ISOM 2700: Operations Management

Session 21. Supply Chain Management III

Bullwhip Effect
Lijian Lu
Dept. of ISOM, HKUST

1
 Last Class
• Responsive Supply Chain
• Make-to-stock (Push) vs Make-to-order (Pull)
• Quick response with reactive capacity

• Operational Hedging
• Flexible Manufacturing and Product Pooling

2
 Exercise: Sarah’s Wedding
Sarah is planning her wedding. She and her fiancé have signed a contract with a
caterer to tell the caterer the number of guests one month before the wedding
date. This “final number” will determine how much they pay, and they pay $60
per guest. For any additional guest without booking costs $85. (for example, if D
↓ ↓
they reserve 90 seats, they pay $5,400 even if only 80 guests show up and pay
↓ ↓
$5,570 if 92 guests show up (original 5400+2*85)). D
I
↓
S

The problem Sarah faces is that she still does not know the exact number of
guests to show up (for example, her brother may or may not bring their
girlfriend). She forecasts number of guests has a mean of 100 and could be any
number from 88 to 112 with equal likelihood.

How many guests should Sarah commit to the caterer?

Co(S =
D)
=

$60
Cu(S <
D) =
$85 $60 $25
-
=

3
 Exercise: Sarah’s Wedding
↓

• Overage cost: Co = $60

• Underage cost: Cu = $25 (=85 – 60)
• Critical fractile = 0.29
• Optimal number of guests to book: 95
PrcDIS) =
CF =
0 . 29

4
 Exercise: Sarah’s Wedding
• Suppose now the caterer is willing to refund Sarah if committed
guests do not show up. In particular, Sarah only has to pay $35
for each “no-show”.

• How many guests should they commit?

• Overage cost: Co = $35

• Underage cost: Cu = $25 (=85 – 60)
• Critical fractile = 0.42
• Optimal number of guests to book: 98

5
 This Class
• Supply Chain Management
• Information distortion: Bullwhip effects
• What is the Bullwhip effects
• Causes of the Bullwhip effects
• How to alleviate Bullwhip effects

6
 Bullwhip Effects
demand variability is amplified as it moves a

supply chain from retailers to manufacturers

Order Volatility

Customer Retailer Manufacturer Supplier

7
 History of Bullwhip

“The Bullwhip Effect in Supply Chains”, by Lee, Padmanabhan, and Whang,

in Sloan Management Review, Spring 1997.

Bullwhip Effect
https://www.youtube.com/watch?v=lfZ6I8watQk

8
 Supply Chain Demand and Variability
• Over the long run the average inflow to a firm must equal
the average outflow
Outflow
Inflow Firm (sales)

• However, the volatility of the inflow can differ

substantially from the volatility of the outflow

Inflow Outflow
Firm (sales)

9
 What is the Bullwhip Effect
• Demand variability increases as you move up the supply
chain from customers towards supply
Order Order Order Order Order

Equipment Tier 1 Supplier Factory Distributor Retailer Customer

Delivery Delivery Delivery Delivery Delivery

10
 The Bullwhip Effect
Large swings
1000
Manufacturer’s Production at the tip
800

600

400
Delivery
Manufacturer 200

33
35
37
39
41
11
13
15
17
19
21
23
25
27
29
31
1
3
5
7
9
Delivery 1000

Order 800 Wholesaler’s Orders to Manufacturer

600
Order 400 Delivery
200

0
Wholesaler

31
33
35
37
39
41
11
13
15
17
19
21
23
25
27
29
1
3
5
7
9

Delivery 1000

Order 800
Retailer’s Orders to Wholesaler
600
Order 400

200 #I
Retailer Delivery
Small (
0
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
1
3
5
7
9

Order perturbation at
Delivery 1000
the handle
800 Consumer Demand at Retailer
Order 600

Customers 400

200

0
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
1
3
5
7
9

11
Bullwhip Effect: Autos to Machine Tools
(more stable
80% Autos Machine tools

60%

40%

20%

0%

-20%

-40%

-60% GDP = solid line

-80%
Source:Anderson, Fine and Parker (1996)
12
 Bullwhip Effect: US PC Industry
Changes in
demand
80%

60%

Semiconductor
40% Equipment

20% PC

0%

-20% Semiconductor

-40%

1995 1996 1997 1998 1999 2000 2001

Annual percentage changes in demand (in $s) at three levels of the semiconductor supply chain:
personal computers, semiconductors and semiconductor manufacturing equipment.
13
 Consequences of Bullwhip Effect
• Poor customer service due to stock-outs I

• Inefficient production or excessive inventory

• High supply-demand mismatch cost

– Holding, backlogging # A I M
– Transportation
↑M
– High costs for corrections (expedited shipments and
overtime)

• Waste of production capacity ("∗ = $ + &') 14

 Causes of Bullwhip Effect
• Order synchronization

• Order batching #It

Supplier
Delivery
• Trade promotions and forward buying Order

Retailer

• Reactive and over-reactive ordering

Customers
• Rationing and shortage gaming

15
 Order Synchronization
• Customers order on the same order 70

cycle, e.g., every Monday. 60

depends on how many retailers order on the same day 50

• The graph shows simulated daily 40

Units
consumer demand (solid line) and 30
supplier demand (squares) when
20
retailers
70 order weekly and most of

them favor
60
beginning of a week. 10

0
50
Time (each period equals one day)

