You are on page 1of 58

Supply Chain Management

Managing the Supply Chain


Key to matching demand with supply
Cost and Benefits of inventory

Economies of Scale
Inventory management of a retailer: EOQ + ROP
Levers for improvement

Safety Stock
Hedging against uncertainty
Role of lead time

Improving Performance
Centralization & Pooling efficiencies
Postponement
Accurate Response

What is a Supply Chain?


The Procurement
or supply system
Raw Material
supply points Movement/
Transport

The Operating
System

Raw Material
Movement/
Storage
Transport

The Distribution System


Manufacturing
PLANT 1

PLANT 2

PLANT 3

Movement/
Transport

Finished Goods
Movement/
Storage
Transport
WAREHOUSE

WAREHOUSE

WAREHOUSE

C
MARKETS

What makes a good Supply Chain?

Corporate Finance
CISCO

SELECTRON

Current Assets

BS/ 2000

BS/2000

Cash &
Equivalents

4,234

1,475.5

Short-term investments

1,291

958.6

AR

2,299

2,146.3

Inventories

1,232

Deferred Income

2,054

260.5

Total Current Assets

11,110

8,628.2

%11

3,787.3

%44

Inventories represent about %34 of current assets for a typical


company in the US; %90 of working capital.
For each dollar of GNP in the trade and manufacturing sector,
about 40cents worth of inventory was held.
Average logistics cost = 21c/dollar = %10.5 of GNP.

Solectrons executive officer


compensation plan
Base compensation
Bonuses
The Bonus Plan provides for incentive compensation ...
based on certain worldwide, site and individual
performance measures. Worldwide and site performance
are measured based on targets w.r.t. profit before taxes,
inventory turns, days sales outstanding and return on assets.
The Compensation Committee believes that these factors
are indicative of overall corporate performance and
shareholder value. [compensation committee report FY 96]

Long Term Incentive Compensation

Supply Chain Metrics

Costs of not Matching


Supply and Demand
Cost of overstocking
liquidation
obsolescence
holding

Cost of under-stocking
lost sales and customer goodwill
lost margin

Why hold Inventory?


Economies of scale
Fixed costs of ordering/manufacturing
Quantity discounts
Trade Promotions

Uncertainty
Information Uncertainty
Supply/demand uncertainty

Seasonal Variability
Strategic
Flooding, availability

Cycle/Batch stock

Safety stock
Seasonal stock
Strategic stock

Cost of Inventory
Physical holding cost
(out-of-pocket)
Financial holding cost
(opportunity cost)
Low responsiveness
to demand/market
changes
to supply/quality changes

Holding cost (H)

Costs Associated with Batches


Ordering costs (S)
Changeover of production line (Set-up)
Transportation (Delivery)
Receiving

Holding costs (H = r C)
Physical holding cost
Cost of capital (r)
Cost of obsolescence

Pal Gear: Retail Inventory Management


Annual jacket revenues at a Pal Gear retail store are roughly
$1M. Pal jackets sell at an average retail price of $325, which
represents a mark-up of 30% above what Pal Gear paid its
manufacturer. Being a profit center, each store made its own
inventory decisions and was supplied directly from the
manufacturer by truck. For each order up to 5000 jackets, the
manufacturer charges a flat fee of $2,200 for delivery. To
exploit economies of scale, stores typically orders 1500 jackets
each time it places an order. (Pals cost of capital is
approximately 20%.) What order size would you recommend
for a Pal store in current supply network?

Pal Gear: Evaluation of current policy

of ordering 1500 units each time


1. What is average inventory I?
I=
Annual cost to hold one unit H =
Annual cost to hold I =

2. How often do we order?


Annual throughput R =
# of orders per year =
Annual order cost =

3. What is total cost?


TC =

Can we do better ?

