CONFIDENTIAL

Introduction to inventory management

This report is solely for the use of client personnel. No part of it may be
circulated, quoted, or reproduced for distribution outside the client
organization without prior written approval from McKinsey & Company.
This material was used by McKinsey & Company during an oral
presentation; it is not a complete record of the discussion.
AGENDA DRAFT

 Introduction to inventory management

Working Draft - Last Modified

 Standard analysis used during an inventory
management engagement

 Using the inventory tool – demo

 “Advanced” topics in inventory management

Printed
1
INVENTORY CAN BE HELD AT VARIOUS STAGES IN THE SUPPLY CHAIN

DC
Supplier DC C
DC

Working Draft - Last Modified

U
S
Supplier DC T
DC O
Warehouse Manu- M
facturing DC
Supplier E
• Raw material • WIP R
inventory DC DC S
inventory

Printed
Supplier DC

• Finished goods
inventory

Each addresses a different supply and demand matching

• Raw material: Supplier and manufacturing
• WIP: Manufacturing and DC
• Finished goods: DC and customer 2
WHY DO WE NEED INVENTORY IN OUR PROCESS? Primary focus
Secondary focus

 What kind of inventory should we maintain?

Working Draft - Last Modified

– Only raw materials/components (build-to-order)
– Finished goods and raw materials (build-to-stock)
– Intermediates (production postponement)
 Where should we stock inventory in the distribution
network?

Printed
 How much inventory should we maintain at each
location?

 How sensitive is my inventory policy to changes in

business conditions?

Source: McKinsey
3
INVENTORY BUFFERS THE MISMATCH BETWEEN DEMAND
AND SUPPLY PROCESSES

Working Draft - Last Modified

Supply Demand

Inventory

Printed
Inventory root causes

Supply variability Other factors Demand variability

• Manufacturing variability • Manufacturing and distribution • Demand volatility

(yield, cycle time) cycle times • Forecast error
• Supplier variability • Information flow lead times • Seasonality
• Capacity constraints • Purchase discounts
• Special purchases (e.g. EOL)

4
LEVEL OF INVENTORY HELD IS FUNDAMENTALLY A RESULT
OF CROSS-FUNCTIONAL TRADE-OFFS
Warehouse:
Marketing: Finance: “I’m out of
“I can’t sell from an “Where am I going space”
empty wagon” to get the money?”

Working Draft - Last Modified

Cost of
holding
Product inventory
availability
and service

Printed
Manufacturing: • Inventory
“Large lot sizes problem is cross-
and advance notice functional
can help us reduce
unit costs” • Balance is the
paramount
consideration

5
SEVERAL TYPES OF INVENTORY EXIST Focus of discussion

Typical inventory layers

Working Draft - Last Modified

• Average inventory level between placement of an order
Cycle stock and receipt of product
• Buffer inventory to protect against uncertainties
Safety stock (protect against possibility that expected demand exceeds
available inventory)

Seasonal stock • Extra inventory put in to accommodate seasonal patterns

Printed
• Inventory built up in anticipation of future event, such as
Forward purchase stock industry shortage or expected price increase

Reserve stock
• Inventory on hand, but reserved or “tagged” for specific
customers

Sludge • Error in inventory level due to system, human, or other errors

6
CYCLE STOCK IS THE AVERAGE INVENTORY LEVEL BETWEEN
PLACEMENT OF AN ORDER AND PRODUCT RECEIPT ILLUSTRATIVE

Inventory

100 – Receive
shipmen
90 –

Working Draft - Last Modified

t for 80
80 –
70 – Cycle stock
is driven
60 – largely by
50 – the order
Average quantity
cycle

Printed
40 –
inventory
30 – = 40
20 –
10 –

Place 1 2 3 4 5 6 7 8 9 10 11 12
order Lead time Weeks
for 80 For illustration, assumes constant demand of 20
per week, 4-week lead time with instantaneous
replenishment 7
SAFETY STOCK RAISES THE OVERALL INVENTORY LEVEL ILLUSTRATIVE
AND PROVIDES A BUFFER AGAINST SUPPLY AND DEMAND
UNCERTAINTY

