Managing Uncertainty in a

Supply Chain - Safety

Inventory Dr. V V S N V Prasad
BITS Pilani Faculty in Management
Department of Economics
K K Birla Goa Campus
Learning Objectives (1 of 2)

12.1 Understand the role of safety inventory in a supply

chain.
12.2 Identify factors that influence the required level of safety
inventory.
12.3 Evaluate the appropriate level of safety inventory for a
supply chain.
12.4 Discuss the impact of supply uncertainty on safety
inventory.

BITS Pilani, K K Birla Goa Campus

Learning Objectives (2 of 2)

12.5 Understand how aggregation helps reduce the

required safety inventory in a supply chain.
12.6 Determine the impact of replenishment policies
on safety inventory.
12.7 Identify managerial levers that lower safety
inventory without hurting product availability.

BITS Pilani, K K Birla Goa Campus

The Role of Safety Inventory (1 of 3)

• Safety inventory is carried to satisfy demand that

exceeds the amount forecasted
– Raising the level of safety inventory increases
product availability and thus the margin captured
from customer purchases
– Raising the level of safety inventory increases
inventory holding costs

BITS Pilani, K K Birla Goa Campus

Inventory level

Q
Reorder
point, R

Safety Stock
0
LT LT
Time
BITS Pilani, K K Birla Goa Campus
The Role of Safety Inventory (2 of 3)

Figure 12-1 Inventory Profile with Safety Inventory

BITS Pilani, K K Birla Goa Campus

The Role of Safety Inventory (3 of 3)

• Three key questions

1. What is the appropriate level of product

availability?
2. How much safety inventory is needed for the
desired level of product availability?
3. What actions can be taken to reduce safety
inventory without hurting product availability?

BITS Pilani, K K Birla Goa Campus

Summary of Learning Objective 1

Safety inventory helps a supply chain provide

customers with a high level of product availability in
spite of supply and demand uncertainty. It is carried
just in case demand exceeds the amount forecasted
or supply arrives later than expected.

BITS Pilani, K K Birla Goa Campus

Factors Affecting the Level of Safety
Inventory
• The desired level of product availability
• The uncertainty of demand
• The uncertainty of supply
• Inventory replenishment policies

BITS Pilani, K K Birla Goa Campus

Measuring Product Availability

1. Product fill rate (fr)

– Fraction of product demand satisfied from
product in inventory
2. Order fill rate
– Fraction of orders filled from available inventory
3. Cycle service level (CSL)
– Fraction of replenishment cycles that end with
all customer demand being met
– Replenishment cycle – the interval between
two successive replenishment deliveries

BITS Pilani, K K Birla Goa Campus

Measuring Demand Uncertainty

D = Average demand per period

σD = Standard deviation of demand (forecast error) per period
Lead time (L) is the gap between when an order is placed
and when it is received
σL = Standard deviation of demand during lead time

BITS Pilani, K K Birla Goa Campus

Measuring Supply Uncertainty

Lead time (L) is normally distributed with

L = Average lead time
SL = Standard deviation of lead time

BITS Pilani, K K Birla Goa Campus

Replenishment Policies

1. Continuous review
– Inventory is continuously tracked
– Order for a lot size Q is placed when the inventory
declines to the reorder point (ROP)
2. Periodic review
– Inventory status is checked at regular periodic intervals
– Order is placed to raise the inventory level to a specified
threshold

BITS Pilani, K K Birla Goa Campus

Summary of Learning Objective 2

Safety inventory is influenced by the desired product

availability, demand uncertainty, replenishment lead
times, and lead time variability.
Product availability is measured using the fill rate or
cycle service level.
Demand uncertainty is measured by the forecast
error. The required safety inventory is also influenced
by the inventory policy implemented. Continuous
review policies order a fixed quantity after variable
replenishment intervals. Periodic review policies order
variable quantities after fixed replenishment intervals.

