Copyright 2008 by Kelley School of Business, Indiana University. For reprints, call HBS Publishing at (800) 545-7685.

BH 294
Business Horizons (2008) 51, 419—428

t
os
www.elsevier.com/locate/bushor

rP
Creating emotions via B2C websites
Marilyn Y. Jones a,*, Mark T. Spence a, Christine Vallaster b

a
Faculty of Business, Bond University, Gold Coast, QLD 4229, Australia

yo
b
Marketing Department, University of Giessen, Balanstrasse 16/2, 81669, Muenchen, Germany

KEYWORDS Abstract Compared to offline media communications, business-to-consumer (B2C)

B2C websites; websites possess unique characteristics that affect the likelihood of generating
Emotions; emotional reactions to the web experience itself, the brand, fellow customers,
Affect transfer; and employees of the firm. The emotion-causing antecedents elucidated in this
article are the website’s vividness, interactivity, challenge, interaction speed, ma-
op
Telepresence
chine memory, and allowable social interactions. Depending on how a website per-
forms on these dimensions, positive or negative emotions may result. For example,
using machine memory to automatically generate purchase recommendations based
on prior consumption patterns may be perceived as pleasantly surprising, while a firm
sending unsolicited emails based on a user’s cookie trail may be annoying. Regardless
of the valence of the resultant emotion, the feelings generated may attach to the
brand. Because a goal of brand managers is to get consumers to associate positive
tC

emotions with a brand, a pre-condition to developing and managing a website is

understanding these emotion evoking antecedents and their ramifications.
# 2008 Kelley School of Business, Indiana University. All rights reserved.

1. The online experience to the Internet confer emotion-evoking advantages

that offline media lacks. Prompting and attaching
No

Features that online media, but generally not offline positive emotions to a brand increases customer
media, possess include offering users the opportunity loyalty, trust, and market share (Fournier, 1998;
to interact with the website, effect the challenge Steenkamp & Dekimpe, 1997).
involved during the user-website interaction, control Consumers increasingly turn to the Internet to
content, and connect with other people. In addition, search for, evaluate, purchase and, in some cases,
the web host can learn about the user and capitalize virtually consume products. Online consumer deci-
on this knowledge via machine memory. We start with sion processes can be cognitively demanding and
the supposition that all media can be emotive, not highly emotive. Websites must therefore be care-
always in a positive way, noting that features unique fully managed, and offer an engaging customer
experience. How is this done? Research pertaining
Do

to web use in general has unearthed antecedents

* Corresponding author.
E-mail addresses: [email protected] (M.Y. Jones),
(Novak, Hoffman, & Yung, 2000) leading to flow,
[email protected] (M.T. Spence), which Hoffman and Novak (1996, p. 57) define as ‘‘a
[email protected] (C. Vallaster). seamless sequence of responses facilitated by

0007-6813/$ — see front matter # 2008 Kelley School of Business, Indiana University. All rights reserved.
doi:10.1016/j.bushor.2008.03.009

This document is authorized for use only by Hemangi Bhalerao until March 2013. Copying or posting is an
infringement of copyright. [email protected] or 617.783.7860.
 420 M.Y. Jones et al.

machine interactivity . . .[that is] intrinsically enjoy- one feel excited or playful, which is a positively

t
able . . . accompanied by a loss of self-consciousness, valenced affective state with high levels of arousal.
and . . . self-reinforcing.’’ However, these authors For our purposes, emotions are of short duration

os
caution that their research ‘‘has not considered the and object specific, in contrast to moods which are
specific elements of commercial Web site design of milder intensity, longer-lived, and ‘‘may be eli-
that facilitate a compelling experience, nor how cited and maintained without conscious awareness
this experience is likely to vary across the wide of the feeling state, its cause, or its influence on
range of commercial sites found on the Web current activities’’ (Cohen & Areni, 1991, p. 191);
today’’ (Novak et al., 2000, p. 33). The overarching for example, a person might get out of bed in a bad

rP
objective of the research effort here is to fill this mood. Emotions are characterized by the presence
void. of action tendencies, such as flight; bodily symp-
More specifically, the contribution of this article toms due to arousal, including accelerated heart
is threefold. First, based on an extant review of the rate; and facial expressions, such as smiling (Plut-
literature, we present a conceptual model of six chik, 1994). Significantly, emotions can create af-
characteristics unique to business-to-consumer fective traces that are attached to the object that
(B2C) websites that affect the likelihood of gener- prompted the emotion, thereby affecting an indi-
ating emotional reactions to the web experience vidual’s attitude toward that object. This affective

yo
itself, the brand, fellow customers, and employees trace, unlike the emotion that prompted it, can be
of the firm. These feelings subsequently attach to long-lived. Readers may be familiar with the adage:
the brand, thereby affecting consumer behavior. In I forgot what you said, but I remember how you
addition to highlighting the antecedents causing made me feel.
emotions, we note which are or are not under the Plato and Aristotle made reference to the trilogy
control of the webmaster. Second, while the intent of the human spirit: thoughts, feelings, and desires
of the model is to show how to create positive (see Cohen & Areni, 1991). It is now generally
emotions, we share examples of how underperform- accepted that mental activity involves both cogni-
op
ing on these characteristics can cause negative tions and affect. The former has a stronger effect on
emotions. Indeed, because users are exposed to attitude formation, but the latter has a pervasive
multiple content elements within a website (such and, oftentimes, not subtle effect (Lindgaard,
as FAQs, email enquiries, and payment systems), Fernandes, Dudek, & Brown, 2006). For extant re-
various brand or product announcements (video views of emotions-related literature, see Bagozzi,
clips and brand descriptions), and interface with a Gopinath, and Nyer (1999) and Watson and Spence
variety of actors (customer representatives and (2007).
tC

