Process Analysis

Introd
ductiion / T
The
h three
measures

Prof. Christian Terwiesch

Subway – Sitting in Front of the Store

Prof. Christian Terwiesch

Subway – Sitting in Front of the Store

25 Minutes later….

Prof. Christian Terwiesch

Subway – Sitting in Front of the Store

Prof. Christian Terwiesch

Processes: The Three Basic Measures

• Flow rate / throughput: number of flow units going through the process per unit of time

• Flow Time: time it takes a flow unit to go from the beginning to the end of the process

• Inventory: the number of flow units in the process at a given moment in time

• Flow Unit: Customer or Sandwich

Prof. Christian Terwiesch

Process Analysis: The Three Measures

Immigration department Champagne MBA program Auto company

Applications Bottle of champagne Student Car

Approved or rejected cases Bottles sold per year Graduating class Sales per year

Processing time Time in the cellar 2 years 60 days

Pending cases Content of cellar Total campus Inventory

population

Prof. Christian Terwiesch

Summary

When observing a process

process, always aim to understand the three process measures

• Flow rate / throughput: number of flow units going through the process per unit of time

• Flow Time: time it takes a flow unit to go from the beginning to the end of the process

• Inventory: the number of flow units in the process at a given moment in time

In the next session,

session we will discuss what drives these measures

We will then find out that the three measures are related to each other

Prof. Christian Terwiesch

Process Analysis

Finding the bottleneck

Prof. Christian Terwiesch

Process Analysis

In this session, we will take you INSIDE the black box

Specifically, you will learn how to:

1. Create a process flow diagram

2. Find the bottleneck of the process and determine the maximum flow rate

3 Conduct a basic process analysis

3.

Prof. Christian Terwiesch

Subway – Inside the Store

Prof. Christian Terwiesch

Drawing a Process Flow Diagram

Prof. Christian Terwiesch

Drawing a Process Flow Diagram

Customers Station 1 Station 2 Station 3

Symbols in a process flow diagram

Difference between project management and process management

Prof. Christian Terwiesch

Basic Process Vocabulary

• Processing times: how long does the worker spend on the task?

• Capacity=1/processing time: how many units can the worker make per unit of time
If there are m workers at the activity: Capacity=m/activity time

• Bottleneck: process step with the lowest capacity

• Process capacity: capacity of the bottleneck

• Flow rate =Minimum{Demand rate, Process Capacity)

• Utilization =Flow Rate / Capacity

• Flow Time: The amount of time it takes a flow unit to go through the process

• Inventory: The number of flow units in the system

Prof. Christian Terwiesch

Process Analysis

Labor productivity measures

Prof. Christian Terwiesch

Labor Productivity Measures

Bottleneck
a4 =Idle Time =Processing time
Processing Time

a2

Labor Productivity Measures

a1
• Cycle time CT= 1/ Flow Rate
a3 Directt L
Di Labor Cont t 1+p2+p3+p4
C tent=p
If one worker per resource:
Direct Idle Time=(CT-p1) +(CT-p2) +(CT-p3)

1 2 3 4 • Average lab
bor uttili
ilizatition
labor content
Review of Capacity Calculations 
labor content  direct idle time
Number of Resources i
• Capacityi =
Processing Time i
• Cost of direct labor
• Process Capacity=Min{Capacityi}
Total wages per unit of time

• Flow Rate = Min{Demand, Capacity} Flow Rate per unit of time

Flow Rate
• Utilizationi=
Capacity i
Prof. Christian Terwiesch
Example: Assembly Line with Six Stations

3 min/unit 5 min/unit 2 min/unit 3 min/unit 6 min/unit 2 min/unit

Prof. Christian Terwiesch

Insert Excel analysis of Subway line here

Prof. Christian Terwiesch

The Role of Labor Costs in Manufacturing:
The Auto Industry
100% Other
90% Overhead
Warranty
80% Quality
70%
60% Assembly and other
Labor costs
50% Purchased Parts and
40% parts and material
assemblies costs Logistics costs
30%
20%
10% Material costs
0%
Final
Fi l Includi
l ding IIncluding
l di R ll d
Rolled-up
Assembler’s Tier 1 Tier 2 Costs over
cost Costs Costs ~ 5 Tiers

• While labor costs appear small at first, they are important

- look
look relative to value added
- role up costs throughout the value chain

• Implications
- also hunt for pennies (e.g. line balancing)
- spread operational excellence through the value chain
Source: Whitney / DaimlerChrysler

Prof. Christian Terwiesch

Process Analysis

Little’s Law

Prof. Christian Terwiesch

Processes: The Three Key Metrics

Prof. Christian Terwiesch

Little’s law: It’s more powerful than you think...

What it is: Inventory (I) = Flow Rate (R) * Flow Time (T)

How to remember it: - units

Implications:
• Out of the three fundamental performance measures (I,R,T), two can be chosen by
management, the other is GIVEN by nature
• Hold throughput constant: Reducing inventory = reducing flow time

Given two of the three measures, you can solve for the third:
• Indirect measurement of flow time: how long does it take you on average to respond to an email?
You write 60 email responses per day
You have 240 emails in your inbox

Prof. Christian Terwiesch

Examples for Little’s Law Applications

In a large Philadelphia hospital, there are 10 births per day.

80% of the deliveries are easy and require mother and baby to stay for 2 days
20% of the cases are more complicated and require a 5 day stay

What is the average occupancy of the department?

