, or are just plain invisible.

3. INVISIBLE CONTROLS AND CRYPTIC ICONS

Toolbars, icon bars, and menus with icons can certainly help people who can
recognize and or recall the icon. Locating an icon can enable people to perform tasks
in a single click, which is especially great for common or repetitive tasks. But when
an icon has issues, it may as well not exist.
Additionally, showing controls or icons for a few moments than taking them away is
a trend that seems to appear on mobile apps more and more, and is creeping in to

300 [email protected] Executive Summary

 desktop apps. This kind of UI is confusing and makes it difficult for users to find
commands.
Some related issues we saw include:
• Showing controls for a few seconds then hiding them
• Tiny graphics
• Blurry graphics
• Recognizable but confusing elements, sometimes overly using
metaphors
• Poor contrast between the icon and its background
• Multiple similar icons
• Pretty icons but that don’t help support the user’s core tasks
Attractive as they may be, icons need to support core tasks. Designers sometimes
didn’t seem to spend enough time on task analysis to figure out what steps people
need to take, and instead added attractive icons which resulted in a misallocation
screen real estate. Users are much better off with a plain and simple interface where
the navigation closely matches the tasks, they don’t have to spend time looking at
the navigation, and the content is the star of the show.
On a related note, we are also still seeing foreign elements such as expand/ collapse
arrows just hanging in space, and buttons that don’t look like buttons (clickable.)

4. HIDEOUS FORMS
Alignment and balance go an incredibly long way in form design and usage. So why
do designers still come up with forms that look like a game on Jenga 5? Users have
trouble dealing with forms with right-aligned labels, and fields that are also right
aligned, not aligned at all, or arranged in multiple sloppy columns.
And while stacked fields are typically better than multiple columns of fields, and
sectioned steps are better than a potpourri of fields, cutting a form into too many
screens, like mini-forms, makes the process take longer and can make the user feel
constrained. Some of the non-winning apps shied away for long forms and broke
their questions into too many steps.

5. UNCLEAR DIFFERENCES WHEN COMPARING CHOICES

Especially in an e-commerce environment people should be able to immediately tell
the difference between items. But even with compare-type features sometimes it is
difficult to decipher the difference between items, as images look the same and
small-print in lists makes it difficult to find the differences.

6. INDISTINGUISHABLE CURRENT SELECTION STATUS

From accordions to scrollbars to buttons to preview elements, rampant in application
design is the lack of feedback about the currently selected item. The issues almost
always surround the visual representation of the selected item. Something as subtle

5
From Wikipedia: Jenga is a game of physical and mental skill created by Leslie Scott, and currently
marketed by Parker Brothers, a division of Hasbro. During the game, players take turns to remove a block
from a tower and balance it on top, creating a taller and increasingly unstable structure as the game
progresses.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 301

 as the color of a tab not matching the related content below, or an expanded
accordion menu heading which has a line below it that visually separates it from the
sub-menu choices can completely baffle users.

7. BRANDING FEATURES OR MENU COMMANDS

Sure, sometimes your content or a link name may need to match a popular
advertising promotion at an organization, but making up catchy phrases for features
that are part of a task makes them confusing for people to find, understand, and
use. Leave the branding to the advertising campaigns and name features based on
tasks and common terms people use.

8. ONLY THE PROMISE OF CONTENT

Portals and dashboards are meant to deliver information as well as links to more
details, but some apps only provide the latter. On a page with several sections
displaying information, users should be able to decipher something, some answers,
even from the small amount of visible content. Designers must use discretion and
not cram the area full of too much content for the space allocated. But giving people
generic terms that constitute only the promise of content is not only boring but
dangerous, as users may not click something they need because the content on the
front page is too vague.

9. VIEWS AND MODES THAT TRAP PEOPLE

A view is meant to help people see information in particular ways and eliminate the
information not needed at the moment. A mode allows people to focus on one task
only, by restricting access to unrelated features. Designers employing either often
have good intentions, and both types of UI’s can work quite well. But when users are
not sure what a mode is or its purpose, they can have a terrible time. Also, if a
person is not aware that he filtered out information or is viewing a subset, again
things can get pretty messy pretty quickly.
Ideally, make modes and views understandable, easy to tell when you are viewing
content a particular way or in a mode for a certain task, and easy to get out of and
back to another state. Even better, allow people to choose commands within a mode
and make the mode forgiving by taking people out of the mode and to the area they
actually need to be in. Be very cautious about hiding important features while in a
certain mode or view.