Inventory

Working Draft - Last Modified

100 –
90 –
80 –
70 –
Expecte
60 – d

Printed
Possible demand
50 –
demand
40 – Average inventory = cycle stock + safety stock
60 = 40 + 20
30 –
20 –
10 – Safety stock = 20

1 2 3 4 5 6 7 8 9 10 11 12
Weeks
8
SEASONAL STOCK IS USUALLY BUILT UP IN ANTICIPATION
OF PREDICTED SEASONAL RAMP-UP IN VOLUME

• Build up
Inventory
inventory during
250 – April-July

Working Draft - Last Modified

• Accommodate
high seasonal
200 – demand
Inventory October through
level November and
bleed down
150 – seasonal
Demand
inventory

Printed
Build up
seasonal • Required only if
stock Production production and
100 – (100 units/ distribution
month max) capacity cannot
handle seasonal
50 – peaks for
demand
J F M A M J J A S O N D
Months

9
COMBINING CYCLE, SAFETY, AND SEASONAL STOCKS
PROVIDES AN OVERVIEW OF BASIC INVENTORY MANAGEMENT
AND STOCKING LOGIC
ILLUSTRATIVE
Inventory
100 –

90 –

Working Draft - Last Modified

80 –
1
Order 70 – 2 Reorder point (function
of lead time and expected
quantity 60 – demand)
(function Expected 7 Average inventory =
of product 50 – Possibl
demand
e cycle stock + safety stock
carrying
cost and 40 – demand

Printed
setup cost)
30 –

20 – 6
5 Safety stock Safety stock
10 –
Cycle stock
(order 1 2 3 4 5 6 7 8 9 10 11 12
Weeks
quantity ÷ 2)
3 With 3-week 4 Receive Primary inventory drivers
lead time, shipment
• Replenishment lead time (2, 3, 4)
place order
• Demand/supply variability (2, 4, 6)
here
• Customer service levels (1, 5)
10
TWO FUNDAMENTAL QUESTIONS IN INVENTORY MANAGEMENT

Working Draft - Last Modified

 When do I order?

 How much do I order?

Printed
11
 Probability of meeting demand
OPTIMAL SAFETY STOCK DEPENDS ON THE Probability of shortage
REQUIRED SERVICE LEVEL
As we increase the service level …
50% service level 85% service level 99% service level

Working Draft - Last Modified

… the safety stock increases dramatically and inventory turns fall
Safety stock

Printed
Units
2.50
2.00
1.50
1.00
0.50
0.00
50% 60% 70% 80% 90% 100%
Customer service level
Percent
12
HOWEVER, INVENTORY COST NEEDS TO BE BALANCED WITH
INCREASED REVENUE DRIVEN BY HIGHER SERVICE LEVELS

Working Draft - Last Modified

Revenue margin through
increased revenue
Profit function

• Optimal “profit”
point
Profit curve • Optimal customer
service level

Printed
70% 75% 80% 85% 90% 95% 100%

Customer service level

Inventory carrying cost

through increased inventory
levels

13
WITH SAFETY STOCK, REORDER WHEN INVENTORY LEVEL
EQUALS THE DEMAND DURING LEAD TIME PLUS SAFETY STOCK

Inventory
level

Working Draft - Last Modified

New OP = average demand
New OP
Order quantity

during lead time + the level

of safety stock

Average

Printed
demand during
lead time

Level of safety
stock Time

14
THE TWO MODELS THAT THE OMC USES MOST FREQUENTLY TO
CALCULATE SAFETY STOCK UTILIZE ALPHA AND BETA SERVICE LEVELS
Results in a
higher -Service Level (used in many -Service Level
inventory level models)

Working Draft - Last Modified

• Probability of stock-out event Quantity delivered on time
Definition during replenishment cycle
Requested quantity
• Limited practical value, because • High practical value, because
Application it tells nothing about duration of usually a better measure for
stock-outs or backlogged service related (opportunity) cost
volume • Usually the base of KPI
• No consideration of order-line- definitions (e.g., Fill rate, OTIF)