Find the most economical order quantity


Method 1 : Enumerate
Number of units Number of
per order/batch Batches per
Q
Year: R/Q
50
62
100
31
150
21
200
15
250
12
300
10
350
9
400
8
450
7
500
6
510
6
520
6
530
6
540
6
550
6
600
5
650
5
700
4
750
4
800
4
850
4
900
3
1000
3

Annual
Setup Cost
$ 135,385
$
67,692
$
45,128
$
33,846
$
27,077
$
22,564
$
19,341
$
16,923
$
15,043
$
13,538
$
13,273
$
13,018
$
12,772
$
12,536
$
12,308
$
11,282
$
10,414
$
9,670
$
9,026
$
8,462
$
7,964
$
7,521
$
6,769

Annual
Holding Cost
$
1,250
$
2,500
$
3,750
$
5,000
$
6,250
$
7,500
$
8,750
$
10,000
$
11,250
$
12,500
$
12,750
$
13,000
$
13,250
$
13,500
$
13,750
$
15,000
$
16,250
$
17,500
$
18,750
$
20,000
$
21,250
$
22,500
$
25,000

Annual
Total Cost
$ 136,635
$
70,192
$
48,878
$
38,846
$
33,327
$
30,064
$
28,091
$
26,923
$
26,293
$
26,038
$
26,023
$
26,018
$
26,022
$
26,036
$
26,058
$
26,282
$
26,664
$
27,170
$
27,776
$
28,462
$
29,214
$
30,021
$
31,769

$160,000

Setup Cost

$140,000

Holding Cost

$120,000

Total Cost

$100,000
$80,000
$60,000
$40,000
$20,000
$0

100 200 300 400 500 600 700 800 900 1000
Order (batch) size Q

Economies of Scale:
Inventory Build-Up Diagram
R: Annual demand rate
(units/yr),
Q: Number of jackets per
replenishment order
Number of orders per
year = R/Q.
Average number of
jackets in inventory =
Q/2 .

Inventory

T = Q/R

Q
R = Demand rate

T
time between
orders
Order placed,
Order placed,
arrives immediately
Lead time=0

arrives immediately
Lead time=0

Time t

Economic Order Quantity


EOQ
R
S
H
Q

Demand per year,

Setup or Order Cost ($/setup; $/order),

Marginal annual holding cost ($/per unit per year), H

Order quantity.

2SR
QEOQ
H
C
r

Cost per unit ($/unit),

Cost of capital (% / yr),

=rC
Total annual
costs

2 SRH
H Q/2: Annual
holding cost

S R /Q:Annual
setup cost

QEOQ

Batch Size Q

What do we learn from the EOQ formula ?


How does the ordering policy change if
The product is a success and the demand picks up,
now we are selling 4 times the original demand
The interest rates double up, so does our unit
inventory holding costs
After investing in IT, we manage to reduce our fixed
ordering cost by half

Optimal Economies of Scale:


For a Pal Gear retailer
R = 3077 units/ year
r = 0.20/year

C = $ 250 / unit
S = $ 2,200 / order

Unit annual holding cost = H = $50/unit-yr


Optimal order quantity = EOQ = 520 units
Number of orders per year = R/Q = 3077/520 = 5.91
Time between orders = Q/R = 8.78 weeks
Annual order cost = (R/Q)S = $13,018/yr
Average inventory I = Q/2 = 260 units
Annual holding cost = (Q/2)H = $13,018/yr

Take-Aways I
Batching & Economies of Scale
Increasing batch size of production (or purchase) increases
average inventories (and thus cycle times).
Average inventory for a batch size of Q is Q/2.
The optimal batch size trades off setup cost and holding cost.
To reduce batch size, one has to reduce setup cost (time).
Square-root relationship between Q and (R, S):
If demand increases by a factor of 4, it is optimal to
increase batch size by a factor of 2 and produce (order)
twice as often.
To reduce batch size by a factor of 2, setup cost has to be
reduced by a factor of 4.

Role of Supply Leadtime L:


Pal Gear cont.
The lead time from when a Pal Gear retailer places an
order to when the order is received is two weeks. If
demand is stable as before, when should the retailer
place an order?
I-Diagram:
ROP
Order 2wks
placed

ROP =

Order
arrives

Two Key Decisions of Inventory


Management
1.

How much to order ?


Answer: EOQ

2. When to place an order ?


Answer: ROP Reorder Point
represents the amount of inventory on hand when we place a new order .
Delivery Lead Time = 0 (Instantaneous Delivery) then ROP = .
Delivery Lead Time > 0 then ROP = .
ROP is driven by:
Delivery Lead Time
Demand Uncertainty
Customer Service Level

A Key to Matching Supply and


Demand
When would you rather place your bet?