10. BAD DIALOG

Sometimes a dialog box or page was obviously created as an afterthought and
nobody truly considered what was being said to the user, or what he is allowed to
reply.
If you tell a person something negative in a dialog box but present OK as the only
button, then you deserve whatever comes next. It’s incredibly annoying to users to
have to click OK when things are simply not okay. Consider using Close in those
situations.

302 [email protected] Executive Summary

 About the Authors

Kara Pernice is the Managing Director at Nielsen Norman Group and is an expert
with more than 20 years of experience in user experience (UX) research and design.
The Wall Street Journal called Pernice an "intranet guru." Since joining Nielsen
Norman Group, Pernice has led many major intercontinental research studies and
authored many research reports and guidelines, and is coauthor of the
book Eyetracking Web Usability (2009.) She has lectured around the world on a wide
range of topics, and her client work spans many businesses and industries. Before
joining NN/g, Pernice gained invaluable experience about building and managing UX
in an assortment of development environments and established several successful
user-experience programs. A champion for usability, Pernice chaired the Usability
Professionals' Association 2000 and 2001 conferences, and served as 2002
conference advisor. She holds an M.B.A. from Northeastern University and a B.A.
from Simmons College.
Kathryn Whitenton is a User Experience Specialist with Nielsen Norman Group. She
works with clients to evaluate the usability and information architecture of websites
in a variety of industries including technology, telecommunications, and media, as
well as corporate intranets. She has conducted usability research, eyetracking user
research, and studies of users on mobile devices in the United States, Europe, Asia,
and Australia. Her user studies have included general audiences as well as specific
consumer types, business segments, children, and seniors. She teaches seminars
on Information Architecture, Principles of Human–Computer Interaction, and
Fundamental Guidelines for Web Usability.
Prior to joining NN/g, Whitenton worked as a Usability Engineer with the University
of Washington Libraries, where she led user research and usability testing for the
Libraries' website. She also worked on developing the user experience for agile web
application startup companies, and as a Psychology researcher at the University of
Texas at Austin, where she managed a clinical research study funded by the National
Institute of Health. Whitenton holds Masters in Library and Information Science from
the University of Washington, and a B.A. in Psychology and Plan II from the
University of Texas at Austin.
Patty Caya (www.pattycaya.com) is a multi-platform journalist, editor, and award-
winning digital media producer. In her business writing, she specializes in topics
relating to usability, social media, and the business and technology of the web. For
NN/g she has co-authored the Intranet Design Annuals since 2008 and she wrote the
3rd and 4th editions of the report on Intranet Portals. She is the author of the report,
Enterprise 2.0: Social Software on Intranets.
For over a decade, Caya has split her time between journalism projects and web
consulting. She is an experienced content strategist and information architect. She
has consulted for many of Boston’s top interactive agencies, leading web and
intranet development projects as well as usability testing, research, and design
initiatives. Her client roster spans a wide range of industries and includes leading
national brands alongside mission-driven non-profits. She has a BFA from New York
University’s Tisch School of the Arts and has completed coursework in the User
Experience Program at Bentley University.
Dr. Jakob Nielsen (www.useit.com) is a principal of Nielsen Norman Group. He is
the founder of the “discount usability engineering” movement, which emphasizes fast

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 303

 and efficient methods for improving the quality of user interfaces. Nielsen, noted as
“the world’s leading expert on Web usability” by U.S. News and World Report and
“the next best thing to a true time machine” by USA Today, is the author of the best-
selling book Designing Web Usability: The Practice of Simplicity (1999), which has
sold more than a quarter of a million copies in 22 languages. His other books include
Hypertext and Hypermedia (1990), Usability Engineering (1993), Usability Inspection
Methods (1994), International User Interfaces (1996), Homepage Usability: 50
Websites Deconstructed (2001), Prioritizing Web Usability (2006), and Eyetracking
Web Usability (2009.) Nielsen’s Alertbox column on web usability has been published
on the Internet since 1995 and currently has about 200,000 readers. From 1994 to
1998, Nielsen was a Sun Microsystems Distinguished Engineer. His previous
affiliations include Bell Communications Research, the Technical University of
Denmark, and the IBM User Interface Institute. He holds 79 US patents, mainly on
ways of making the Internet easier to use.

