22

REVIEW: HOW TO ANALYZE A PROCESS?

Resources
OPIM 404
Business Process Inputs Outputs
Analysis & Design
§ How many customers can the service handle per hour?
§ How long will it take to serve a customer?
Analyzing Process Flows – § What change is needed to expand capacity?
§ How much does the process cost?
Deterministic – Part 2

Tevhide Altekin

22-1

OUTLINE Recall: Resources and Resource Pools

§ Part 1: § In every process, activities are performed by capital and labor
− Recall: Business Process resources
− Each activity may require one or more resources
− Process Flowcharting
− Each resource may be allocated to one or more activities
− Process Flowcharting as Asking Questions
− Measuring Process Flows § Resource pool is a collection of interchangeable resources
that can perform an identical set of activities
§ Note: Through cross-training and process improvement might
§ Part 2:
be possible to make the separate resource pools to handle
− Review: How to Analyze a Process tasks performed by each other à Resource pooling
− Little’s Law − Involve combining of separate resource pools flexible to handle tasks
performed by each other.
− Powerful operational concept à Can significantly affect the flow rate,
capacity and flow time
 22-2 22-3

Recall: Load Batch Effect of Product Mix on Unit Load

§ Resources may process several flow units simultaneously § Capacity of a process is a function of the product mix
during the operation time § Unit Load for Product Mix à Average the unit loads of
§ Oven baking 10 loaves of bread in a tray simultaneously à the individual products using the weights determined in
− Load batch = 10 loaves the product mix
− Example: Health Maintenance Organization
§ A machine packing breads one by one
• Process: Physician claims & hospital claims
− Load Batch = 1 loaf • Product mix: 60% Physician claims & 40% hospital claims
§ A recipe for 100 loaves of bread
− Load Batch = 100 loaves
Unit Load Unit Load Unit Load (60% - 40% Mix)
§ The higher the load batch, the higher the capacity. Resource Pool (Physician) (Hospital) (min/claim)
− Why? Mailroom clerk 0.6 min/claim 1.0 min/claim 0.6x60% + 1.0x40% = 0.76
# of Identical Resources Data-entry clerk 4.2 min/claim 5.2 min/claim 4.2x60% + 5.2x40% = 4.60
60 (min/hr) Load
CapacityResourcePool = x x in Resource Pool
(units/hr) Unit Load (min) Batch Claims processor 6.6 min/claim 7.5 min/claim 6.6x60% + 7.5x40% = 6.96
Claims supervisor 2.2 min/claim 3.2 min/claim 2.2x60% + 3.2x40% = 2.60

22-2* 22-4

Unit Load Example – Recall: Flow Time, Flow Rate

Health Maintenance Organizations
§ Flow Time: Average time for a unit to move through the
Mailroom Data Entry Initial Inspection Final system. Includes
(Arriving claims (check, if OK Processing Processing
are opened & store, o/w − Value added time: “Considers” time unit spent actually
date stamped) send back) being worked on (Sum of activity operation times if
0.6 min/claim 4.2 min/claim 4.8 min/claim 2.2 min/claim 1.8 min/claim activities are in series)
Mailroom Data-entry Claims Claims Claims
− Time spent waiting in buffer (i.e. waiting in a queue)
clerk clerk processor supervisor processor
Assume there is one unit of each resource and claims are processed one by one § Flow Rate: Output rate the process is expected to produce
over time
Unit Load (Tp)
− Flow Rate = Min{Demand, Supply, Process Capacity}
Resource Pool (p) (min/claim)
Mailroom clerk 0.6
Data-entry clerk 4.2 ... ... ... ... ...

Claims processor 4.8+1.8 = 6.6 Flow Rate

Claims supervisor 2.2 [units/min]
Flow Time [min]
 23 25

Cycle Time Operator Loading Chart

§ Cycle Time à Average time between completions of two § Sometimes, we are given a goal of T time units to
successive units (=1/Flow Rate) produce all the units demanded.
Cycle Time [min] § We can check whether that goal can be met or not in 3
... ... ... ... ...
different ways.
− Process Capacity >= Demand: Goal can be met
Flow Rate
[units/min] − Cycle Time <= Takt Time: Goal can be met
Flow Time [min]
− Time required to
§ What is the difference between cycle time and flow time? complete one cycle at
each stage <= Takt time:
Goal can be met
• Draw and check
Operator Loading Chart