A
A:
B:
C:
D:

B
A month before start of Derby
The Monday before start of Derby
The morning of start of Derby
The winner is an inch from the finish line

C D

Demand uncertainty and


forecasting
Year
1
2
3
4
5
6

Demand
323
258
303
304
284
285

Forecast Error

Demand uncertainty and forecasting


Forecasts depend on
historical data
market intelligence
Forecasts are usually (always?) wrong.
A good forecast has at least 2 numbers (includes a
measure of forecast error, e.g., standard deviation).
The forecast horizon must at least be as large as the
lead time. The longer the forecast horizon, the less
accurate the forecast.
Aggregate forecasts tend to be more accurate.

Pal Gear:
Service levels & Inventory management

In reality, a Pal Gear stores demand fluctuates from week to


week. In fact, weekly demand at each store had a standard
deviation of about 30 jackets assume normally distributed.
Recall that average weekly demand was about 60 jackets; the
order lead time is 2 weeks; fixed order costs are $2,200/order
and it costs $50 to hold one jacket in inventory during one year.

Questions:
1. If the retailer uses the ordering policy discussed before
(ROP =120), what will the probability of running out of
stock in a given cycle be?
2. The Pal retailer would like the stock-out probability to be
smaller. How can she accomplish this?
3. Specifically, how does it get the service level up to 95%?

Example: say we increase ROP to


140 (and keep order size at Q = 520)
1. On average, what is the stock level when the
replenishment arrives?
2. What is the probability that we run out of stock
before a delivery arrives?

3. How do we get that stock-out probability down to


5%?

Lead Time Demand


Consider the lead time between the placement of an
order and the arrival of the order. L=2 weeks.
Demand per week, R ~ (60, 30)
The mean demand during the lead time = 2*60 = 120
Standard deviation of demand during lead time

(30)2 + (30)2

30

D lead time ~ N (120, 30

IF ROP=120
Probability of stock out (i.e.,
Probability that the demand during
lead time is greater than 120 ) = ?

Lead time
demand distribution

120
Mean Demand
During Lead Time

SAFETY STOCK
Safety stock increase
with the service level.

Increase reorder point


above average demand
during lead time, lower
the probability of
stockout! BETTER
SERVICE LEVEL!

e.g., % 95

ROP

Less demand variability


Small variance / SD
%95

120

Higher Variability
Larger SD/ variance

ROP

Safety stock increases


with the variance (SD) of lead time demand

%95

120

ROP

Safety Stock
I s z * R L
Safety stock increases (decreases) with an increase
(decrease) in:
demand variability or forecast error,
delivery lead time for the same level of service,
delivery lead time variability for the same level of
service.

Pal Gear: Determining the required


Reorder Point for 95% service
DATA:
R = 60 jackets/ week

R = 30 jackets/ week

H = $50 / jacket, year


S = $ 2,200 / order

standard deviation of weekly demand


L = 2 weeks

QUESTION:

What should safety stock be to insure a desired cycle service level of 95%?
ANSWER:
1. Determine lead time demand

42.42

2. Required # of standard deviations z*


= 1.64
3. Reorder Point = 120+1.64*42.42= 190 jackets
4.

Safety stock

Is = 1.64*42.42=70 jackets

Review of Probability
The standard normal distribution F(z)
Transform X = N() to

z = N(0,1)
z = (X - ) / .
F(z) = Prob ( N (0,1) < z)

F(z)
0

Transform

back,

knowing z*:
X* = + z*.

z
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
2.0
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
3.0
3.1
3.2
3.3

0.00
0.5000
0.5398
0.5793
0.6179
0.6554
0.6915
0.7257
0.7580
0.7881
0.8159
0.8413
0.8643
0.8849
0.9032
0.9192
0.9332
0.9452
0.9554
0.9641
0.9713
0.9772
0.9821
0.9861
0.9893
0.9918
0.9938
0.9953
0.9965
0.9974
0.9981
0.9987
0.9990
0.9993
0.9995