304 [email protected] Executive Summary

 Acknowledgements

Thank you to all of the organizations and designers who submitted their applications
for this content. Special thanks to the winners in this report, for their wonderful apps
and for sharing them with us.
Also thank you to Susan Pernice for her assistance with this report.

© NIELSEN NORMAN GROUP WWW.NNGROUP.COM 305

About Nielsen Norman Group
Nielsen Norman Group (NN/g) is a consulting and research company that is solely
focused on user experience. We are not a Web design shop—we will tell you what
your customers want and how to vastly increase the business value of your site or
intranet, but we won’t build the new site for you. We are independent of technology
vendors and design agencies, so we can report the unvarnished truth about what
works and what doesn’t.

For a full list of services and current prices, please see www.nngroup.com/services
CONSULTING
• Independent expert review of the user experience of most designs, such as
a website or intranet: $38,000. (Lower prices for small focused reviews, like a
mobile app.)
• User testing: typically $25,000 to test a website or intranet; $45,000 for a
competitive study. Less for a mobile app or other small UI.
TRAINING
We produce an annual conference series where world-class experts teach the latest
findings about the usability of websites, intranets, mobile sites/apps, and email
newsletters. We also teach user research methodology so that you can hone your
skills and conduct your own usability projects with more success than if you use
weaker methods.

NN/g is the only company that presents high-end usability training events bringing
the same courses to the United States, Europe, and Australia. For the current
conference program, see http://www.nngroup.com/events
IN-HOUSE PRESENTATIONS
Most of our conference seminars are available for in-house delivery at your location.
Special training events are optimized for leveraging your own design questions:
• 3-day Learning-by-Doing Usability Testing ($23,000). We take you
through a user test of your own design, teaching usability principles with your
own project as the case study.
• Intranet Usability ($23,000). Combines a full-day tutorial with the lessons
from our testing of 27 intranets and a full-day workshop about your own
intranet’s usability, based on our review of your design.
• Application Usability ($16,000). Two days intensive course on everything from
screen design (buttons, field labels, widgets) to feature and workflow design.
• Writing for the Web ($9,000). A writing workshop using your own sample
content for the rewrite exercises.
• Fundamental Guidelines for Web Usability ($9,000). What everybody should
know about users’ online behavior and how to design better sites.
• User Experience (UX) Basic Training ($9,000). Bring your team up to speed
on all the most important concepts and methods of the field.
PRICES
Prices are stated in U.S. dollars and were valid when this report was published. Travel
expenses are extra for all training seminars and for many other services; prices are
higher outside the United States. Prices are subject to change without notice: for
current prices, please see http://www.nngroup.com/services
NEWSLETTER
Free e-mail newsletter published every two weeks with summaries of our newest
research and thinking. To subscribe: http://www.useit.com/alertbox/subscribe.html
TWITTER
Follow us at @NNgroup (highly abbreviated feed; missing much newsletter info).
Reports by Nielsen Norman Group
For a full list and to download reports, please see http://www.nngroup.com/reports
WEB USABILITY
• E-commerce user experience: series of 13 reports
• Corporate Image: presenting company information in a site’s “About Us” area
• PR section of corporate sites: supporting journalists
• Investor Relations area of corporate website: supporting investors
• Site map usability FREE
• B2B: guidelines for converting business users into leads and customers
INTRANET USABILITY
• Intranet usability guidelines: 10 reports based on user testing of 27 intranets
• Intranet information architecture (IA)
• Intranet design annual: published every year about that year’s 10 best intranets
• Sector-specific intranets: financial services, manufacturing, technology, retail,
knowledge-intensive, and government agencies
• Intranet portals: report from the trenches
• Social networking and collaboration features on intranets
MOBILE DEVICES
• Mobile websites and apps
• iPad usability FREE
• WAP phones — field study findings FREE