BITS Pilani, K K Birla Goa Campus

Determining the Appropriate Level of Safety
Inventory (1 of 8)
• Evaluating Safety Inventory Given a Reorder Point

Expected demand during lead time = D × L

Safety inventory, ss = ROP − D × L

BITS Pilani, K K Birla Goa Campus

Determining the Appropriate Level of Safety
Inventory (2 of 8)
Average demand per week, D = 2,500
Standard deviation of weekly demand, sD = 500
Average lead time for replenishment, L = 2 weeks
Reorder point, ROP = 6,000
Average lot size, Q = 10,000

Safety inventory, ss = ROP −DL = 6,000 −5,000 = 1,000

Q 10,000
Cycle inventory = = = 5,000
2 2

BITS Pilani, K K Birla Goa Campus

Determining the Appropriate Level of Safety
Inventory (3 of 8)
Average inventory= cycle inventory + safety
inventory
= 5,000 + 1,000 = 6,000
average inventory 6,000
Average flow time = = = 2.4 weeks
throughput 2,500

BITS Pilani, K K Birla Goa Campus

Determining the Appropriate Level of Safety
Inventory (4 of 8)
• Evaluating Cycle Service Level Given a Reorder
Point

CSL = F(ROP, DL, σL) = NORMDIST(ROP, DL, σL, 1)

BITS Pilani, K K Birla Goa Campus

Determining the Appropriate Level of Safety
Inventory (5 of 8)

Q = 10,000, ROP = 6,000, L = 2 weeks

D = 2,500/week, σD = 500

DL = D �L = 2 �2,500 = 5,000
s L = Ls D = 2 �500 = 707

CSL = F(ROP, DL, σL) = NORMDIST(ROP, DL,

σL, 1)
= NORMDIST(6,000, 5,000, 707, 1) =
0.92
BITS Pilani, K K Birla Goa Campus
Determining the Appropriate Level of Safety
Inventory (6 of 8)
• Evaluating Required Safety Inventory Given a Desired
Cycle Service Level
Desired cycle service level = CSL
Mean demand during lead time = DL
Standard deviation of demand during lead time = σL

BITS Pilani, K K Birla Goa Campus

Determining the Appropriate Level of Safety
Inventory (7 of 8)

ss = FS–1(CSL ) �s L = FS–1(CSL ) � Ls D
= NORMSINV (CSL ) � Ls D

BITS Pilani, K K Birla Goa Campus

Determining the Appropriate Level of Safety
Inventory (8 of 8)

BITS Pilani, K K Birla Goa Campus

Evaluating Fill Rate Given a Reorder Point (1
of 4)

• Expected shortage per replenishment cycle

(ESC) is the average units of demand that are not
satisfied from inventory in stock per
replenishment cycle
• Product fill rate
ESC (Q – ESC )
fr = 1 – =
Q Q

BITS Pilani, K K Birla Goa Campus

Evaluating Fill Rate Given a Reorder Point (2
of 4)

BITS Pilani, K K Birla Goa Campus

Evaluating Fill Rate Given a Reorder Point (3
of 4)

Lot size, Q = 10,000

Average demand during lead time, DL =5,000
Standard deviation of demand during lead time,
s L = 707

Safety inventory, ss = ROP − DL = 6,000−5,000 =

1,000
ESC = –1,000[1 – NORMDIST (1,000 707,0,1,1)]
+ 707 NORMDIST (1,000 707,0,1,0) = 25

fr = (Q - ESC) Q = 110,000 – 252 10,000 = 0.9975

BITS Pilani, K K Birla Goa Campus

Evaluating Safety Inventory Given Desired
Fill Rate (2 of 4)
Desired fill rate, fr = 0.975
Lot size, Q = 10,000 boxes
Standard deviation of ddlt, sL = 2 �500 = 707

ESC = (1−fr)Q = (1 − 0.975)10,000 = 250

BITS Pilani, K K Birla Goa Campus

Evaluating Safety Inventory Given Desired
Fill Rate (3 of 4)