fellow chat room guests), it is reasonable to assume Some scholars are skeptical about the ability
that web environments are particularly prone to of websites to evoke emotions (Leong, Huang, &
evoking multiple feelings–—some good and some Stanners, 1998). However, web practitioners count-
not. Insights on mixed emotions are shared. Third, er by noting (a) the multimedia richness of the
we present managerial implications. Collectively, medium (Carton, 2002; Philport & Arbittier, 1997);
these insights are directed toward brand managers, (b) evidence supporting the emotive potential of
promotion coordinators, and IT staff responsible for combining words, pictures, and sound (Allen, 2001);
No

website design and maintenance. and (c) the ability to form relationships via person-
alization and interactivity (Lindstrom, 1999). In a
series of empirical studies there was evidence that
2. Causes and consequences of opinions concerning a website’s visual appeal are
emotions formed within 50 milliseconds, and that this near
spontaneous reaction can carry over to assessments
Studies on emotions within marketing usually begin of other features on the website (Lindgaard et al.,
by distinguishing emotions from closely related 2006), conceivably causing an individual to disre-
concepts such as affect, mood, and attitude. Brief- gard or discount evidence not consistent with that
ly, affect is a reference to a valenced feeling state. individual’s initial reaction. This finding speaks vol-
Do

Both moods and emotions have valence, as well as umes for the need to carefully manage a website.
arousal, and can therefore be thought of as specific An example of a website that evokes positive
examples of affect. Cohen and Areni (1991, p. 191) emotions for many visitors is that of BMW
note, ‘‘The distinction most commonly made be- (www.bmw.com). Prospective owners can configure
tween moods and emotions is that emotions are the car of their dreams, a task that is exhilarating for
more intense and stimulus specific than moods.’’ some. In addition, the website also showcases BMW
Thus, downloading a stimulating video may make art cars from the last 30 years, offers a Lifestyle

This document is authorized for use only by Hemangi Bhalerao until March 2013. Copying or posting is an
infringement of copyright. [email protected] or 617.783.7860.
 Creating emotions via B2C websites 421

section where current and prospective owners can are elucidated shortly. What is important at this

t
purchase BMW-themed products, and provides access juncture is that each action can result in a positive
to the acclaimed BMW films. In short, the website or negative emotion. This figure is for illustrative

os
presents a convenient opportunity for BMW owners purposes only; by no means are the actions and
and interested parties to indulge their fascination emotions shown meant to be exhaustive. The resul-
with the brand. tant emotions, ephemeral though they are, affect
A goal of brand managers is to get consumers to brand evaluations either by a direct affect transfer
associate positive emotions with a brand. For ex- (where the residual affective trace attaches to a
ample, Pepsi’s latest attempt to reinvigorate the brand or organization) or by triggering cognitive

rP
flagship brand is to roll out new themes every few elaborations (such as when a prompt reply to an
weeks, each with its ‘‘own website with video clips enquiry results in pleasant surprise, which may
and other enticements to engage consumers. The generate a lasting cognition regarding the helpful-
aim is to represent the ‘fun, optimistic and youth- ness of the company).
ful’ spirit of Pepsi’’ (‘‘Trouble Brewing,’’ 2007). Individuals appraise aspects of the web experi-
Interacting with websites like those of BMW and ence (i.e. objects, actors and events) much like one
Pepsi can generate feelings directed at an object reacts to a service encounter where the actor,
(the brand), actors (interactions with a customer object, and event are all visible. For example, an

yo
representative or with other customers), or the individual’s overall evaluation of a particular hotel
event (interfacing with the website; e.g. down- is likely to be affected by interactions with the
loading video clips or entering personal data). reception staff, which is a critical service encounter,
These emotion-evoking avenues operate indepen- but one that is tangential to the core product. Just
dently. Therefore, an individual might find watch- as contact with the hotel receptionist constitutes a
ing the video clips provided on the website moment of truth which provides an opportunity to
pleasurable, but be annoyed by comments posted evaluate an organization and its offerings, a website
in a chat room. experience–—being a form of two-way communica-
op
In Figure 1, we provide an emotion for the type of tion–—is also a moment of truth, and must therefore
actions that can occur when an individual is inter- be managed accordingly. Based upon an extant
acting with a website. The antecedents in the figure review of the literature, we present Figure 2, a
tC
No
Do

Figure 1. Examples of website actions and resultant emotions

This document is authorized for use only by Hemangi Bhalerao until March 2013. Copying or posting is an
infringement of copyright. [email protected] or 617.783.7860.
 422 M.Y. Jones et al.

