Supply Chain Integration and Push-Pull Strategies

Supply Chain Integration and
Push-Pull Strategies
David Simchi-Levi
Professor of Engineering Systems

Massachusetts Institute of Technology
Tel: 617-253-6160
E-mail: dslevi@mit.edu
©Copyright 2021 D. Simchi-Levi
Outline
• A new Supply Chain Paradigm
• Matching Products with Strategies

The Future is Not What it Used to Be
• A new e-Business Model
– Reduce cost
– Increase service level
– Increase flexibility
– Increase Profit

Reality is Different…..
• On-line Furniture Industry
– After an investment of $70M Living.com declared bankruptcy
in August 2000
– Logistics brought the recent downfall of Furniture.com after
about 20 months of operation
• On-line Grocery Industry

– Shoplink.com, Streamline.com, Web van recently retired
from this market
– Peapod escaped this fate when it was sold to Royal Ahold
after heavy losses

Reality is Different…..
• Amazon.com Example
– Founded in 1995
– 1996: $16M Sales, $6M Loss
– 1999: $1.6B Sales, $720M Loss
– 2000: $2.7B Sales, $1.4B Loss
– Last quarter of 2001: $50M Profit
• Total debt: $2.2B

Reality is Different….
• Dell Example:
– Dell Computer has outperformed the competition
in terms of shareholder value growth over the
eight years period, 1988-1996, by over 3,000%
(see Anderson and Lee, 1999)

Channel Type: Retail vs. Online
Retail Online
Low High
Product Variety
Customization Limited High
Forecast Accuracy High Low

Volume by
High Low
Configuration
Cost of Lost Sale Low High
PC-Retail PC-Direct
A new Supply Chain Paradigm
• A shift from a Push System...
– Production decisions are based on forecast
• …to a Push-Pull System

From Make-to-Stock Model….
Suppliers Assembly Configuration

Demand Forecast
• The three principles of all forecasting
techniques:
– Forecasts are always wrong
– The longer the forecast horizon the worst is the
forecast
– Aggregate forecasts are more accurate
• The Risk Pooling Concept


Push-Pull Supply Chains
The Supply Chain Time Line
Suppliers Customers
PUSH STRATEGY PULL STRATEGY
Low Uncertainty High Uncertainty

Push-Pull Boundary

– Parts inventory is replenished based on
forecasts
– Assembly is based on accurate customer
demand

….to Assemble-to-Order Model
Suppliers Assembly Configuration

Business models in the Book
Industry
• From Push Systems...
– Barnes and Noble
• ...To Pull Systems
– Amazon.com, 1996-1999
• And, finally to Push-Pull Systems
– Amazon.com, 1999-present
• 7 warehouses, 3M sq. ft.,

Direct-to-Consumer:Cost Trade-Off
Cost Trade-Off for BuyPC.com
$20
$18
$16
Cost ($ million)
$14 Total Cost

$12 Inventory
$10 Transportation
$8
Fixed Cost
$6
$4
$2
$0
0 5 10 15
Number of DC's

Business models in the Grocery
Industry
• From Push Systems...
– Supermarket supply chain
• ...To Pull Systems
– Peapod, 1989-1999
• Stock outs 8% to 10%
• And, finally to Push-Pull Systems
– Peapod, 1999-present
• Dedicated warehouses
• Stock outs less than 2%
Business models in the Grocery
Industry
• Key Challenges for e-grocer:
– Transportation cost
• Density of customers
– Very short order cycle times
• Less than 12 hours
– Difficult to compete on cost
• Must provide some added value such as convenience

Less than 300,000 shoppers
Number of Average Delivery charges
customers order
Webvan 21000 $71 $4.95 for < $50
free for > $50
Peapod 140000 $120 $7.95 per order
HomeGrocer.com 50000 $110 $9.95 < $75 free

for > $75
NetGrocer.com 60000 $70 $2.99 for < $50
$4.99 for > $50
ShopLink.com 3300 $98 $25 monthly
Streamline.com 3400 $100 $30
Source: D. Ratliff ©Copyright 2021 D. Simchi-Levi

Business Models in the Retail
Industry
• Brick-&-Mortar companies establish Virtual
retail stores
– Wal-Mart, K-Mart, Barnes and Noble
• Use a hybrid approach in stocking
– High volume/fast moving products for local
storage
– Low volume/slow moving products for browsing
and purchase on line
• Channel Conflict Issues

Matching Supply Chain Strategies
with Products
Demand
uncertainty
(C.V.)
Pull H
?
I IV
Delivery cost
III II Unit price
Push L
L H Economies of
Scale
Pull Push
with Products
Demand
uncertainty
(C.V.)
Pull H
?
I IV
Delivery cost
III II Unit price
Push L
L H Economies of
Scale
Pull Push
Case Study: Positioning Inventory
in the Supply Chain
• Manufacturer of consumer packaged goods
– Household goods
• Large and complex network