� �
�ss � �ss �
ESC = 250 = -ss �
1 – Fs � � �+ s L fs � �
� �sL �
� �sL �

� �ss �� �ss �
= – ss �
1 – Fs � � � + 707fs � �
� �707 �
� �707 �

1 – NORMDIST ( ss 707,0,1,1) �
250 = – ss �
� �
+ 707NORMDIST (ss 707,0,1,0)

Use GOALSEEK to find safety inventory ss = 67

boxes

BITS Pilani, K K Birla Goa Campus

Impact of Desired Product Availability, Lead
Time, and Demand Uncertainty
• As desired product availability goes up the
required safety inventory increases

Table 12- Required Safety Inventory for Different Values of Fill Rate

Fill Rate Safety Inventory

97.5% 67
98.0% 183
98.5% 321
99.0% 499
99.5% 767

BITS Pilani, K K Birla Goa Campus

Impact of Desired Product Availability and
Uncertainty
• Goal is to reduce the level of safety inventory
required in a way that does not adversely affect
product availability
1. Reduce the supplier lead time L
2. Reduce the underlying uncertainty of demand
(represented by σD )

BITS Pilani, K K Birla Goa Campus

Benefits of Reducing Lead Time and
Demand Uncertainty
D = 2,500/week σD = 800, CSL = 0.95

ss = NORMSINV (CSL ) � Ls D
= NORMSINV (.95) � 9 �800 = 3,948

• If lead time is reduced to one week

ss = NORMSINV (.95) � 1 �800 = 1,316

• If standard deviation is reduced to 400

ss = NORMSINV (.95) � 9 �400 = 1,974

BITS Pilani, K K Birla Goa Campus

Summary of Learning Objective 3 (1 of 2)

Given a desired cycle service level CSL, a lead time

L, and a standard deviation of periodic demand σD ,
the required safety inventory ss for a continuous
b
review
ss =policy is given
NORMSINV (CSL ) � Ls D
y
ROP,
Given a reorder a lead time L, a standard deviation of
point
periodic demand σD, and periodic demand D, the resulting
cycle service level is given by
CSL = NORMDIST (ROP , D �L, Ls D ,1)

BITS Pilani, K K Birla Goa Campus

Summary of Learning Objective 3 (2 of 2)

Given a level of safety inventory, one can evaluate

the resulting fill rate.
Given a desired fill rate, one can evaluate the
required safety inventory.
The required safety inventory increases with an
increase in desired product availability, lead time, and
uncertainty of periodic demand.

BITS Pilani, K K Birla Goa Campus

Impact of Supply Uncertainty on Safety
Inventory
We incorporate supply uncertainty by assuming that lead time is
uncertain
D : Average demand per period
σL : Standard deviation of demand per period
L : Average lead time for replenishment
sL : Standard deviation of lead time

DL = DL s L = Lσ D2 + D 2sL2

BITS Pilani, K K Birla Goa Campus

Impact of Lead Time Uncertainty on Safety
Inventory (1 of 3)
Average demand per period, D = 2,500
Standard deviation of demand per period, σD = 500

Average lead time for replenishment, L = 7 days

Standard deviation of lead time, sL = 7 days
Mean ddlt, DL = DL = 2,500 × 7 = 17,500

Standard deviation of ddlt s L = Ls D2 + D 2sL2

= 7 �5002 + 2,5002 �72
= 17,500

BITS Pilani, K K Birla Goa Campus

Impact of Lead Time Uncertainty on Safety
Inventory (2 of 3)
• Required safety inventory

ss = FS–1(CSL ) �s L = NORMSINV (CSL ) �s L

= NORMSINV (0.90) �17,500
= 22,491 tablets

BITS Pilani, K K Birla Goa Campus

Impact of Lead Time Uncertainty on Safety
Inventory (3 of 3)
Table 12-2 Required Safety Inventory as a Function of Lead Time
Uncertainty

sL σL ss (units) ss (days)
6 15,058 19,298 7.72
5 12,570 16,109 6.44
4 10,087 12,927 5.17
3 7,616 9,760 3.90
2 5,172 6,628 2.65
1 2,828 3,625 1.45
0 1,323 1,695 0.68