E-MAIL USABILITY
• Email newsletters
• Transactional email and confirmation messages
APPLICATION USABILITY
• Application design showcase: case studies of award-winning app UIs
• Customization usability
• Flash and Rich Internet Applications (RIA) FREE
SPECIAL USER SEGMENTS
• Usability of websites for children (age 3–12)
• Teenagers on the Web (age 13–17)
• College students (age 18–24)
• Web usability for senior citizens (age 65+)
• Beyond ALT text: improving usability for users with disabilities FREE

USER-CENTERED DESIGN METHODOLOGY

• Agile usability: Best practices for user experience on Agile development projects
• Return on investment (ROI) for usability
• Paper prototyping: a how-to video (40 minutes on DVD or Blu-ray)
• 230 tips to improve the way you run user tests
• Recruiting test participants for user testing FREE
• Testing users with disabilities
• How to conduct eyetracking studies FREE
OTHER
• Social media postings FREE
• Non-profit and charity websites: attracting donations and volunteers
The Usability Week 2012 Conference Who Will You Meet
at the Conference?
Many conferences offer cavernous exhibit halls, brief seminars on second-hand
Companies that sent the most
discoveries, and a sense of anonymity that can be truly alienating. people in 2010 and 2011:
Usability Week takes a different approach.
Accenture
ADP
Knowledge, Directly from the Source American Express
In place of scattered, shallow talks, Usability Week offers 5 or 6 days of deep Apple
AQA (Assessment and
learning as international experts lead full-day training courses on topics such as:
Qualifications Alliance)
• UX Basic Training AT&T
• Fundamental guidelines for Web usability Autodesk
• Information architecture (IA) principles BMC Software
• Writing for the Web Canada Post
CBC
• Application design
Channing Bete Company
• Designing usable social features Chevron
• The human mind (how your users think) Church of Jesus Christ of Latter Day
• Mobile websites and touchscreen/gesture apps Saints
Cisco Systems
The College Board
Course levels range from introductory to advanced; you can sign up for as few Cox Communications
as 1 or 2 days or as many as 6. Deloitte
Deutsche Lufthansa AG
Many sessions include hands-on training exercises that let you apply what you Discover Financial Services
eBay
learn immediately; all ensure that you’ll learn tools you can use to improve your Ericsson
website, intranet, or application as soon as you get home. Euro RSCG Life MetaMax
Fidelity Investments
Gap Inc Direct
Networking Opportunities GE
Because you’ll spend each day in a group with in-depth focus on a single topic, Google
you can discuss problems, share solutions, and make lasting connections with Herbalife International
your peers. Hewlett-Packard
ING
Intel
Pay only for the days you need. The more days you attend, the deeper the Intuit
discount. Early bird rates save even more, so sign up early! Itaú Unibanco
John Lewis
Kaiser Permanente
Venues LexisNexis
The conference visits several of the world’s great cities, providing ample Los Angeles Times
incentive to continue networking offsite at world-class restaurants, clubs, and Manulife Financial
attractions. McGraw-Hill Companies
McKesson
McMaster Carr Supply Company
Choose your city at http://www.nngroup.com/events for agendas, location, Microsoft
pricing, and registration information: National Marrow Donor Program
Organisation for Economic Co-
operation & Development
• Amsterdam: April 23-27, 2012 (OECD)
• Washington D.C.: May 14-18, 2012 Overstock.com
• Chicago: June 25-29, 2012 Pearson
• Toronto: July 23-27, 2012 Precision Nutrition
• Sydney: August 12-17, 2012 Qualcomm
RBS Royal Bank of Scotland
• San Francisco: September 23-28, 2012
Research In Motion (RIM)
Rockwell Automation
We Can Come to You Royal Bank of Canada
Sage Software
You can book most of our training courses for in-house presentation at your
Samsung Electronics
company or organization. If you have a large team this is cheaper than sending Sandia National Laboratories
them all to the conference. See http://www.nngroup.com/services/workshops SAP
Sapient
Scotiabank
Sony
State Library of Victoria
Symantec
T. Rowe Price
TD Ameritrade
TFO (Télévision Francophone en
Ontario)
Thomson Reuters
Totaljobs Group
Towers Watson
U.S. Office of Personnel
Management
Vanguard
Verizon
VMware
WebMD
Wells Fargo