Printed
or partial delivery • Accounts for partial delivery
• Often used in retailing where • Often used in industry where fill
backorders cannot be placed rate is a KPI
Inventory • Safety factors can be calculated • Safety-factor/safety stock
Management easily based on -service level. calculation is more sophisticated,
However often wrong usage, which are however absolutely
because KPIs are defined necessary, in the case of KPIs
rather in terms of -service level defined in terms of -service
Source: TEMPELMEIER 1999 page 368 ff., MINNER 2000, page 31, SILVER 1998, McKinsey
level
15
Example of alpha versus beta
Assume we replenish to top up of 20 units of inventory each week

Alpha method
25
23 Stocked out twice

Working Draft - Last Modified

15 15 16 Resulting alpha
14
12 12 service level is
10
8 80%
15 15
10 12 12
8

Beta method

Printed
25 Filled 142 out of
23
5 3 150 orders (“fill
16 rate”)
15 15 14
12 12
10 Beta service level
8 20 20 is 94.6%
15 15
10 12 12
8

Source: Wikipedia 16
Safety stock calculation with α-service level

Safety stock explicitly looks at the probability of … and is calculated based on

stockout based on historical volatility of demand … several parameters*

Safety stock = z · σD · SQRT (Lt)

Working Draft - Last Modified

Lt = Lead time, SQRT due to
convolution of normal distri-
bution to cover >1 periods of
replenishment leadtime
Order σD = Standard deviation demand
qty Cycle Concerned
stock z = Safety factor
about
demand

Printed
greater than Service level Z
Safety the mean 95.0% 1.65
stock
99.0% 2.33
=NORMSINV(0,99) in Excel
Lead time Possible
shortage
Stochastic treatment of leadtime event Stochastic leadtime
not reasonable, because it does
= z  lt s d2 + d s lt2
2
not follow a normal distribution Safety stock

* α-service level definition

Source: McKinsey

17
Safety stock calculation with α-service level NON-EOQ METHOD

n
Target inventory = (L  D) + (z  sd * L) + (x/7  D) + [  (Di-D) / n ]
i=1

= Demand during + Demand side + Supply side + Cycle stock (promotional/

lead time safety stock safety stock seasonal stock)

Working Draft - Last Modified

Symbol Name Description
L • Replenishment lead time • The time from when the DC sends the plant its replenishment request
to the moment the product which fills that request in unloaded at the
DC (see the following pages for a detailed example)
D • Average forecasted base • Average of forecasted orders for weeks in the upcoming 52-week
weekly orders period in which there are neither promotions nor seasonal effects
z • Demand safety factor • Number of standard deviations required to meet a specified service
level
• Determined by the cumulative normal distribution function F(z), where

Printed
F(z) = service level (see following pages for details on F(z))
sd • Standard deviation of base • Statistical standard deviation of historic base weekly orders over the
weekly orders past 52-week period
x • Supply safety factor • Number of days of inventory the BU manager feels are necessary to
cover shipments in the event of an unplanned increase in
n replenishment lead time
 • Sum • Summation function of a series of terms from term 1 to the nth term
i=1
(i, i+1, i+2, i+3, . . . , n)
i • Promotion/season week • Promotion/season week number (i will range from 1 to n)
number
n • Total number of weeks • Total number of weeks for the promotion (n is the length of the
for the promotion/season promotion in weeks)
Di • Forecasted weekly orders • Forecasted weekly orders in week i of the promotion/season (for a
for week i 4-week promotion there would be D1, D2, D3, and D4)
18
Example calculation for simplified Beta service level model for safety
stock
Calculation formula

E(kss) = q (1 - SL) / σLT σLT Demand variance during lead time

q Reorder lot size

Working Draft - Last Modified

SL -Service level (0,00 - 1,00)

Example

• σLT = 100 pieces within LT=1 month Expectation

value of stock- Safety
• q = 400 pieces out amount factors,
•  -SL set at 98% E(k)* kss

Printed
0.39894 0.00
► E(kss) = 400 (1 - 0.98)
= 0.08 0.28611 0.25
100 0.19713 0.50
0.13037 0.75
• Corresponding kss for E(kss) of 0.08 = 1.00 0.08285 1.00
• kss is a safety factor that must be looked up in …
accompanying table 0.00864 2.00
► Safety stock = σLT * kss = 100 x 1.00 = 100 pieces ...