0.01
0.5040
0.5438
0.5832
0.6217
0.6591
0.6950
0.7291
0.7611
0.7910
0.8186
0.8438
0.8665
0.8869
0.9049
0.9207
0.9345
0.9463
0.9564
0.9649
0.9719
0.9778
0.9826
0.9864
0.9896
0.9920
0.9940
0.9955
0.9966
0.9975
0.9982
0.9987
0.9991
0.9993
0.9995

0.02
0.5080
0.5478
0.5871
0.6255
0.6628
0.6985
0.7324
0.7642
0.7939
0.8212
0.8461
0.8686
0.8888
0.9066
0.9222
0.9357
0.9474
0.9573
0.9656
0.9726
0.9783
0.9830
0.9868
0.9898
0.9922
0.9941
0.9956
0.9967
0.9976
0.9982
0.9987
0.9991
0.9994
0.9995

0.03
0.5120
0.5517
0.5910
0.6293
0.6664
0.7019
0.7357
0.7673
0.7967
0.8238
0.8485
0.8708
0.8907
0.9082
0.9236
0.9370
0.9484
0.9582
0.9664
0.9732
0.9788
0.9834
0.9871
0.9901
0.9925
0.9943
0.9957
0.9968
0.9977
0.9983
0.9988
0.9991
0.9994
0.9996

0.04
0.5160
0.5557
0.5948
0.6331
0.6700
0.7054
0.7389
0.7704
0.7995
0.8264
0.8508
0.8729
0.8925
0.9099
0.9251
0.9382
0.9495
0.9591
0.9671
0.9738
0.9793
0.9838
0.9875
0.9904
0.9927
0.9945
0.9959
0.9969
0.9977
0.9984
0.9988
0.9992
0.9994
0.9996

0.05
0.5199
0.5596
0.5987
0.6368
0.6736
0.7088
0.7422
0.7734
0.8023
0.8289
0.8531
0.8749
0.8944
0.9115
0.9265
0.9394
0.9505
0.9599
0.9678
0.9744
0.9798
0.9842
0.9878
0.9906
0.9929
0.9946
0.9960
0.9970
0.9978
0.9984
0.9989
0.9992
0.9994
0.9996

0.06
0.5239
0.5636
0.6026
0.6406
0.6772
0.7123
0.7454
0.7764
0.8051
0.8315
0.8554
0.8770
0.8962
0.9131
0.9279
0.9406
0.9515
0.9608
0.9686
0.9750
0.9803
0.9846
0.9881
0.9909
0.9931
0.9948
0.9961
0.9971
0.9979
0.9985
0.9989
0.9992
0.9994
0.9996

0.07
0.5279
0.5675
0.6064
0.6443
0.6808
0.7157
0.7486
0.7794
0.8078
0.8340
0.8577
0.8790
0.8980
0.9147
0.9292
0.9418
0.9525
0.9616
0.9693
0.9756
0.9808
0.9850
0.9884
0.9911
0.9932
0.9949
0.9962
0.9972
0.9979
0.9985
0.9989
0.9992
0.9995
0.9996

0.08
0.5319
0.5714
0.6103
0.6480
0.6844
0.7190
0.7517
0.7823
0.8106
0.8365
0.8599
0.8810
0.8997
0.9162
0.9306
0.9429
0.9535
0.9625
0.9699
0.9761
0.9812
0.9854
0.9887
0.9913
0.9934
0.9951
0.9963
0.9973
0.9980
0.9986
0.9990
0.9993
0.9995
0.9996

0.09
0.5359
0.5753
0.6141
0.6517
0.6879
0.7224
0.7549
0.7852
0.8133
0.8389
0.8621
0.8830
0.9015
0.9177
0.9319
0.9441
0.9545
0.9633
0.9706
0.9767
0.9817
0.9857
0.9890
0.9916
0.9936
0.9952
0.9964
0.9974
0.9981
0.9986
0.9990
0.9993
0.9995
0.9997

Safety Stocks
Inventory on hand
I(t)

EOQ

order

EOQ
order

order

ROP
R

mean demand during


supply lead time:

= R L
Is

safety stock s

Time t

Comprehensive Financial Evaluation:


Inventory Costs of Pal Gear
1. Cycle Stock (Economies of Scale)
1.1 Optimal order quantity
1.2 # of orders/year
=

1.3 Annual ordering cost per store

= $13,009

1.4 Annual cycle stock holding cost.

= $13,009

2. Safety Stock (Uncertainty hedge)


2.1 Safety stock per store = 70
2.2 Annual safety stock holding cost

= $3,500

3. Total Costs for 5 stores = 5 (13,009 + 13,009 + 3,500)


= 5 x $29,500 = $147.5K.