t
os
Figure 2. Factors affecting the development and transfer of emotions to brands in an online environment

rP
conceptual model of web-enabled emotion causing Krampf, and Palmer (2001) show that the more vivid
antecedents. The antecedents elucidated within the online media, the greater the involvement with
this article are the website’s vividness, interactivi- the web experience.
ty, challenge, interaction speed, machine memory, Creating telepresence is therefore an objective
and allowable social interactions. Within this sec- for those seeking to stimulate emotions via a web-
tion we clarify the components of the diagram, site. However, several issues impede the ability to
starting from the top left. do so. Most notable is that while vividness is under

yo
the control of the webmaster, interacting with a
2.1. Engage the senses website can only be encouraged, not forced. Ma-
chine interactivity is explored next.
Studies pertaining to offline media show support for
the effects of four factors on the demand for an 2.2. Encourage interaction with the
individual’s mental resources: its vividness or rich- website
ness (Shiv & Fedorikhin, 1999), the amount of com-
peting stimuli (Edell & Keller, 1989), the copy style The likelihood and ability to interact with a website
op
used, and the ad layout (Peracchio & Meyers-Levy, is influenced by task characteristics and individual
1997). In the web domain, a primary driver is tele- differences. Addressing task characteristics first, an
presence. Telepresence is jointly affected by the individual’s ability to interact with a website comes
combination of vividness (media richness) and ma- from the (a) range of controllable elements, such as
chine interactivity, often referred to as user control. temporal display or spatial arrangement, (b) speed
Vividness is the focus within this section. Vividness of response to interactive commands, and (c) map-
arises from the variety and verisimilitude of sensory ping; that is, the fit between the action required and
tC

modes used (Coyle & Thorson, 2001; Klein, 2003; the natural way of producing a result (Steuer, 1992).
Steuer, 1992). Sources of media richness include flash For example, to drag and drop a file into an image of
animation, video clips, the breadth of sensory ele- a garbage can has a higher level of mapping than
ments available or soon to be available (i.e., visual, pressing the delete key; therefore, the former is
aural, tactile, olfactory, body balance/orientation), more engaging. Some ways that viewers exercise
improved computer displays, and the ability to make control over content and presentation are to en-
convincing digital images. large, reduce, and rotate images; choose the order
No

Telepresence is defined as ‘‘the compelling sense of viewing brand messages; adjust the sound; and
of being present in a mediated virtual environment’’ determine the type and amount of information they
(Novak, Hoffman, & Duhachek, 2003, p. 9). High desire via linked sources. Consider the last point:
levels of telepresence serve to mimic natural shop- providing an abundance of easy-to-follow product
ping environments: consumers can react to products information could be pleasing, whereas providing
and information much as they do in-store. There is long-winded technical descriptions in response to a
ample offline evidence demonstrating emotional straightforward enquiry could be annoying.
responses to atmospheric, spatial, and assortment Interactivity is not unique to the Internet; offline
characteristics within retail stores (Donovan, media can also permit consumers to exercise some
Rossiter, Marcoolyn, & Nesdale, 1994; Machleit & control. For example, there are ways to interact with
Do

Eroglu, 2000), and growing evidence for similar television programs by voting or purchasing. But
reactions online (Eroglu, Machleit, & Davis, 2003). overall, TV is considered a lean back medium, where
For example, including avatars as sales agents on a viewers are passive absorbers of one-way, mass com-
website has been shown to increase satisfaction munication, while the Internet is a lean forward
with the retailer, improve attitudes toward interactive medium, where viewers determine what
the product, and increase purchase intentions is seen and what actions occur (Dijkstra & van Raaij,
(Holzwarth, Janiszewski, & Neuman, 2006). Griffith, 2001). Consumers appreciate the opportunity to

This document is authorized for use only by Hemangi Bhalerao until March 2013. Copying or posting is an
infringement of copyright. [email protected] or 617.783.7860.
 Creating emotions via B2C websites 423

influence brand-related communications (Levine, and machine interactivity (collectively, telepre-

t
Locke, Searls, & Weinberger, 2000), something that sence), along with challenge and interaction speed
is not feasible with print ads and television. The are empirically supported drivers of flow (Novak

os
challenge for marketing on the Internet is to offer et al., 2000; see also Csikszentmihalyi, 1990). In
machine interactivity and user control that produces this section, we address challenge. We take the
a satisfying electronic experience while achieving an position that in B2C websites, it is unlikely that
effective marketing outcome. an individual would experience extreme feelings
To address individual differences, we start by of ‘‘flow [where] time may seem to stand still and
noting that just like offline shopping, online shop- nothing else seems to matter’’ (Novak et al., 2000,

rP
pers exhibit a variety of orientations. Brown, Pope, p. 23). But websites can be engaging, and challenge
and Voges (2003) empirically unearthed five orien- affects the likelihood of them being so.
tations, ranging from individuals seeking conve- Challenge reflects an individual’s computer liter-
nience to those deeply involved with, and acy and desire to engage mental resources. The
enjoying, the shopping experience itself. It is rea- relationship between the valence of emotions and
sonable to assume these different orientations challenge is an inverted U: a level of challenge that
would be reflected in the time spent online, number is too low or too high is likely to result in a negative
of links examined, frequency of visits, and so on. For emotion, while positively valenced emotions

