Operations Management Basics: Operations as a Flow Process

C ONTROL OM: Systems

Inputs and Outputs
Info & Feedback Info & Feedback Flow
Transformation
Flow Flow
I NPUTS T RANSFORM O UTPUTS Feedback and Control

–1– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 1/2

Example from the text: P RESBYTERIAN H OSPITAL P HILADELPHIA

Inputs & Outputs = Patients

Transform = Angiography
Patients wait when they “compete” for the same limited resource.
Patients wait due to unpredictable nature of activities.

–2– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 2/2
Example: A flow unit is a single call

S WITCH C OMPLETE

Abandon

–3– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 3/2

Inputs: Calls

I NPUTS (C ALL I NITIATION )

S WITCH C OMPLETE

Abandon

–3– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 3/2
Input, transform

P RODUCT C OR VIP
I NPUTS (C ALL I NITIATION )

R OUTING RULES

B
S WITCH C OMPLETE

A, B OR C

Abandon

–3– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 3/2

Input, transform, output

P RODUCT C OR VIP
I NPUTS (C ALL I NITIATION ) O UTPUTS (C ALL R ESOLUTION )

R OUTING RULES

B
S WITCH C OMPLETE

A, B OR C

Abandon

–3– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 3/2
Input, transform, output, feedback

Feedback S UPERVISOR

P RODUCT C OR VIP
I NPUTS (C ALL I NITIATION ) O UTPUTS (C ALL R ESOLUTION )

R OUTING RULES

B
S WITCH C OMPLETE

A, B OR C

Abandon

–3– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 3/2

Input, transform, output, feedback, control

Feedback S UPERVISOR
C ONTROLS

P RODUCT C OR VIP
I NPUTS (C ALL I NITIATION ) O UTPUTS (C ALL R ESOLUTION )

R OUTING RULES

B
S WITCH C OMPLETE

A, B OR C

Abandon

–3– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 3/2
Call-center example: Typical OM tasks

Forecast Call
Volume &
Customer Behavior

Define Service
Level Goals
Forecasting
System Design
Calculate Required Requirements Planning
No. of Agents
Scheduling
Iterate: Feedback and Control
Efficiently Schedule
All Activities

Evaluate
Performance

–4– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 4/2

Call-center example: Typical OM tasks

Forecast Call
Volume &
Customer Behavior

Define Service
Level Goals
Forecasting
System Design
Calculate Required Requirements Planning
No. of Agents
Scheduling
Iterate: Feedback and Control
Efficiently Schedule
All Activities

Evaluate
Performance

–4– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 4/2
Competition, Strategy, Productivity

Organizations strive for increased productivity. Example: Call Center

Measurement and tracking of productivity goals:

1. Total call throughput
2. Number of calls handled per worker

Associated performance measures

1. 30-second average speed of answer (ASA)
2. Abandonment rate (AR) of 1% or less
3. % of callers with > 60s ASA

Systems analysis
1. ASA “much higher” during peak hours.
2. ASA “much higher” at certain centers (bottlenecks).

–5– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 5/2

Three “Little” Measures of Process Performance

Start thinking of the process (at the highest level) as a “black box”. Inputs are processed
into outputs. What is the flow unit? (e.g. Patient, phone call, manufactured item.)

1. The number of flow units in the process (system) is called the inventory.

2. The time it takes the flow unit to enter and exit the process is called the flow time.

3. The rate at which the process delivers output is called the flow rate.

Customers care about flow time! Inventory in a production environment is called work in
process (WIP). The flow rate is also commonly referred to as the throughput rate.

Process Flow Unit Flow/throughput Rate Flow Time Inventory

H&R Block Tax Returns Returns Processed 45,000 per day 0.25 days 11,250 returns
Landyatz Skateboard Manufacture Boards 1400 boards per week 0.1 weeks per board 140 boards
Expedia Travel Call center Calls 90 callers per hour 0.2 hours per call 18 callers
Business School OM301 Students 675 students per year 1/3 year 225 students.

–6– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 6/2
Simple call-center example: Flows and Little’s Law

Incoming Calls → Call Center (holding+in-service=inventory) → Serviced Calls

What (interesting) questions can we ask about a process?

Inventory: I
Flow Time: T
Flow Rate: R

LITTLE’s LAW, what is it?