Source: McKinsey

Source: McKinsey 19
Key demand and inventory inputs for Alpha and Beta diagnostic tools
Description Data request
• SKU number and category fields provided – • Planning group
SKU Information customize to add additional fields interview

Working Draft - Last Modified

• Weekly (or monthly) demand for previous 12 weeks – • IT or planning group
Demand use weekly/monthly version as appropriate data request to pull
national-level sales by
item

• Client uses a forecast to generate production/order • Planning group

Manufacturing quantity (Make to Forecast): interview
strategy – Key “1” into Forecast Available field

Printed
– Enter Forecast Error (Root Mean Squared Error) as
%
• Client uses past demand to generate production/order
quantity (Make to Stock):
– Key “0” into Forecast Available field

• Weekly (or monthly) average inventory for previous 12 • IT or planning group

Inventory weeks data request to pull
• Average inventory preferred over end-of-period national-level
inventory inventory by item

SOURCE: McKinsey
20
Key demand and inventory inputs for Alpha and Beta diagnostic tools
Description Data request

• Supply leadtime defined as the duration from an order • IT/planning pull if

Leadtime is placed to the time the order is received. available or planning
group interview
(approximate by group
of SKUs)

Working Draft - Last Modified

• Cost/case • Planning group data
Inventory value request

• To use Economic Order Quantity method to determine • Planning group

EOQ parameters order quantity: interview
– Yearly carrying cost

Printed
– Order setup cost – fixed charge, time spent on
ordering
• To use required order quantity: • Planning group
Reorder quantity – Key “2” into order quantity method interview
– Enter required order quantity

Use required order quantity to modify EOQ for

volume discounts, pallet rounding, scheduled
vendor order times

SOURCE: McKinsey
21
TWO FUNDAMENTAL QUESTIONS IN INVENTORY MANAGEMENT

Working Draft - Last Modified

 When do I order?

 How much do I order?

Printed
22
THE OPTIMAL ORDER QUANTITY IS BASED ON TRADE-OFFS
OF SEVERAL COSTS

Cost

Working Draft - Last Modified

Total cost

Minimum Total Cost

Carrying cost

Product costs
Ordering costs

Printed
EOQ Order quantity

EOQ minimizes the total cost associated with the order quantity

D = annual demand
O = variable cost of placing an order
EOQ = q* = 2DO q = order quantity
Ci
i = inventory carrying rate
C = unit cost of the item

23
3 TYPES OF REPLENISHMENT APPROACHES EXIST
Replenishment approaches

Continuous • When to order? – when inventory falls below target level

review (discussed • How much to order? – EOQ
earlier)

Working Draft - Last Modified

• When to order? – place an order at fixed intervals
Fixed interval
ordering • How much to order? – order what has been consumed since
the last order was placed

Printed
Optimal • When to order? – review conducted at fixed intervals
replenishment • How much to order? – if inventory is below pre-determined
system (s, S) level(s), then order amount necessary to bring inventory up to S

24
BASIC INVENTORY REVIEW AND REPLENISHMENT POLICIES

Review how How much

often? When to order? to order? Notation* Comments

Continuous Reorder point Fixed order (S, Q) • Easy to understand

quantity • Less change of errors occurring
EOQ (S, Q) • Predictability of production requirements on the

Working Draft - Last Modified

part of the supplier
• Does not easily cope with demand larger than Q
(can be dealt with in several ways)
Order-up-to-S- (s, S) • Frequently encountered in practice
level • s and S are arbitrarily set and usually erroneous
• The computation effort to find the best values for
s and S is difficult

• Preferred system for coordination the

Printed
Periodic Every R time Order-up-to-S- (R, S)
units level replenishment of related items
• Offers a regular opportunity to adjust the order-
up-to-level S
• Inventory carrying costs are usually higher
Check if Order-up-to-S- (R, s, S) • Offers the lowest total cost of ordering, carrying,
inventory is level and shortage costs of all systems
below reorder • Computational effort to obtain the best value of
point s, every R R, s, and S is difficult
time units • Most difficult to understand relative to the other
systems