24 26

Takt Time LITTLE’S LAW

§ An indication of how fast processes must be in order to § Flow unit

meet demand in allotted time − A customer; a dollar; a ton of steel; an R&D project
§ Amount of time available / Units demanded − A bank transaction to be processed; a student; a shopper
§ Suppose you had to produce 600 units in 80 hours to § Three questions:
meet the demand requirements of a product. What is − On average, how many flow units pass through the
the takt time? process per unit time?
In 80 hours we have 4,800 minutes − On average, how much time does a typical flow unit
spend within process boundaries?
(60 minutes/hr x 80 hrs)
− On average, how many flow units are within process
Takt time = 4,800 min/600 units boundaries at any given time?
= 8 min btw consecutive units § Answers: Average flow rate (throughput rate),
flow time (throughput time), inventory
 26 28
Analyzing Processes with Little’s Law
Little’s Law
§ Recall: We assume system is in steady state (Ri ≈ Ro)
Inventory I − We know the flow rate (R)
... ... ... [units] ... ...
Flow Rate R § Recall: A process is a network
[units/hr] of activities & buffers
Flow Time T [hrs] § Thus, we can decompose any process:
− For the queues (buffers)
§ Assumes the process is stable
ž It is easy to observe number of units in queue (I), and
(Input Rate ≈ Output Rate)
ž We know the flow rate (R)
§ Little’s Law ž Using Little’s Law find the average waiting time as T = I / R
Inventory = Flow Time x Flow Rate − For the activities
I = T x R ž Average operation times (activity durations) are known (T)
ž We know the flow rate (R)
§ Applications (Manufacturing, Waiting lines, Invoice
ž Using Little’s Law find the average number of units being
processing, Legal office transactions, Accounts processed (I) as I = T x R
receivable processing)
27 29
Little’s Law Example 2
Little’s Law Example 1
Consider an automobile factory 8,000 batteries
A fast-food restaurant processes an average of 5,000 lb. with flow rate R = 25 cars/hours Automobile
RM
of hamburgers per week. Assume 1 battery per automobile Manufacturing
Each battery costs $45 12 hours
The typical inventory of raw meat is 2,500 lb.
Assume a stable process
What is the average hamburger’s flow time? Average time a battery spends in raw materials (RM) inventory?
Inventory I Average number of batteries in the automobile factory?
... ... ... 2,500 lb ... ... Average worth of batteries in the automobile plant?
Flow Rate R Recall: R I T
5,000 lb/week
Little’s Law Flow Rate Inventory Average Time
Flow Time T I=TxR (cars/hour) (batteries) (hours)
RM Inventory 25 8,000 8,000/25 = 320
I = 2,500 lb
Automobile Manufacturing 25 12 x 25 = 300 12
R = 5,000 lb/wk
Average Number of Batteries in the Plant = 8,300 batteries
Little’s Law: I = T x R Average Worth of Batteries in Plant = 8,300 batteries x $45/battery
2,500 lb = T x 5,000 lb/wk T = 0.5 weeks = $373,500
 30 31

Analyzing Processes with Little’s Law Little’s Law Example 3

Consider an upscale hotel with Leisure Travelers
§ Recall: We assume system is in steady state (Ri ≈ Ro) flow rate R = 135 guests/night 3.6 nights $210/night
− We know the flow rate (R) Two guest types:
Business Travelers
§ Hence, we have conservation of the flow rate (R) Leisure Travelers
Business Travelers 1.8 nights $250/night
R1
R2 R = R1 + R2 Average number of guests of each type?
Average total revenue per night?
Ri = R1 + R2 Ro = R1 + R2 Average fee per night per guest?
1=α+β+δ
Recall: T R I
Average Time Flow Rate Inventory
R1 = α x R Little’s Law
R1 = 0.4 x R α I=TxR (nights) (guests/night) (guests)
R R R2 = β x R Leisure Travelers 3.6 1/3 x 135 = 45 45 x 3.6 = 162
R2 = 0.6 x R β Business Travelers 1.8 2/3 x 135 = 90 90 x 1.8 = 162
δ
Ro = 0.4 x R + 0.6 x R R2 = δ x R Average Number of Guests = 324

Average Total Revenue per Night = 162 x $210 + 162 x $250 = $74,520
Ri = R Ro = R Ri = R Ro = R
Average Fee per Night per Guest = $74,520 / 324 = $230

30* 32
Analyzing Processes Example 4
Analyzing Processes Using Little’s Law
§ Consider the insurance claims processing example
§ We assume the system is in steady state introduced earlier
− A system is in equilibrium if the average number of flow − Number of claims processed per day, on average, is 29.9
units in the system stays the same over time. à R = 29.9 claims/day
− For a system to be in equilibrium, the average rate at − Time to process a claim, on avg., is 10 days à T = 10
which flow units enter into the system must be days
approximately equal to the average rate at which flow
units leave the system (i.e. Ri ≈ Ro). − How many claims are there in the system à I = ?
− Little’s Law à I =T x R = 10 days x 29.9 claims/day
§ By extension, for a system with multiple stages to
be in equilibrium, the throughput rate at each stage I = 299 claims
must be the same. § Suppose the customers are unhappy with the average
throughput time of 10 days.
§ Thus, we have conservation of throughput rate
§ Management realizes that in order to maintain the
R1 R1 = 0.4 x R current level of demand, the firm will need to improve
40%
R2 R = R1 + R2 responsiveness.
R R2 = 0.6 x R
60%
 33 35
Analyzing Processes Example 4 (Cont’d) Analyzing Processes Example 4 (Cont’d)
§ The management team is aware that Recall: R I T
Little’s Law Thr. Rate Inventory Average Time
− 80% of the claims are fairly standard I=TxR (claims/day) (claims) (days)
− 20% of the claims require complex evaluation Registration 29.9 60 60 / 29.9 = 2.01
§ Standard claims are more profitable. Standard 0.8 x 29.9 = 40 40 / 23.92 = 1.67
Claims
§ In the current process both types of claims are routed Underwriting
23.92
through the same process. Complex
0.2 x 29.9 = 20 20 / 5.98 = 3.34
R = 29.9 claims/day Claims
5.98
Underwriting
I = 299 claims Rating 29.9 50 50 / 29.9 = 1.67
Claims Policy Policy Writing 29.9 30 30 / 29.9 = 1.00
R Registration Rating
Underwriting Writing