yo
a person seeking a high level of engagement, a emerge when the challenge is consistent with an
media-rich site that offers many links and resources, individual’s skill and desire to engage. But the desire
for example, would be likely to produce a positive to engage is an individual characteristic, and–—con-
emotion. Conversely, an individual who wants to sistent with the aforementioned comment on the
quickly confirm brand-related information but finds variety of online shopping orientations–—it should be
he or she has to negotiate multiple links, could assumed there are grades of engagement. Further-
become annoyed. Depth of content should therefore more, even within an individual, it is reasonable to
be layered on; a good heuristic is never require a assume that his or her willingness to engage varies
op
user to go to more than three links to achieve an with activity. Thus, within B2C websites levels of
objective. challenge can, and probably should, be varied. FAQs
The ability to interact with a website provides an and payment systems should be simple, while offer-
opportunity to relate to an organization and, impor- ing an opportunity to create, for example, a brand
tantly, to its brands. Fournier (1998, p. 346) main- by attribute matrix will necessarily be more de-
tains that for a brand relationship to exist, ‘‘the manding. There is anecdotal evidence which sug-
partners must collectively affect, define, and rede- gests that, despite the popularity of electronic
tC

fine the relationship’’–—a nice quote capturing the games, an increasing number of college-aged stu-
utility of user control. Researchers have found sup- dents are finding computer games frustrating rather
port for relationships between elements of user than entertaining, due to the effort involved to
control and positive feelings about the website, learn the rules; the challenge involved is too high.
as well as positive feelings about the online shopping Emotional rewards and frustrations from interacting
experience (Ariely, 2000; Koufaris, 2002). Using with computers are well-established (Cockton,
Fournier’s (1998) Brand Relationship Quality scale, 2002).
No

Thorbjornsen, Supphellen, Nysveen, and Pedersen

(2002) demonstrated a link between interacting 2.4. Whatever you do, don’t be slow
with a website and the sense of brand relationship,
especially among more experienced web users. User control implies an experience where the speed
For good examples of vividness and machine and display of feedback constitutes an exchange
interactivity in particular, visit www.pepsi.com. that can be either pleasing, such as quickly access-
That site’s homepage options include an abundance ing information sought, or annoying, such as data
of brand-related material, as well as entertainment intensive links that require a long time to download.
and fashion sections, with a one-click link to ‘‘youth Evidence so far suggests there is a threshold effect:
empowerment and diversity issues’’. At www.nike. slow interaction speed has a negative effect on the
Do

com, users can tailor-design shoes if they so desire. evaluation of a website, whereas fast speed has a
null effect (Dellaert & Kahn, 1999). We can probably
2.3. Be reasonable in your expectations all relate to rebuking ourselves for naively clicking
on a data-intensive, and therefore time-intensive,
Assuming an individual wants to interact in a media- download. To mitigate the negative affect due to
rich environment, other antecedents can come into slow download speed, the BMW website allows users
play that affect the resulting emotions. Vividness to select whether they are connected via broadband

This document is authorized for use only by Hemangi Bhalerao until March 2013. Copying or posting is an
infringement of copyright. [email protected] or 617.783.7860.
 424 M.Y. Jones et al.

or modem, in which case the content displayed instant messaging, and bulletin boards are the most

t
adjusts accordingly. Users are forewarned that the common ways for users to interact with one anoth-
modem version has less rich content, but the benefit er, although there may be live human assistance and

os
is faster download. In some instances file size is interactive decision aids.
noted, as well, so users can decide if they wish to Social interaction about brands ranges from sim-
proceed with the download. ple word of mouth recommendations to physical
participation in brand communities. One study of an
2.5. Machines have memory –— Use it! offline brand community interaction where Jeep
Camp was studied showed the ability of active

rP
When discussing a brand relationship it is necessary social participation to move customers from nega-
to distinguish between an established relationship tive or weakly-held emotions to strong positive
and an emerging one. A growing portion of the emotions, such as happiness or love, about a brand
Internet population uses a limited set of websites (McAlexander, Schouten, & Koenig, 2002). In addi-
on a daily basis (Forsyth, Lavoie, & McGuire, 2000). tion to stimulating emotional responses, social in-
An opportunity available to webmasters where re- teractions serve to integrate participants (Kemper,
curring visits are common is to offer processing 1978), thereby creating a sense of community or
shortcuts, thereby reducing the cognitive resources belongingness.