* s is the reorder point, Q is the order quantity, S is the maximum level of inventory allowed, and R is the time between each order
placement 25
THESE REPLENISHMENT POLICIES HAVE DIFFERENT APPLICATIONS

Optimal Continuous review (fixed

replenishment Fixed interval quantity

• Efficient batch • Efficient batch operations • Allows continuous

Advantages operations of data of data processing monitoring of inventory

Working Draft - Last Modified

processing • Groups orders, incurring position
• Groups orders, one setup charge • Adjusts order quantity to
incurring one setup • Appropriate when reflect changes in
charge continuous review is forecast
• Avoid uneconomical costly or inconvenient • Usually requires less
small orders safety stock than fixed
• Usually requires lowest interval model

Printed
levels of inventory

• Requires systems and • Replenishment orders • Continuous review may

Disadvantages process sophistication as may be uneconomically be infeasible
well as discipline small • Incurs high ordering costs
• Usually requires more
safety stock, since orders
are placed at fixed
intervals

26
AGENDA DRAFT

 Introduction to inventory management

Working Draft - Last Modified

 Standard analysis used during an
inventory management engagement

 Using the inventory tool – demo

 “Advanced” topics in inventory management

Printed
27
 FOR EACH INVENTORY CATEGORY, IDENTIFY CLIENT EXAMPLE
“SLUDGE” INVENTORY…
Raw material inventory
Turns per year 100% 100% 100%
0
Other* 9 6 5

Working Draft - Last Modified

> 8 turns 7
24 30
3-8 turns 13

About 70% inventory

1-3 turns 20 consists of parts
29 having less than 3
turns

Printed
42
< 1 turn “sludge”
(products with very
low inventory 51
turns) 41
18

5
Products/ Inventory Volume
SKUs
* Includes parts with zero inventory 28
AND INVENTORY REDUCTION POTENTIAL FOR CLIENT EXAMPLE
PRODUCTS WITH HIGH INVENTORY
Raw material inventory
Current inventory Optimal inventory*
Item number Dollars Dollars

Working Draft - Last Modified

1 84,127 18,219
2 57,385 15,786
3 46,568 27,743
4 36,568 48,131
5 33,985 57,155

Printed
6 21,866 8,046
7 1,361 1,677
8 1,144 957 Average 37%
reduction
9 1,088 599 based on
10 633 2,188 sample
11 270 206

* Based on 98% service-levels

29
ONCE TARGET INVENTORY LEVELS HAVE BEEN SET, CLIENT EXAMPLE
ESTIMATE OVERALL INVENTORY REDUCTION POTENTIAL BY
SUMMARIZING RESULTS FOR VARIOUS INVENTORY CATEGORIES
$ Millions Current Opportunity
inventory Percent Impact Rationale
25

Working Draft - Last Modified

3
Purchase
40-60% 1-2
• Network optimization
parts A • Vendor managed inventory
5 • Drop shipments to Service Centers
Purchase
40-60 2-3
• Fact-based inventory logic
parts B • Improved delivery lead time and reliability

10
• Reduced sludge inventory

Printed
Finished goods 20-40 2-4
• Eliminated slow moving nonstorage SKUs

2
WIP 50-60 1-2
• Reduced lot sizes
• Reduced queuing in front of machines
5 • Fact-based inventory logic
Raw material 30-40 1-2 • Improved delivery lead time and reliability

More than $7 million-13 million

inventory reduction potential
30
PRODUCT SEGMENTATION BASED ON CRITICALITY AND USAGE
PROVIDES A DESIGN FRAMEWORK FOR OPTIMIZING INVENTORY
Product criticality vs. usage segmentation

High • Maintain “insurance” • Develop statistically-

level to provide backup based targets with

Working Draft - Last Modified

• Use historical levels of explicit service-level Statistical inventory
usage as guide and safety stock calculations can
calculation establish target
inventory levels for
each segment
Criticality