BITS Pilani, K K Birla Goa Campus

Summary of Learning Objective 4

An increase in supply uncertainty significantly

increases the amount of safety inventory required for
a given level of product availability.
Lead time uncertainty has a more significant impact
on the required safety inventory than lead time itself.
A reduction in supply uncertainty can help to
dramatically reduce the required safety inventory
without hurting product availability.

BITS Pilani, K K Birla Goa Campus

Impact of Aggregation on Safety Inventory (1
of 5)

• How does aggregation affect forecast accuracy

and safety inventories

Di: Mean periodic demand in region i, i = 1, …, k

σi : Standard deviation of periodic demand in region i, i =,
…, k
ρij: Correlation of periodic demand for regions i, j, 1 �i �j �k

BITS Pilani, K K Birla Goa Campus

Impact of Aggregation on Safety Inventory (2
of 5)

k
Total safety inventory in
decentralized option = � S (CSL) � L �s i
F –1

i=1

k
DC = �i =1 Di ; var ( DC ) = �s i2 + 2�rij s i s j ;
k

i =1 i> j

s DC = var ( DC )

DC = kD s DC = ks 2 + k ( k – 1) rs 2

Simplified to D = ks D
s C

BITS Pilani, K K Birla Goa Campus

Impact of Aggregation on Safety Inventory (3
of 5)

Require safety inventory on aggregation

k
= �FS–1(CSL ) � L �s DC
i =1

Holding – cost savings on aggregation per unit sold

FS–1 (CSL ) � L �H �k C �
=
D C
���
�i =1
si – sD �
�

BITS Pilani, K K Birla Goa Campus

Impact of Aggregation on Safety Inventory (4
of 5)

• The safety inventory savings on aggregation increase

with the desired cycle service level CSL
• The safety inventory savings on aggregation increase
with the replenishment lead time L
• The safety inventory savings on aggregation increase
with the holding cost H
• The safety inventory savings on aggregation increase
with the coefficient of variation(sof
D Ddemand
)

The safety inventory savings on aggregation decrease as the

correlation coefficients increase

BITS Pilani, K K Birla Goa Campus

Impact of Aggregation on Safety Inventory

BITS Pilani, K K Birla Goa Campus

Impact of Correlation on Value of
Aggregation (1 of 3)
Standard deviation of weekly demand =25, σD =5
Replenishment, L = 2 weeks; Decentralized CSL = 0.9

Total required safety inventory, ss = k �Fs ( CSL ) � Lσ� D

–1

= 4 �Fs–1 ( 0.9 ) � 2 �5
= 4 �NORMSINV ( 0.9 ) � 2 �5 = 36.25 cars

Aggregate ρ = 0
Standard deviation of weekly
demand at central outlet,
ss = Fs–1 ( 0.9 ) � Lσ� DC NORMSINV
= ( 0.9 ) � 2 �10 = 18.12
BITS Pilani, K K Birla Goa Campus
Impact of Correlation on Value of
Aggregation (2 of 3)
Table 12-3 Safety Inventory in the Disaggregate and Aggregate Options

Disaggregate Aggregate
ρ Rho
Safety Inventory Safety Inventory
0 36.25 18.12
0.2 36.25 22.93
0.4 36.25 26.88
0.6 36.25 30.33
0.8 36.25 33.42
1.0 36.25 36.25

BITS Pilani, K K Birla Goa Campus

Impact of Aggregation on Safety Inventory (5
of 5)

Figure 12-4 Square-Root Law

BITS Pilani, K K Birla Goa Campus

Impact of Correlation on Value of
Aggregation (3 of 3)
• Two possible disadvantages to aggregation