yo
required to choose or use the website. Machine There are numerous ways to interact socially with
interactivity, discussed previously, is therefore en- other customers on a website, ranging from reading
hanced by the capacity of electronic media to re- the ratings provided by persons either known or
member something about the customer and make unknown to the user, to chat rooms that include
use of that knowledge in future transactions. This video as well as audio streaming. Most online social
requires machine memory. groups take the form of message boards or typed
Machine memory stems from two sources: what is chat rooms. Online communities show we intention,
stored on website databases and linked to an indi- which is a sense of group involvement (Bagozzi &
op
vidual user by a cookie or activated via a log-in Dholakia, 2002), and virtual social interactions
process, and by what a user is willing to share. At give rise to a sense of us versus others (Holland &
the most basic level the website database should be Menzel-Baker, 2001). Popular online communities,
able to retrieve the user’s browsing patterns and such as Bayer’s support groups for users of its phar-
transaction outcomes, such as timing and frequency maceutical products, testify to the value of con-
of visits, and monetary value of transactions. With necting people.
more self-explication from the user, the database Bagozzi and Dholakia (2002) show that participa-
tC

may contain personal and financial details as well as tion in online groups is motivated at least partly by
preference information. Amazon.com has been a anticipated emotional experiences. Consciously or
trendsetter in harnessing the customer relation- otherwise, social interactions prompt evaluations of
ship-building potential afforded by intelligent ma- others’ behaviors and attributions regarding their
chine memory: by tracking and categorizing past intentions. Depending on the outcome of this ap-
book purchases, the website can provide like-mind- praisal process, emotions such as pride, shame, and
ed purchase recommendations. But utilitarian web- admiration may occur (Watson & Spence, 2007). For
No

sites like Turbotax.com have also capitalized on the example, after reading comments from satisfied
ability to retrieve, for a fee, personal information customers, an individual may be proud because
from previous years’ tax returns, thereby expediting they, too, have chosen to use the brand.
tax form preparation. There is considerable variation in the realism of
the interactions, ranging from direct, real-time
2.6. The good and the bad of web-hosted interaction to more distant exchanges, such as
discourse reading other users’ online ratings. In the case of
Lands’ End’s Live help feature there is temporal
Traditionally, potential and current customers in- synchronicity (Hoffman & Novak, 1996), meaning
teract with a company via the telephone, by mail, that feedback and results from either the company
Do

at a service counter, or by using the product. In this or other customers occurs immediately. An individ-
way, they have developed a relationship to the ual anticipating temporal synchronicity in an envi-
brand and the company. Similarly, customers inter- ronment that does not support it is likely to be
act with other customers in informal settings away annoyed.
from the company. Websites offer excellent oppor- Overall, the area of social interaction is an area
tunities for relationship building, whether with the of great opportunity, but also one of great vulnera-
company or with other customers. Chat rooms, bility. If interactions are with fellow website users,

This document is authorized for use only by Hemangi Bhalerao until March 2013. Copying or posting is an
infringement of copyright. [email protected] or 617.783.7860.
 Creating emotions via B2C websites 425

the webmaster has little control over content; as Traditionally, website-generated emotions were

t
such, it should come as no surprise that employees regarded as potentially valuable, but difficult to
may provide blogs and participate in chat rooms, produce and manage. However, as we have endeav-

os
oftentimes unbeknownst to other users. Firms ored to achieve by elucidating unique character-
should carefully monitor postings by customers on istics of B2C websites–—the rich media, the ability to
their websites for both positive and negative com- connect with other consumers, the capacity to in-
ments. In the event that an electronic version of a teract directly with the website, and the user’s
negative word-of-mouth comment appears on the control over actions and content–—this is no longer
site, the webmaster cannot stop it, although he or the case. Websites can now closely approximate the

rP
she can endeavor to initiate a dialogue to mitigate gamut of real-life consumer experiences, increasing
the customer’s frustration, or at least identify the the chances that a consumer will feel emotions.
cause of the frustration so corrective actions can be Additionally, websites can perform functions that
taken to prevent a reoccurrence. When interactions may not be feasible in-store, as the BMW and Pepsi
are with an employee of the firm, this constitutes a examples illustrated. We acknowledge that TV and
moment of truth, meaning a contact point that print stimuli also produce measurable emotional
affords the customer or user the opportunity to reactions, but they lack the immediacy of personal
gauge the goodness of the provider. An unsatisfac- experience with the actions of the company and the

yo
tory or slow response will be directly attributed to process of choosing, buying, and using the brand.
the organization sponsoring the website, and will Therefore, TV and print ads may be better at re-
likely result in feelings ranging from annoyance to inforcing emotion than creating it for the first time;
outright anger. however, we posit the latter is better achieved
through a compelling web experience.

3. Attaching emotions to a brand 3.1. Extreme feelings have a

disproportionate effect
op
Experienced emotions are short lived, but the
residue of emotion, its affective trace, can be Websites are multi-faceted exposures to various
stored in long-term memory, as can brand-specific events, brands or products, and actors; as such, it
cognitions prompted by the emotion. Studies on is reasonable to assume the web environment is
the effect of advertisements on brand evaluations particularly prone to evoking multiple feelings.
support this position. Appraisals of a stimulus, or Mixed emotions do occur, although supporting evi-
an advertisement in prior studies, produced an dence is not web specific. For example, in one study,
tC

attitude toward the stimulus that transferred to subjects were asked to rate the extent to which they
the brand. This peripheral route to persuasion has felt six different emotions that they experienced as
been widely supported, but is generally thought to a result of reading about an intentionally emotion-
have a weaker affect on attitudes toward the evoking incident. It turned out that the reaction to
brand than the central route, where brand-specific some, but not all, of the scenarios produced pleas-
cognitions affect attitudes towards the brand di- ant surprise as well as anger, which are emotions of
rectly, not indirectly via the stimulus (Petty & opposite valence (Watson & Spence, 2007).
No