• Carry minimal, if not • Carry enough buffer to

Printed
zero, inventory provide usage over
• Provide only through lead time Replenishment
planned frequency and quantity
usage/demand can also be
established for each
segment

Low
Low High

Frequency of usage/demand
* Includes whether a shortage of an item with will cause great disruption in operations 31
PRODUCTS CAN ALSO BE SEGMENTED BASED ON CLIENT EXAMPLE
CUSTOMER ATTRACTIVENESS AND REQUIRED SERVICE LEVELS
Customer attractiveness vs. service-level required segmentation

ABC should consider differentiated customer service-level targets

High 89-92% 93-97% 98-100%
• AAFES • ABC • Kmart

Working Draft - Last Modified

• Toys “R” Us • WalHMart
• WW
Grainger
Customer attractiveness to client

85-88% 89-92% 93-97%

• Ace Establish tailored
• Cotter inventory levels
• Lowes
• Home Depot and replenishment
• JC Penney

Printed
approaches for
• Meijer each segment
• Sam’s Club
80-84% 85-88% 89-92%
• Ames
• Caldor
• Service
Merchandise

Low
80% 92% 97% 100%
Customer line-fill service level required

32
AGENDA DRAFT

 Introduction to inventory management

Working Draft - Last Modified

 Standard analysis used during an inventory
management engagement

 Using the inventory tool – demo

 “Advanced” topics in inventory management

Printed
33
FUNCTIONALITY OF THE INVENTORY TOOL
Input Data For Individual Product Source of data

• Segment to which it belongs (this data comes the Segmentation Tool) - Segmentation tool
• Average demand per day and demand variability - User entered
• Lead time in days, lead time variability and planning cycle time in days - User entered
• Current service levels in percentage - User entered

Working Draft - Last Modified

• Lower and upper limits on the service level - User entered
• Inventory value, product revenue, inventroy carrying cost in USD/unit - User entered
• Underage impact in percentage - User entered
• Current inventory level in units and current fill rate - User entered

Perform Inventory Calculations

Printed
(calculate optimal service levels)

Output For Individual Product

• Optimal service level • Safety stock in USD
• Safety stock in units • Cycle stock in USD
• Cycle stock in units • Total stock in USD
• Total stock in units • Inventory carrying cost in USD/year
• Total stock in days • Total stock in days
• Comparison with current scenario • Safety stock in USD

34
AGENDA DRAFT

 Introduction to inventory management

Working Draft - Last Modified

 Standard analysis used during an inventory
management engagement

 Using the inventory tool – demo

 “Advanced” topics in inventory

Printed
management

35
ADVANCED TOPICS IN INVENTORY MANAGEMENT

Working Draft - Last Modified

 Specialized inventory management (low
volume flows, MRO parts)

 Challenges in execution – “If it’s so simple,

why do clients need our help?”

 Inventory management outside

organizational walls – VMI and Consignment

Printed
 Multi-echelon inventory management and
production postponement

36
SPECIALIZED TYPES OF INVENTORY MANAGEMENT NOT EXHAUSTIVE

Type Areas of application Characteristics Status in OMC

Low  Manufacturing (low  Intermittent spiky demand  Poisson model

volume volume SKUs) – impossible to forecast available in Excel
flows  Distribution  Very large SKU count
 After market spares

Working Draft - Last Modified

MRO  Refineries and  Poor inventory visibility  Unknown
other Capital  Ownership spread
intensive industries across multiple locations
 Airplane  Inventory budgets
maintenance  Potentially rotable parts
 Military and reverse logistics
 MTBR/ MTBF data
available

Printed
Limited  Medical equipment  Limited product life –  Simulation model
product  Pharmaceuticals inventory levels limited available in Excel
life  Fresh foods by potential for product and Arena
scrap

Product  High tech  Rapid shift from new  Unknown

life cycle products driven by
manage marketing forecast to
-ment EOL and demand
transition to newer
technology
37
CHALLENGES IN EXECUTION FOR DISCUSSION

 Inventory planning often not linked to strategic objectives

– Setting of corporate inventory targets often based on financial
requirements and not on operational capability
– Lack of alignment of supply chain segmentation objectives

Working Draft - Last Modified

 Inventory ownership is often only on one side of the S&OP
(sales/marketing or production) leading to unbalanced objective.