T = 10 days
§ Management decides to split the claims underwriting
into two distinct groups by hiring a standard claims
underwriter.
34 35-2
Analyzing Processes Example 4 (Cont’d) Analyzing Processes Example 4 (Cont’d)
Recall: R I T
Standard Little’s Law Thr. Rate Inventory Average Time
80%
Claims I=TxR (claims/day) (claims) (days)
R = 29.9 claims/day
Underwriting
Registration 29.9 60 60 / 29.9 = 2.01
R ISCU = 40 claims Policy Standard
Registration Rating 0.8 x 29.9 = 40 40 / 23.92 = 1.67
Writing Claims 23.92
IREG = 60 claims IRAT = 50 claims Underwriting
Complex IPW = 30 claims
Claims Complex 20 20 / 5.98 = 3.34
20% 0.2 x 29.9 =
Underwriting Claims
Average number of claims in 5.98
Underwriting
ICCU = 20 claims the process: I = 200 claims
Rating 29.9 50 50 / 29.9 = 1.67
§ After trying out the new system for a few months, the Policy Writing 29.9 30 30 / 29.9 = 1.00
management finds the average number of claims at Avg. Flow Time for Standard Claims = 2.01 + 1.67 + 1.67 + 1.00 = 6.35 days
each stage, which are either
− In process or
− Waiting.
§ Have these changes resulted in a better service?
 35-3 36
Analyzing Processes Example 4 (Cont’d) Analyzing Processes Example 4 (Cont’d)
Recall: R I T Avg. FT S = 6.35 days
Little’s Law Thr. Rate Inventory Average Time Standard
80% Avg. FT C = = 8.02 days
I=TxR (claims/day) (claims) (days) Claims
R = 29.9 claims/day
Underwriting
Registration 29.9 60 60 / 29.9 = 2.01
Standard R ISCU = 40 claims Policy
0.8 x 29.9 = 40 40 / 23.92 = 1.67 Registration Rating
Claims Writing
23.92
Underwriting IREG = 60 claims IRAT = 50 claims
Complex IPW = 30 claims
Complex 20 20 / 5.98 = 3.34 Claims
0.2 x 29.9 = 20%
Claims Underwriting
5.98 Average number of claims in
Underwriting
ICCU = 20 claims the process: I = 200 claims
Rating 29.9 50 50 / 29.9 = 1.67
Policy Writing 29.9 30 30 / 29.9 = 1.00 § Clearly, with the new system the overall responsiveness
Avg. Flow Time for Complex Claims = 2.01 + 3.34 + 1.67 + 1.00 = 8.02 days has improved.
− Previous average flow time was 10 days
− New average flow time is 6.7 days:
− 33.3% reduction in flow time (36.5% for profitable S claims)
− Responsiveness improved for both (S=6.35 & C=8.02 days)

35-4
Analyzing Processes Example 4 (Cont’d)
Recall: R I T
Little’s Law Thr. Rate Inventory Average Time
I=TxR (claims/day) (claims) (days)
Registration 29.9 60 60 / 29.9 = 2.01
Standard 0.8 x 29.9 = 40 40 / 23.92 = 1.67
Claims 23.92
Underwriting
Complex 20 20 / 5.98 = 3.34
0.2 x 29.9 =
Claims
5.98
Underwriting
Rating 29.9 50 50 / 29.9 = 1.67
Policy Writing 29.9 30 30 / 29.9 = 1.00

Avg. Flow Time for Standard Claims = 2.01 + 1.67 + 1.67 + 1.00 = 6.35 days
Avg. Flow Time for Complex Claims = 2.01 + 3.34 + 1.67 + 1.00 = 8.02 days
Avg. Flow Time for Claims = 0.8 x 6.35 + 0.2 x 8.02 = 6.684 days
Check: Overall I = 200 claims R = 29.9 claims/day T = I / R
Avg. Flow Time for Claims: T = 200 / 29.9 = 6.689 days