Cacioppo, 1981). MacKenzie and Lutz (1989) ex- An explanation for such outcomes is provided by
panded our understanding of these routes to per- Fredrickson and Kahneman (1993), who postulate
suasion by empirically showing that should there that memories of emotion-laden experiences are
be high levels of involvement with the stimulus, reconstructed as snapshots in time, not as a contin-
which is entirely feasible if a user is engaged with a uous flow of events. These snapshots are indepen-
website for a non-trivial period of time, the atti- dent events, each with a unique response. Their
tude toward the stimulus prompts deeper cogni- work, in conjunction with that of others (Ross &
tions about the brand itself. For example, feeling Simonson, 1991; Varey & Kahneman, 1992), con-
happy about quickly locating important product cludes that specific snapshots in time are dispropor-
information on a website might result in long- tionately weighted when forming an overall
Do

lasting cognitions about the company being good evaluation of the experience. In other words, an
and helpful, clearly a desirable cognition to attach individual’s perception of the overall experience is
to a brand. Generating positive emotions during not an average of the sum of its parts. There is
the website interaction experience therefore con- strong evidence for a ‘‘preference for happy end-
fers benefits to a brand, either directly via affect ings’’ effect (Ross & Simonson, 1991), meaning that
transfer or indirectly via heightened brand-specific the last emotion felt has the strongest effect on an
cognitions. individual’s overall evaluation of the experience.

This document is authorized for use only by Hemangi Bhalerao until March 2013. Copying or posting is an
infringement of copyright. [email protected] or 617.783.7860.
 426 M.Y. Jones et al.

There is also evidence that the most extreme expe- We focused on six: vividness, interactivity, chal-

t
rience felt will have a disproportionately strong lenge, interaction speed, machine memory, and
effect. Collectively, these insights are referred to allowable social interactions. We are not discount-

os
as the Peak-End rule, in reference to the dispropor- ing the relevance of findings unearthed in relation
tionately strong effect that the peak emotional to offline media, such as the amount of competing
experience and the end state have on overall eval- stimuli (in short, too much is bad), copy style (con-
uations. This insight has clear practical import: sistency is good) and ad layout (there are consistent
consumer research should be conducted which patterns of search), that affect an individual’s
tracks moment-to-moment responses to interac- processing effort. Rather, in addition to these is-

rP
tions with elements of a website, to identify and sues, the unique characteristics of websites eluci-
mitigate actions that cause negative affect as well dated herein can be and are themselves emotive,
as to identify strong, positive affect-evoking events. examples of which were shared in Figure 1.
Ideally, one of the latter should be moved to the end Consumers increasingly turn to B2C websites to
state, if feasible. search for, evaluate, purchase and, in some cases,
Unfortunately, what is little understood is what virtually consume products. In doing so, it is increas-
happens in the case where there is a series of mild ingly likely they will engage in multiple actions, such
emotions, both positive and negative. For example, as reading brand descriptions, downloading a video

yo
a user might be annoyed with the slow download clip, or entering a chat room. Given the multiplicity
speed, but pleasantly surprised with the informa- of actions, websites are prone to evoking a variety of
tion provided about the brand’s performance. Nei- emotions. Evidence to date suggests that the domi-
ther of these emotions is highly arousing. Would one nant or peak emotion felt, or the last emotion
emotion dominate, or would these opposing emo- experienced, is likely to be best remembered,
tions neutralize each other? There is a paucity of and therefore have a disproportionate effect on
research addressing this situation. But discourse overall evaluations of the web experience as well
amongst scholars within this research domain leads as the brand. It is the resultant feeling that attaches
op
us to postulate that, in the case of a series of mild to a brand, either directly via affect transfer or
emotions, these will blend into a single overall indirectly via cognitive elaboration. It therefore
feeling. Therefore, on balance, what will be the behooves a brand manager to be cognizant of these
overall feeling? The recalled memory will tend to six characteristics, and to manage them appropri-
become more extreme over time. Memories are not ately to mitigate the possibility of a lingering nega-
veridical (Hubbard, 1994). If this supposition is tive feeling. In some cases, this is best achieved by
correct, then in the case of a website that lacks removing functions within a website–—better to do a
tC

extreme emotion-laden events or a particularly function right or not do it at all.

noteworthy ending, this will work to the webmas- The array of purposes and functionalities of B2C
ter’s advantage, assuming the overall experience is websites is vast, as is the variance in capabilities and
at least mildly pleasant. interest levels of website users. Nevertheless, be-
cause interacting with a website involves multiple
activities, it should be assumed that a series of
4. Final remarks and managerial independent emotions unfolds. Techniques now
No

implications exist for tracking these moment-to-moment re-

sponses and their influence on overall evaluations.
The benefits of attaching positive feelings to Baumgartner, Sujan, and Padgett (1997) used what
a brand are well documented: it increases brand they called a feelings monitor, whereby participants
loyalty, trust and, ultimately, market share moved a cursor horizontally left or right to reflect
(Fournier, 1998; Steenkamp & Dekimpe, 1997). strong negative feelings, neutral feelings, or strong
There are many ways of encouraging positive feel- positive feelings. In that case, the technique was
ings, the most meaningful of which is for a customer used to track feeling states while watching adver-
to have a satisfying experience using the product; tisements. This technique can be modified for use in
this assumes a firm has gotten that far in the con- a website testing lab by using, for example, a
Do

sumer’s decision making process. Within this arti- galvanic skin test or EEG to provide proxy measures
cle, we have highlighted a comparatively new for feeling states, where accelerated heart rate and
approach to generating feelings toward an organi- moist skin indicate heightened arousal. Even with-
zation and its brands; namely, through an engaging out this level of sophistication it should be apparent
business-to-consumer website experience. B2C that, like advertisements, websites should be pre-
websites possess characteristics that differentiate tested for their efficacy, both in terms of the cog-
them from other forms of media communications. nitions generated and the emotions evoked. Given