Inventory(I) = Flow Rate(R) × Flow Time(T )

# of Calls in the system(I) = Call Rate(R) × Avg. time in system(T )

Implications of Little’s Law: Out of the three fundamental performance

measures (I,R,T), two can be chosen by management, and the other is
GIVEN by nature.
EXAMPLE: If we hold throughput constant: Reducing flow time → Reduces inventory
–7– c Samuel K. Eldersveld – OM301 – April 7, 2008
!

OM301 – p. 7/2

Call-center example: Flows and Little’s Law

Example: Suppose that a call center services an average of 75 callers per hour (and there
are no abandonments). Also, suppose that it takes, on average, a customer waits or is
speaking with a customer service representative for 12.8 minutes, the average time a
customer spends “in the system”. Then, we have:

12.8 minutes per call

Flow Time: T = 60 minutes per hour
= 0.213 hours per call

Flow Rate: R = 75 calls per hour.

Now, what is I?

–8– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 8/2
Call-center example: Flows and Little’s Law

Little’s Law: I =R×T

Using Little’s Law we have the average number of calls in the call center system as:

No. of Calls in the system(I) = Call Flow Rate(R) × Avg. Time in System(T )
= 75 × 0.213
= 16.

Hence, if there are M = 10 call service agents working, the expected (average) number of
callers waiting to have their call answered at any time is (I − M ) = (16 − 10) = 6 callers.

–9– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 9/2

Inventory Turns and Little’s Law

Flow rate (R) = Cost of Goods Sold (COGS).

Inventory (I) = Cost of items on hand (average inventory).

Inventory
Flow Time (T ) = Avg. time a unit is in the system. =
COGS

This is data is available from a company’s 10K report

“TURNS”: The number of times a company “turns over” its inventory

each period:

1 R COGS
TURNS = ≡ ≡
T I I

–10– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 10/2
Example of Inventory Turns (“things go better(?)”) with Coke?

The Coca-Cola Company and Subsidiaries (2004 10K report)

Compound Growth Rates Year Ended December 31,

(In millions $ except per
share data and growth rates)
5 Years 10 Years 2004 2003
SUMMARY OF OPERATIONS

Net operating revenues 5.5% 4.2% $ 21,962 $ 21,044

Cost of goods sold 4.9% 2.2% 7,638 7,762
Gross profit 5.9% 5.5% 14,324 13,282
http://www2.coca-cola.com/investors/form_10K_2004.html

Coca-Cola (KO) revenues in 2004: $21.9B

The cost of goods sold was 35% of revenue

–11– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 11/2

Inventory from financial statements: Turnover with KO

Balance Sheet - Excerpt (in $millions)
2004 2003
Inventories (avg.) $1,491 $1,252
Income Statement (Excerpt)
Cost of Goods Sold (flow rate) $7,638 $7,762

I R T

Data avg: 2003–2004: Inventory ($) COGS ($) Time (Yrs.)

I
$1.37B $7.7B = 0.178
R

Turns = Flow1Time = 1
T
COGS = COGS = 5.6.
= Inventory I

Days for coke to go through its inventory: 365

5.61 = 65.1.
Q: Why is Coke’s turn rate lower than industry beverage average (≈ 8)?

–12– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 12/2
Practice Problem (see page 28 of the course notes)

For next time: Visit a web-site such as

http://finance.yahoo.com
and search for financial statements for two “competing” companies.
Download COGS (or costs of sales) from the income statement and
inventory from the balance sheet. Compute inventory turns:

COGS
T urns =
Inventory

This is the type of preliminary research a “stock-picker” might conduct to

determine company-wide trends while performing a competitive analysis.

What does the analysis using your choices of companies tell you?

–13– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 13/2

Isn’t Inventory BAD? If so, why do we have it?

1. Pipeline Inventory
Since it takes time for an flow unit to traverse the system, at any point in time units make
up inventory in transit or WIP. Examples: Grocery items moving from distribution to retail
or a caller to a call center waiting or with a service rep.
2. Seasonal Inventory
When capacity is fixed, but demand is variable. Campbell’s Chicken Noodle Soup.
Agricultural seasonality plays a role in raw materials inventories of wheat and corn being
higher at production facilities in the fall.
3. Cycle Inventory
Economies of scale (and set-up times) often make it economical to produce products in
batches. Even when demand is constant, when a batch is completed, inventory will be
present at the start of the demand cycle.
4. Decoupling Inventory and Buffers
Inventory between process steps act as buffers: especially important when subsequent
process steps have uncertain completion times or management desires steps to operate
independently or to allow independent breaks.
5. Safety Inventory
To allow for uncertain, unpredictable (stochastic = probabilistic) demand.
–14– c Samuel K. Eldersveld – OM301 – April 7, 2008
!