40
Units

30
What if daily ordering?
Note: 20 retailers with 9 order on Monday, 5
20 on Tuesday, 1 on Wednesday, 2 on Thursday
and 3 on Friday
10
cycle time o order volatility ↓
0
Time (each period equals one day) 16
 Order Batching
• Retailers may be required to order in integer multiples of
some batch size, e.g., full truck load, case quantities,
pallet quantities, etc. truck load with other retailers
70
F share 70

What if small
60 orders?
60

50
50
40

Units
40
Units

30
30
20

20
10

10
small orders f order volatility ↓
0
Time (each period equals one day)
0
Time (each period equals one day)
Note: Simulated daily consumer demand (solid line) and supplier demand (squares) when retailers order in 17
batches of 15 units, i.e., every 15th demand a retailer orders one batch from the supplier that contains 15 units.
 Trade Promotions and Forward Buying
• Supplier gives retailer a temporary discount, called a trade promotion
• Retailer purchases enough to satisfy demand until the next trade promotion
#irm wait until promotional period buy item at low price I
:
,

• Example: Campbell’s Chicken Noodle Soup over a one year period

A retailer’s purchase Total shipments and consumption

7000

6000 ① summer Gend of winter

Shipments
5000

4000
Cases

Cases
3000
Consumption
2000

1000

0
Jul
Jan

Jun

Aug

Nov
Feb

May

Sep

Oct
Apr
Mar
Dec

L L Time (weeks)
18
promotion promotion
 Reactive and Over-reactive Ordering
· believe
• How should a firm respond to a ① Random shock · actual
12
“high” demand observation? or demand shift?
10
· ⑧ ·
• Signal of higher future demand 8
N

(i.e., demand shift) or just

6
random shock (temporary)? · ⑧ ·
4
• Hedge by assuming this signals
2
higher future demand, i.e. order
more than usual 0
0 2 4 6 8 10 12
Days

• Rational reactions at one level propagate up the supply chain

• Unfortunately, it is human to over-react, thereby further
increasing the bullwhip effect

19
 -
nF - 4 #FEERITE
TE F T
Rationing and Shortage Gaming
• Setting:
– Retailers submit orders for delivery in a future period.
– Supplier produces.
– If supplier production is less than orders,
orders are rationed, i.e., retailers are “put on allocation”.

• … to secure a better allocation, the retailers inflate their orders, i.e., order
more than they need… firm's measure : consumer pay shipping fee
& cancel after several days -> penalty
• … So retailer orders do not convey good information about true demand …

• This can be a big problem for the supplier, especially if retailers are later
able to cancel a portion of the order:
– Orders that have been submitted that are likely be canceled are called phantom
orders.

20
 Strategies to Combat the Bullwhip Effect
• Information sharing
– Electronic Data Interchange (EDI): track inventory information
– Radio Frequency Identification (RFID): track demand information
• Smooth the flow of products
– Coordinate with retailers to spread deliveries evenly
– Reduce minimum batch sizes
Video: RFID
https://www.youtube.com/watch?v=gEQJxNDSKAE

– Smaller and more frequent replenishments Video: Amazon Go

• Channel alignment https://www.youtube.com/watch?v=NrmMk1Myrxc

– Eliminate forward buying (e.g., everyday low price)

– Order allocation based on past sales in case of shortage
– Restrict returns and order cancellations
• Centralized decision
– Vendor-Managed Inventory (VMI): delegation of stocking decisions
– Used by Barilla, P&G/Wal-Mart and others 21
 An Antidote to the Bullwhip Effect
• Is there any force in a supply chain that counteracts the
bullwhip effect?

• Yes: If demand is seasonal (i.e., there are anticipated peaks

and valleys in demand), then use production smoothing:
– If the firm’s orders are correlated with its production, the firm’s
suppliers will see orders that are smoother than the firm’s demand.

Demand exceeds
production, drawing
Production down inventory
Quantity

Production exceeds
demand, building
Demand inventory

22
Time
 Production Smoothing
U.S. Monthly Trade Sector Sales and Production

Inflow and Outflow of Goods to U.S. General Merchandisers (Walmart, Target, Kohl’s) 23
 Beer Game

“The Beer Game is a role-playing simulation

developed at Massachusetts Institute of Technology
(MIT) in the 1960's to clarify the advantages of
taking an integrated approach to managing the
supply chain; it particularly demonstrates the
value of sharing information across the various
supply chain components.”

http://web.mit.edu/jsterman/www/SDG/beergame.html
Playing Beer Game: http://scgames.bauer.uh.edu/
Video: Beer Game 24
https://www.youtube.com/watch?v=Pp4v0K5mRJI
 Takeaways
• Orders are interpreted as signals of demand in the
supply chain.
– Information gets distorted as it moves upstream

• Sharing demand/inventory information wins half the

battle

• Centralized decision-making achieves further

coordination
Bullwhip Effect
https://www.youtube.com/watch?v=v3WK1Sgmu50

25