This document is authorized for use only by Hemangi Bhalerao until March 2013. Copying or posting is an
infringement of copyright. [email protected] or 617.783.7860.
 Creating emotions via B2C websites 427

the heterogeneity of shopping orientations (Brown Dellaert, B. G. C., & Kahn, B. (1999). How tolerable is delay?

t
et al., 2003), pre-testing should be done on at least Consumers’ evaluations of internet web sites after waiting.
Journal of Interactive Marketing, 13(1), 41—54.
two groups: one with low to moderate levels of

os
Dijkstra, M., & van Raaij, F. (2001). Media effects by involve-
interest or involvement, and another with high ment under voluntary exposure: A comparison of television,
levels of interest or involvement. Ideally, what is print and static internet. Journal of Euromarketing, 11(2),
presented on the website will appeal to both. Low 1—21.
involvement users want quick solutions–—a Manage- Donovan, R. J., Rossiter, J. R., Marcoolyn, G., & Nesdale, A.
(1994). Store atmosphere and purchasing behavior. Journal of
ment Summary, if you will–—whereas high involve- Retailing, 70(3), 283—294.
ment individuals want to be engaged longer and Edell, J. A., & Keller, K. L. (1989). The information processing of

rP
therefore desire more content. coordinated media campaigns. Journal of Marketing Re-
We are not so presumptuous to believe that we search, 26(3), 149—163.
Eroglu, S., Machleit, K. A., & Davis, L. (2003). Empirical testing of
have presented the panacea for website success.
online store atmospherics and shopper responses. Psychology
Our objective was to integrate and elaborate upon & Marketing, 20(2), 139—150.
six characteristics relevant to websites that, to Forsyth, J. E., Lavoie, J., & McGuire, T. (2000). Segmenting the e-
date, have been little understood or overlooked. market. McKinsey Quarterly, 4, 1—4.
These characteristics can cause emotions, both pos- Fournier, S. (1998). Consumers and their brands: Developing
itive and negative, which affect evaluations of relationship theory in consumer research. Journal of Consum-

yo
er Research, 24(40), 343—373.
brands. Therefore, the insights herein have practi- Fredrickson, B. L., & Kahneman, D. (1993). Duration neglect in
cal import to brand managers, promotion coordina- retrospective evaluations of affective episodes. Journal of
tors, and IT staff responsible for website design and Personality and Social Psychology, 65(1), 45—55.
maintenance. Griffith, D. A., Krampf, R. F., & Palmer, J. W. (2001). The role of
interface in electronic commerce: Consumer involvement
with print versus on-line catalogs. International Journal of
Electronic Commerce, 5(4), 135—153.
Hoffman, D. L., & Novak, T. P. (1996). Marketing in hypermedia
References computer-mediated environments: Conceptual foundations.
op
Journal of Marketing, 60(4), 50—68.
Allen, C. (2001, January 2). Emotions trigger the right moves. The Holland, J., & Menzel-Baker, S. (2001). Customer participation in
ClickZ Network. Retrieved February 1, 2008, from http:// creating brand site loyalty. Journal of Interactive Marketing,
www.clickz.com/showPage.html?page=834621 15(4), 34—45.
Ariely, D. (2000). Controlling the information flow: Effects on Holzwarth, M., Janiszewski, C., & Neuman, M. M. (2006). The
consumers’ decision making and preferences. Journal of Con- influence of avatars on online consumer shopping behavior.
sumer Research, 27(2), 233—248. Journal of Marketing, 70(4), 19—36.
Bagozzi, R. P., & Dholakia, U. M. (2002). Intentional social action Hubbard, T. L. (1994). Memory psychophysics. Psychological Re-
tC

in virtual communities. Journal of Interactive Marketing, search, 56, 237—250.