Source: Graves and Little

Prof. Christian Terwiesch
Little’s law: Some remarks

Not an empirical law

Robust to variation, what happens inside the black box

Deals with averages – variations around these averages will exist

Holds for every time window

Shown by Professor Little in 1961

Prof. Christian Terwiesch

Process Analysis
Inventory Turns / Inventory costs

Prof. Christian Terwiesch

Inventory Turns

Cost of Goods sold: 20,000 mill $/year Cost of Goods sold: 25,263 mill $/year
Inventory: 391 mill $ Inventory: 2,003 mill $

Inventory Turns
Computed as: COGS
Inventory turns=
Inventory

Based on Little’
Little s la
law
w
Careful to use COGS, not revenues

Prof. Christian Terwiesch

Inventory Turns At Dell

100

90

80

70

60

50

40

30

20

10

0
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Prof. Christian Terwiesch

 Inventory Turns in Retailing
and Its Link to Inventory Costs

Inventory Cost Calculation

Compute per unit inventory

costs as:
Annual inventoryy costs
Per unit =
it Inventtory costts=
Inventory turns

Example::
Example

• Annual inventory costs=30%

• Inventory turns=6
30% per year
Per unit Inventory costs=  5%
6 turns per year

Source: Gaur, Fisher, Raman

Prof. Christian Terwiesch

Process Analysis

Buffer or Suffer

Prof. Christian Terwiesch

Simple Process Flow – A Food Truck

Food Truck Every five minutes:

- You get 0, 1, or 2 orders with equal probability
- You have a capacity of 0, 1, or 2 with equal probability
- It is not possible to make a sandwich before the order
- Customers are not willing to wait

=> How many sandwiches will you sell per five minute slot?

Prof. Christian Terwiesch

Variability Will Be a Key Factor in Waiting Time

Scenario Demand Capacity

A 0 0
B 0 1
C 0 2

D 1 0
E 1 1
F 1 2

C 2 0
H 2 1
1 2 2
Average 1 1

Why variability does not always average itself out

Buffer-or-suffer strategy

Buffering is easier in production settings than in services (make to order vs make to stock)
Preview two different models: Queue and Newsvendor

Prof. Christian Terwiesch

Difference Between Make-to-Order and Make-to-Stock

McDonald’s Subway
1. Make a batch of sandwiches 1. Customer orders
2. Sandwiches wait for customer orders 2. Customer waits for making of sandwich
3. Customer orders can filled immediately 3. Customer orders can filled w
with
ith delay
=> Sandwich waits for customer => Customer waits for sandwich

Which approach is better?

Make-to-Stock advantages include:

+ Scale economies in production
+ Rapid fulfillment (short flow time for customer order)

Make-to-Order advantages include:

+ Fresh preparation (flow time for the sandwich)
+ Allows for more
more customization
customization (you can’t hold all
all versions
of a sandwich in stock)
+ Produce exactly in the quantity demanded

Prof. Christian Terwiesch

Examples of Demand Waiting for Supply
Service Examples
 ER Wait Times: 58-year-old Michael Herrara of Dallas died of a heart attack
after an estimated 19 hours in the local Hospital ER
Some ER’s now post expected wait times online / via Apps
 It takes typically 45 days do get approval on a mortgage; Strong link

between wait times and conversion

 Waiting
W iti times
ti h att McDonald’s:
ffor drive-through
d i th M D ld’ 159 seconds;
d Long queues
deter customers to join

Production Examples
• Buying an Apple computer
• Buying a Dell computer

=>
> Make
Make-to-order
to order vs Make
Make-to-Stock

to Stock

[Link]

Prof. Christian Terwiesch

Five Reasons for Inventory

Pipeline inventory: you will need some minimum inventory because of the flow time >0

Seasonal inventory: driven by seasonal variation in demand and constant capacity

Cycle inventory: economies of scale in production (purchasing drinks)

Safety inventory: buffer against demand (Mc Donald’s hamburgers)

Decoupling inventory/ buffers: buffers between several internal steps

Source: De Groote
Prof. Christian Terwiesch
Process Analysis

Multiple flow units

Prof. Christian Terwiesch

Processes with Multiple Flow Units

Foreign Dep.
m=2
20 min/app
Contact faculty/
other persons
Foreign acc.
3 cases per hour File
Regular
11 cases per hour m=1 File
4 cases per hour EZ form 3 min/app
Department
Contact prior 1 Print invoice
m 3
m=3
employers m=1
Confirmation
15 min/app letter2 min/app

Department 2
Benchmark
gradesm=2
8 min/app

Prof. Christian Terwiesch

Approach 1: Adding-up Demand Streams

Prof. Christian Terwiesch

Approach 2: A Generic Flow Unit (“Minute of Work”)

Prof. Christian Terwiesch

Steps for Basic Process Analysis with Multiple
Types of Flow Units

1. For each resource, compute the number of minutes that the resource can
produce
2. Create a process flow diagram, indicating how the flow units go through
the process
3. Create a table indicating how much workload each flow unit is consuming
at each resource
4. Add up the workload of each resource across all flow units
Add units.
5. Compute the implied utilization of each resource as

The resource with the highest implied utilization is the bottleneck

Note: you can also find the bottleneck based on calculating capacity for each step
and then dividing the demand at this resource by the capacity
Prof. Christian Terwiesch
Processes with Attrition Loss

Where is the Bottleneck?

70/500 20/70 6/20 2/6

500 ideas

per year

Pitches Scripts Pilots Shows New

Series

Processing time 2 days 10 days 30 days 70 days 200 days

Resources 5 judges 3 script writers 2 pilot teams 2 Series crews 1 Main crew
(250 days per year)

Prof. Christian Terwiesch