1. Increase in response time to customer order

2. Increase in transportation cost to customer

BITS Pilani, K K Birla Goa Campus

Impact of Aggregation

 If number of independent stocking locations decreases

by n, the expected level of safety inventory will be
reduced by square root of n (square root law)
 Many e-commerce retailers attempt to take advantage of
aggregation (Amazon) compared to bricks and mortar
retailers (Borders)
 Aggregation has two major disadvantages:
– Increase in response time to customer order
– Increase in transportation cost to customer
– Some e-commerce firms (such as Amazon) have reduced
aggregation to mitigate these disadvantages

BITS Pilani, K K Birla Goa Campus

Trade-Offs of Physical Centralization (1 of 2)

Use four regional or one national distribution center

D = 1,000 week,s D = 300, L = 4 weeks, CSL = 0.95

• Four regional centers

Total required
safety inventory, ss = 4 �F -1
s (CSL ) � L �sD

= 4 �NORMSINV ( 0.95 ) � 4 �300 = 3,948

BITS Pilani, K K Birla Goa Campus

Trade-Offs of Physical Centralization (2 of 2)

• One national distribution center, r = 0

Standard deviation
s DC = 4 �300 = 600
of weekly demand,
ss = Fs-1 ( 0.95 ) � L �s DC
= NORMSINV ( 0.95 ) � 4 �600 = 1,974

Decrease in holding costs = (3,948 - 1,974) �$1,100 �0.2

= $394,765
Decrease in facility costs = $150,000
Increase in transportation = 52 �1,000 �(13 - 10)
= $624,000
BITS Pilani, K K Birla Goa Campus
Information Centralization

• Online systems that allow customers or stores to

locate stock (Virtual aggregation)
• Information system that allows access to current
inventory records in all warehouses from each
warehouse
• Most orders are filled from closest warehouse
• In case of a stock out, another warehouse can fill
the order
• Better responsiveness, lower transportation cost,
higher product availability, but reduced safety
inventory
BITS Pilani, K K Birla Goa Campus
Specialization (1 of 2)

• Stock all items in each location or stock different

items at different locations?
• Different products may have different demands in different locations (e.g.,
snow shovels)

• There can be benefits from aggregation

• Benefits of aggregation can be affected by:

• coefficient of variation of demand (higher cv yields greater reduction in
safety inventory from centralization)

• value of item (high value items provide more benefits from centralization)

BITS Pilani, K K Birla Goa Campus

Product Substitution

• The use of one product to satisfy demand for a different

product
1. Manufacturer-driven substitution

Allows aggregation of demand

Reduce safety inventories
Influenced by the cost differential, correlation of
demand
2. Customer-driven substitution
Allows aggregation of safety inventory

BITS Pilani, K K Birla Goa Campus

Component Commonality

• Without common components

– Uncertainty of demand for a component is the
same as for the finished product
– Results in high levels of safety inventory
• With common components
– Demand for a component is an aggregation of
the demand for the finished products
– Component demand is more predictable
– Component inventories are reduced

BITS Pilani, K K Birla Goa Campus

Value of Component Commonality

BITS Pilani, K K Birla Goa Campus

Postponement (1 of 2)

• Delay product differentiation or customization until closer to

the time the product is sold
– Have common components in the supply chain for most
of the push phase
– Move product differentiation as close to the pull phase
of the supply chain as possible
– Inventories in the supply chain are mostly aggregate

BITS Pilani, K K Birla Goa Campus

Postponement (2 of 2)