How to find service level (given ROP)?


How to find re-order point (given SL)?

L
= Supply lead time,
D =N(RR) = Demand per unit time is normally distributed
with mean R

and

DL =N() = Demand during the lead time is normally distributed


with mean = RL

1.

2.

standard deviation R

and

standard deviations RL

Given ROP, find SL = Cycle service level = P(stock out)


= P(demand during lead time >ROP)
= 1-F(z*= (ROP- )/)
[use table]
= 1- NORMDIST(ROP, , , True) [or Excel]
Given SL, find ROP = + Is
= + z* [use table to get z* ]
= NORMINV(1-SL, , )
[or Excel]
Safety stock Is

= z* Reorder point ROP = + Is

Take-Aways II
Demand Uncertainty & Inventory Management
Cycle Cost : EOQ: How much to order?
Balance the fixed costs of ordering with the average
inventory holding cost.
Total Cycle Cost are quiet robust to misestimating the
parameters
Growth brings scale economies, adjust operating policies as
markets conditions change.
Safety Stock Cost: Reorder Point : When to place an order?
Uncertainty is nothing but forecasting error.
Determined by the Service Level P(D(lead time)>ROP)
ROP=RxL+Safety Stock

Improving The Supply Chain


How can we reduce supply chain costs without sacrificing
customer service?
How can we improve customer service without increasing
supply chain costs?

Distribution Centralization
Product Postponement (HP)
Process Postponement (Benetton)
Capacity Analysis (Benetton)

Improving Supply Chain Performance:


1. The Effect of Pooling/Centralization
Decentralized Distribution

Is=100

Centralization Distribution

Is=100

I =400
Is=
400
s

Is=100

Is=100

Centralized vs. Decentralized Distribution


Pros

Cons

Pal Gears Internet restructuring:


Centralized inventory management

Weekly demand per store


= 60 jackets/ week
with standard deviation = 30 / week
H = $ 50 / jacket, year S = $ 2,200 / order
Supply lead time L = 2 weeks
Service level : 95% availability.

Pal Gear now is considering restructuring to an

Internet store. So 5 local stores will be closed and a


National DC will be opened to distribute direct to
customers.

Compare the safety stock in the decentralized


and centralized systems
Decentralized
Demand R per week
for each store

s
s

Demand during Supplier


lead time (L=2)

ltd
ltd
Safety Stock for each store
(%95 availability)

Centralized 5 stores
Demand R per week
for the centralized warehouse

c
c

Demand During Supplier Lead Time

(L=2)
ltd
ltd
Safety Stock for the centralized warehouse

(%95 availability)
Total Safety Stock

Pal Gears Internet restructuring:


comprehensive financial inventory evaluation
1. Cycle Stock (Economies of Scale)
1.1 Optimal order quantity
=
1.2 # of orders/year
=
1.3 Annual ordering cost of e-store

= $29,089

1.4 Annual cycle stock holding cost

= $29,089

2. Safety Stock (Uncertainty hedge)


2.1 Safety stock for e-store=
2.2 Annual safety stock holding cost
3. Total Costs for consolidated e-store
= $65,980

= $7,800.
= 29,089 + 29,089 + 7,800

Learning Objectives:
centralization/pooling
Different methods to achieve pooling efficiencies:

Physical centralization
Information centralization
Specialization
Raw material commonality (postponement/late
customization)

Cost savings are sqrt(# of locations pooled).