16(2), 2—21. Kemper, T. D. (1978). A social interactional theory of emotions.
Bagozzi, R. P., Gopinath, M., & Nyer, P. (1999). The role of New York: Wiley.
emotions in marketing. Journal of the Academy of Marketing Klein, L. R. (2003). Creating virtual product experiences: The role
Science, 27(2), 184—206. of telepresence. Journal of Interactive Marketing, 17(1),
Baumgartner, H., Sujan, M., & Padgett, D. (1997). Patterns of 41—55.
affective reactions to advertisements: The integration of Koufaris, M. (2002). Applying the technology acceptance model
moment-to-moment responses into overall judgments. Jour- and flow theory to online consumer behavior. Information
nal of Marketing Research, 34(2), 219—232. Systems Research, 13(2), 205—223.
No

Brown, M., Pope, N., & Voges, K. (2003). Buying or browsing? An Leong, E. K. F., Huang, X., & Stanners, P. (1998). Comparing the
exploration of shopping orientations and online purchase effectiveness of the website with traditional media. Journal
intention. European Journal of Marketing, 37(11/12), of Advertising Research, 38(5), 44—49.
1666—1684. Levine, R., Locke, C., Searls, D., & Weinberger, D. (2000). The
Carton, S. (2002, July 15). Emotion: The next frontier in online Cluetrain manifesto: The end of business as usual. Cambridge,
advertising? The ClickZ Network. Retrieved February 1, 2008, MA: Perseus Book Group.
from http://www.clickz.com/tech/lead_edge/article.php/ Lindgaard, G., Fernandes, G., Dudek, C., & Brown, J. (2006).
1403711 Attention web designers: You have 50 milliseconds to make a
Cockton, G. (2002). Editorial: From doing to being: Bringing first impression! Behaviour and Information Technology,
emotion into interaction. Interacting with Computers, 25(2), 115—126.
14(2), 89—92. Lindstrom, M. (1999, August 26). Morphing offline into online.
Cohen, J. B., & Areni, C. S. (1991). Affect and consumer behav- Retrieved February 1, 2008, from http://www.clickz.com/
Do

iour. In T. S. Robertson & H. H. Kassarjian (Eds.), Handbook of showPage.html?page=816981

consumer behaviour (pp. 188—240). Englewood Cliffs, NJ: Machleit, K. A., & Eroglu, S. A. (2000). Describing and measuring
Prentice-Hall. emotional response to shopping experience. Journal of Busi-
Coyle, J. R., & Thorson, E. (2001). The effects of progressive ness Research, 49(2), 101—112.
levels of interactivity and vividness in web marketing sites. MacKenzie, S. B., & Lutz, R. J. (1989). An empirical examination
Journal of Advertising, 30(3), 65—78. of the structural antecedents of attitude toward the ad in an
Csikszentmihalyi, M. (1990). Flow: The psychology of optimal advertising pretesting context. Journal of Marketing, 53(2),
experience. New York: Harper and Row. 48—65.

This document is authorized for use only by Hemangi Bhalerao until March 2013. Copying or posting is an
infringement of copyright. [email protected] or 617.783.7860.
 428 M.Y. Jones et al.

McAlexander, J. H., Schouten, J. W., & Koenig, H. F. (2002). Shiv, B., & Fedorikhin, A. (1999). Heart and mind in conflict: The

t
Building a brand community. Journal of Marketing, 66(1), interplay of affect and cognition in consumer decision making.
38—54. Journal of Consumer Research, 26(3), 278—292.

os
Novak, T. P., Hoffman, D. L., & Duhachek, A. (2003). The influence Steenkamp, J. B., & Dekimpe, M. G. (1997). The increasing power
of goal-directed and experiential activities on online flow of store brands: Building loyalty and market share. Long Range
experiences. Journal of Consumer Psychology, 13(1&2), 3—16. Planning, 30(6), 917—930.
Novak, T. P., Hoffman, D. L., & Yung, Y. F. (2000). Measuring the Steuer, J. (1992). Defining virtual reality: Dimensions of deter-
customer experience in online environments: A structural mining telepresence. Journal of Communication, 42(4),
modeling approach. Marketing Science, 19(1), 22—44. 73—93.
Peracchio, L. A., & Meyers-Levy, J. (1997). Evaluating persuasion- Thorbjornsen, H., Supphellen, M., Nysveen, H., & Pedersen, P. E.
enhancing techniques from a resource matching perspective. (2002). Building brand relationships online: A comparison of

rP
Journal of Consumer Research, 24(2), 178—191. two interactive applications. Journal of Interactive Market-
Petty, R. E., & Cacioppo, J. T. (1981). Attitudes and persuasion: ing, 16(3), 17—34.
Classic and contemporary approaches. Dubuque, IA: WC Trouble brewing. (2007, March 1). The Economist. Retrieved
Brown. February 1, 2008, from http://www.economist.com/
Philport, J. C., & Arbittier, J. (1997). Advertising: Brand commu- business/displaystory.cfm?story_id=8776372
nications styles in established media and the internet. Journal Varey, C., & Kahneman, D. (1992). Experiences extended across
of Advertising Research, 37(2), 68—76. time: Evaluation of moments and episodes. Journal of Behav-
Plutchik, R. (1994). The psychology and biology of emotion. New ioral Decision Making, 5(3), 169—185.
York: Harper Collins. Watson, L., & Spence, M. T. (2007). Causes and consequences of

yo
Ross, W. T., & Simonson, I. (1991). Evaluations of pairs of experi- emotions on consumer behaviour: A review and integrative
ences: A preference for happy endings. Journal of Behavioral cognitive appraisal theory. European Journal of Marketing,
Decision Making, 4, 273—282. op 41(5/6), 487—511.
tC
No
Do

This document is authorized for use only by Hemangi Bhalerao until March 2013. Copying or posting is an
infringement of copyright. [email protected] or 617.783.7860.