 Poor data quality

– Forecasts: biased and high error rate
– Manufacturing and distribution cycle times: outdated

Printed
– Supply lead time variability: not measured
– Service levels: rarely measured in consistent and conservative
manner (“how do you define and measure on-time-delivery?” can
lead to surprising answers)

 Inventory management is reactive – leading to sharp swings between

high inventory levels and poor service levels

 Hard to get rid of slow moving inventory

 Others?
38
VMI AND CONSIGNMENT
Primary challenges
• Inventory visibility
• Pricing
• SOX compliance
• Balance sheet improvement may lead to higher costs
across the value chain (key question – which partner
has higher inventory management capability and lower

Working Draft - Last Modified

cost of capital? )
Customer
Inventory management

Traditional Reasons for adoption

• Customer demand (most likely)

• Supplier trying to create barrier to switching (e.g
commodity chemicals)
Consignment
Case studies
Supplier

• Dell VMI

Printed
VMI
– Large VMI hubs outside Dell’s assembly locations
in Austin and Memphis
– Sets inventory targets for suppliers based on
supply and demand variability
Supplier Customer – Pays for inventory carrying cost for required
inventory targets – not on balance sheet but
Inventory ownership compliant with SOX
– Suppliers responsible for any overflow beyond
contracted level.
• J&J Cordis Consignment
– Supplier manages inventory levels at hospitals
and clinics, but partly responsible for product
expiry
– Doctors pay for stent after using it on patient.
39
MULTI-ECHELON INVENTORY MANAGEMENT AND PRODUCTION
POSTPONEMENT – PROBLEM DESCRIPTION

DC

Supplier DC C
DC
U
S
Supplier DC T
DC

Working Draft - Last Modified

O
Warehouse Manu-
DC M
facturing
Supplier E
• Raw material • WIP R
DC DC
inventory inventory S

Supplier DC

• Finished goods
inventory

Input parameters Objective

Printed
• For each product by • Minimize total inventory value
customer segment by Constraints
location – demand, • Supply Chain network - production and distribution flow
demand variability, • Bill-of-Materials, yield, lot size for each manufacturing process
required customer • Supply lead time and variability for distribution arcs and process
service levels, order nodes.
lead time distribution,
offered lead time Variables
• Replenishment policy • Inventory levels for each material (FG, RM, intermediate) at
for each location and each location - Safety stock, lead time demand, reorder point
product • CSL at intermediate nodes

40
KEY CONCEPTS IN MULTI-ECHELON INVENTORY MANAGEMENT – 1/2
Concept Concept Description Impact

Multi-level • Inventory stocked at Upstream inventory

flow multiple locations (Raw level has direct
material at supplier and implication on
client DCs) downstream buffer

Working Draft - Last Modified

• Central DC feeding Potential for
Aggregation regional warehouses reduction in demand
• Single component variability if
makes multiple SKUs covariance is zero or
negative

• Assembly

Printed
Inventory levels at
operations both upstream
Assembly
requiring multiple buffers need to be
components sufficient to support
AND downstream buffer

• Alternate sources Total upstream

Alternate of material inventory driving
sources (production and downstream inventory
distribution) performance
OR
41
KEY CONCEPTS IN MULTI-ECHELON INVENTORY MANAGEMENT – 2/2
Concept Concept Description Impact

Order 80
• Customers place Ability to support
lead time 70 orders with different customers from
60
Order volume
50
lead times (request multiple locations in

Working Draft - Last Modified

40 date – order date) the value chain
30

20

10

0
0 50 100 150

Order lead time (days)

Printed
42
Introduction to
Inventory Management
Muthu Krishnan
Alex Niemeyer
Ramesh Raman
Ting Shen
This presentation contains an abstract. If
you are viewing the downloaded version,
please click on "Notes Page" in the View
menu

September 2005

714463-R
© Copyright 2005-2006 McKinsey & Company. Confidential.
Not for further reproduction or distribution.