Figure 12-6 Supply Chain Flows without and with Postponement

BITS Pilani, K K Birla Goa Campus

Value of Postponement

100 different paint colors, D = 30/week, sD = 10, L = 2

weeks, CSL = 0.95

Total required
s ( CSL ) � L �sD
-1
safety inventory, ss = 100 �F

= 100 �NORMSINV ( 0.95 ) � 2 �10 = 2,326

Standard deviation of base paint weekly demand,

s DC = 100 �10 = 100

ss = Fs-1 ( CSL ) � L �s CD = NORMSINV (0.95) � 2 �100 = 233

BITS Pilani, K K Birla Goa Campus

Summary of Learning Objective 5

Aggregation reduces the required safety inventory as

long as the demand across the aggregated regions is
not perfectly, positively correlated.
The safety inventory savings on aggregation increase
with the desired CSL, the replenishment lead time,
the product holding cost, and the coefficient of
variation of demand.
The safety inventory savings on aggregation decrease
as the correlation of demand across regions
increases.
Firms can aggregate inventories through physical
aggregation, information centralization, product
substitution, component commonality, andBITS Pilani, K K Birla Goa Campus
Impact of Replenishment Policies on Safety
Inventory (1 of 4)
• Continuous Review Policies

D : Average demand per period

σD : Standard deviation of demand per period

L : Average lead time for replenishment

Mean demand during lead time, DL = D×L

Standard deviation of demand during lead time, s L = Ls D

ss = FS–1 (CSL ) �s L = NORMSINV (CSL ) � Ls D ,ROP = DL + ss

BITS Pilani, K K Birla Goa Campus

Impact of Replenishment Policies on Safety
Inventory (2 of 4)
• Periodic Review Policies
– Lot size determined by prespecified order-up-
to level (OUL)

D : Average demand per period

σD : Standard deviation of demand per period
L : Average lead time for replenishment
T : Review interval
CSL : Desired cycle service level

BITS Pilani, K K Birla Goa Campus

Impact of Replenishment Policies on Safety
Inventory (3 of 4)
Probability (demand during L + T �OUL) = CSL

Mean demand during T+L periods, DT +L = (T + L )D

Std dev demand during T+L periods, s T + L = T + Ls D

OUL = DT +L + ss

ss = FS−1 (CSL)  s T +L

BITS Pilani, K K Birla Goa Campus

Impact of Replenishment Policies on Safety
Inventory (4 of 4)

Figure 12-7 Inventory Profile for Periodic Review Policy

with L = 4, T = 7

BITS Pilani, K K Birla Goa Campus

Evaluation Safety Inventory for a Periodic
Review Policy
D = 2,500, σD= 500, L = 2 weeks, T = 4 weeks

Mean demand during T + L periods, DT +L = (T + L )D

= (2 + 4)2,500 = 15,000
Std dev demand
during T + L periods, s T +L = T + Ls D
= ( 4 + 2)500 = 1,225

ss = FS-1 ( CSL ) �s D +L = NORMSINV ( CSL ) �s T +L

= NORMSINV ( 0.90 ) �1,225 = 1,570 boxes
OUL = DT +L + ss = 15,000 + 1,570 = 16,570

BITS Pilani, K K Birla Goa Campus

Summary of Learning Objective 6

Periodic review replenishment policies require more safety

inventory than continuous review policies for the same
lead time and level of product availability.

BITS Pilani, K K Birla Goa Campus

Managerial Levers to Reduce Safety
Inventory
• Reduction of supply uncertainty
– Sharing information
– Coordinated demand
• Reduction of lead times
– Delays contribute more to lead time than
production and transportation time
• Reduction of demand uncertainty
– Reduce information distortion through sharing
– Aggregate demand

BITS Pilani, K K Birla Goa Campus

Summary of Learning Objective 7

The required level of safety inventory may be reduced

and product availability may be improved if a supply
chain can reduce demand uncertainty, replenishment
lead times, and the variability of lead times. A switch
from periodic monitoring to continuous monitoring
can also help reduce inventories. Another key
managerial lever to reduce the required safety
inventories is to exploit aggregation. This may be
achieved by physically aggregating inventories,
virtually aggregating inventories using information
centralization, specializing inventories based on
demand volume, exploiting substitution, using
component commonality, and postponing product
differentiation.
BITS Pilani, K K Birla Goa Campus
 BITS Pilani
K K Birla Goa Campus

Thank you