2. Postponement & Commonality (HP


Laserjet)
Generic Power
Production

Unique Power
Production

Transportation
Europe

Process I: Unique
Power Supply

N. America

Europe

Process II: Universal


Power Supply

Make-to-Stock

N. America

Push-Pull Boundary

Make-to-Order

Benettons Production and Distribution


Network

Tailored Inventories:
Postponement
Simple solution
Produce all garments as Greige goods (Production
cost is 10% higher)

Tailored solution
Base load manufactured from colored thread
(cheaper but long lead time sourcing)
Safety stock held as Greige goods and
manufactured on demand (10% more expensive but
short lead times)

Process Postponement

Dyeing
2 wks

Knitting
2 wks

Make to Stock

Dyeing
2 wks

Dye to Order

Postponement and Re-assortment:


The Advantage to Forecasting
Actual total sales
4000

4000

4000

3500

3500

3500

3000

3000

3000

2500

2500

2500

2000

2000

2000

1500

1500

1500

1000

1000

1000

500

500

500
0

0
0

500

1000 1500 2000 2500 3000 3500 4000

Initial Forecast

500

1000 1500 2000 2500 3000 3500 4000

Updated Forecast after observing 20% of sales

Each data point represents the forecast and the actual


season sales for a particular item (at the style-color level).

500 1000 1500 2000 2500 3000 3500 4000

after 80%

Benetton:
Two types of Production Capacities
Speculative Capacity
Long Lead Time cheap
Commit before
observing the demand
Gamble!

Reactive Capacity
Short Lead Time - expensive
Commit after
observing the early sales data

How do we decide on the size of the speculative


capacity?

Optimal Service Level and Accurate Response to


Demand Uncertainty when you can order only once:
Pal Gears is planning to offer a special line of winter jackets, especially
designed as gifts for the Christmas season. Each Christma-jacket costs the
company $250 and sells for $450. Any stock left over after Christmas would
be disposed of at a deep discount of $195. Marketing had forecasted a demand
of 2000 Christmas-jackets with a forecast error (standard deviation of 500)

How many jackets should Pal Gears order?


Demand forecast for Christmas jackets
18%
16%
16%

14%

16%

13%

12%

13%

10%
10%

8%

10%

6%
6%

6%

4%
2%

3%

1%

3%

1%

0%
800

1000

1%
1200

1400

1600

1800

2000

2200

2400

2600

2800

3000

1%

3200

Optimal Service Level and Accurate Response: with Excel


(1) Performance for a given order Q, say Q = 2000

Optimal Service Level and Accurate Response : with Excel


(2) Performance for all possible Q

200

-55

= F(Q)

Towards the newsboy model


Suppose you placed an order of 2000 units but you are not sure
about whether you should have ordered one more unit.
What is the contribution of ordering an additional unit?
Marginal Benefit
Earn a margin (p-c) = B with
propability (1-p)=

Marginal Cost
Incur an overage cost of (c-s) = C with
probability p = P(D<=2000)

P(D>2000)
=

..

=
Expected contribution of an additional unit
E() = ..

..

So? ... Order more?

Accurate Response:
The newsboy model
In general: at the optimal Q
E() <= 0 no incentives to order more
(1-p)B =
Sell

pC
Do not sell

Equivalently, choose the smallest Q


such that

p = P(D<Q) = F(Q) >= B /

Example:
Critical fractile
B / (B+C) =

230
220
210
(B+C)
o
200
Expected 190
Profit ($k) 180
170
160
150
140
130
u
Order/Stock
120
Quantity Q
10001200140016001800200022002400260028003000

Find Q by rounding up!


Q=

- C = - 55

+ 200 = C

Accurate response:
Find optimal Q from newsboy model

Cost of overstocking by one unit = C


the out-of-pocket cost per unit stocked but not demanded
Say demand is one unit below my stock level. How much did the one unit
overstocking cost me?
E.g.: purchase price - salvage price.

Cost of understocking by one unit = B


The opportunity cost per unit demanded in excess of the stock level provided
Say demand is one unit above my stock level. How much could I have saved
(or gained) if I had stocked one unit more? E.g.: retail price - purchase price.

Given an order quantity Q, increase it by one unit if and only if the


expected benefit of being able to sell it exceeds the expected cost of having
that unit left over.
Marginal Analysis: Order more as long as F(Q) < B / (B + C)

= smallest Q such that service level F(Q) > critical fractile

B / (B + C)

Where else do you find newsboys?

Benefits: Flexible Spending Account decision


ATM
Perishable Products (Newspaper, Medical Supplies,
Fashion Goods)
Weddings, Conferences

How do we deal with uncertainty?