OM301 – p. 14/2
Processes and flows: Five characteristic elements of a process

A process transforms inputs into outputs. Flow Flow

There are five characteristic elements of I NPUTS P ROCESS O UTPUTS
a process:

1. Inputs and outputs

2. Flow units

3. Network of activities and buffers: Process Flow Diagram (PFD)

4. Resources

5. Information structure.

–15– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 15/2

Process Components - Resources and Information structure

4. Organizational Resources: Tangible assets - two catagories:

Capital: Fixed assets such as land, buildings, facilities, equipment, IT systems.
Labor: People such as engineers, operators, Customer service reps, sales staff.

Resources facilitate the transformation of inputs into outputs.

5. Information structure: A structure that shows what information is

needed and is available in order to to perform activities or make
managerial decisions.

A business process is often modeled as a network of activities performed by

resources that transform inputs into outputs. Process flow management is the
set of managerial policies that specify how a process should be operated over
time and which resources should be allocated over time to the activities.

–16– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 16/2
Process Flow Diagrams
The process network of activities and buffers is called a process flow
diagram (PFD).

RM PH LH CFB FB

FG Bri Dis FH

Schematic diagram Process-Flow Diagram

We represent each activity with a box, each buffer, or storage area for
inventory, with a triangle and mark the direction of flow units with arrows.

–17– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 17/2

How to create a PFD

“Be the flow unit (ore)”: Take the perspective of the raw material input.
Envision the journey through the process.
Simplify the process: The PFD.

Activities: Boxes , Flows: →, Inventories/buffers: RM .

–18– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 18/2
Types of Inventories and their representation

Types of inventory:

Raw Materials (raw inputs) RM

Work-in-process W IP

Finished Goods FG

With the PFD in hand: Compute the capacity of each

resource. Determine which is the bottleneck. This
determines the overall process capacity.

–19– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 19/2

Determining Capacities
Capacity is a rate (= units processed per unit time).
It is an upper bound on effectiveness.
Our old friend “Little’s Law” (I = R × T ) comes to help us determine
capacity measurements.
Example: Circored Iron Ore processing plant as described in the text,
we determine that the Fluid-bed reactor step requires 240 minutes to
process 400 tons of inputs. Thus, the capacity per hour of this step is
400
= 1.667 tons per minute,
240
or equivalently,
400 tons 60 minutes
× = 100 tons per hour.
240 minutes 1 hour

–20– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 20/2
Bottlenecks

“Bottleneck”: The activity or resource that determines the flow

(throughput) rate of the entire process.
By speeding up a bottleneck we don’t get rid of all bottlenecks, they will
just shift over to the next slowest activity.
Flow rate determines how much the process actually does produce.

–21– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 21/2

Bottlenecks (visually)

Process capacity (the theoretical maximum). Given a set of n

resources or activities:

Process Capacity = Minimum(Capacity of Resource 1, . . . ,Capacity of Resource n)

The bottleneck of a process: weakest link = the resource with the smallest capacity.

–22– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 22/2
Exogenous factors may determine the flow rate of the process

What happens when...

...the system is starved of raw material inputs?
...if demand is less than the theoretical maximum flow of the system?

Ultimate Process Flow Rate ≤ Minimum{ Supply, Process Capacity, Demand }

Supply → → Demand

–23– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 23/2

Practice Problem: Baking Bread on Two Lines (page 32 course notes)

Bakery produces one type of bread.

The bakery operates two parallel baking lines...
... each line is equipped with a mixer , a proofer and an oven .

There is a single packaging line fed by the two baking lines.

A single RM inventory buffer feeds the two lines.

–24– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 24/2
Bread Baking II

Each baking line consists of 3 resource activities:

MIX: flour, yeast, water, and salt are first mixed.
PROOF: dough is moved to the proofer, a chamber that
encourages fermentation of dough by yeast.
BAKE: proofed dough is placed into pans are baked in the ovens.
After baking is completed on either of the bread-baking lines the loaves feed a single

W IP buffer; followed directly by a packaging activity step. When packaging is

completed, the finished products feed into a single storage area (a finished-goods buffer)
until trucks arrive to pick them up for delivery.

CAN YOU DRAW A PFD of this process?

Remember, there are two parallel lines mixing, proofing and baking...Try it!

–25– c Samuel K. Eldersveld – OM301 – April 7, 2008

!

OM301 – p. 25/2