– Primary and secondary warehouses
• The case study employed InventoryAnalystTM from

LogicTools (www.logic-tools.com)

2-Tier Network for Northeast Region
Suffern, NY (RDC)
feeding forward Whses
Bergen, NJ
Boston, MA
Buffalo, NY
Inventory Optimization Helps Understand the
Bullwhip Effect
Sample Plot of Each SKU by Volume and By Demand Variability
5,000
Order Size Per Week
4,000
3,000
2,000
1,000
0
0.0 0.2 0.4 0.6 0.8 1.0
Order Variability
Inventory Optimization Helps Understand the
Bullwhip Effect

Order Variability is measured using
5,000
COV=STD(Demand)/Avg(Demand)
Order Size Per Week
4,000 Product A B
Volume 10 1000
3,000
STD 1 10
2,000 COV
1,000
0
0.0 0.2 0.4 0.6 0.8 1.0
Order Variability
The underlying demand variability from POS is
significantly smaller than order variability
5,000
4,000
Avg Weekly Sales
3,000
2,000
1,000
0
0.0 0.2 0.4 0.6 0.8 1.0
Demand Variability
Inventory Optimization Determines Specific Hub and
Spoke Strategies by SKU
5,000
4,000
Avg Weekly Sales
3,000
2,000
1,000
0
0.0 0.2 0.4 0.6 0.8 1.0
Demand Variability
This is for a product family of SKU’s within the 40,000 SKU’s.

Each circle represents a SKU
Other drivers include supplier lead time, lead time variability, review period
Inventory Optimization Determines Specific
Hub and Spoke Strategies by SKU

Inventory Optimization
determines that 90% of
5,000
these SKU’s are stored
at the spokes Inventory Optimization
determines that 80% of these
4,000
Avg Weekly Sales
SKU’s are stored at the hubs
3,000
2,000
1,000
Inventory Optimization 0
determines that 55% of 0.0 0.2 0.4 0.6 0.8 1.0
these SKU’s are stored
at the hubs Demand Variability
This is for a product family of SKU’s within the 40,000 SKU’s.

Each circle represents a SKU
Other drivers include supplier lead time, lead time variability, review period
The Long Tail Approach
Popular Products
Popularity
Long Tail
Products
•Is there an opportunity to simplify by managing the tail?

•Does the long tail contribute significantly to revenue?
•Does the long tail include many low margin products?
•What is the hidden cost of products in the long tail?

The “Hidden” Cost of Complexity
R&D Procurement Manufacturing Supply Chain Sales, Mktg,

Support
Spend is inefficient Higher component Higher Higher inventory - Decreasing
prices due to a Manufacturing Finished goods and Customer
decrease in change-over and components in WIP Satisfaction
economies of scale set-up costs
Slower NPI & a Higher Waste and Higher Logistics Increased support
reduction in administration costs obsolescence spend costs, warranty
innovation power to manage the costs
larger portfolio
Resources are Higher prices for Lower utilization of Decrease in Product Increased costs of
scattered products resources Availability poor quality
More resources Increased inbound Increase in quality Poor OTD Brand erosion
required logistics costs control resources performance and among multiple
service levels brands
Lack of visibility to Higher expediting, Broad portfolio is

true costs for handling, & harder to sell
Product warehousing costs
Development Teams

The “Hidden” Cost of Complexity
R&D Procurement Manufacturing Supply Chain Sales, Mktg,

Support
Spend is inefficient Higher component Higher Higher inventory - Decreasing
prices due to a Manufacturing Finished goods and Customer
decrease in change-over and components in WIP Satisfaction
economies of scale set-up costs
Slower NPI & a Higher Waste and Higher Logistics Increased support
reduction in administration costs obsolescence spend costs, warranty
innovation power to manage the costs
larger portfolio
Resources are Higher prices for Lower utilization of Decrease in Product Increased costs of
scattered products resources Availability poor quality
More resources Increased inbound Increase in quality Poor OTD Brand erosion
required logistics costs control resources performance and among multiple
service levels brands
Lack of visibility to Higher expediting, Broad portfolio is

true costs for handling, & harder to sell
Product warehousing costs
Development Teams

Case Study: Demand, variability, and allocated margin
Case Study: High-Tech Company
Tail accounts for

• 35 % of products
• 1.4% of revenue
• 50% of the negative margin products
Allocat
Allocated margin range ed
margin

Lead Time and the Long Tail
For each Category, % of products with a given Lead Time Tail accounts for
• 35 % of products
100% • 1.4% of revenue
90% • 50% of the negative margin products
80%
70% • 10% of total inventory
4_LT>35
60%
50%
• 60% of total cost per unit sold
3_LT: 28-35
2_LT:21-27
40% 1_LT<21
30%
20%
10%
0%
High CV Low CV
Rule: Products in the Long Tail are associated with

Longer Lead Times than other products

High-Tech Transformation
• Retirement of one platform (out of 4)

• Elimination of 65 SKUs (out of 220)
• Inventory down by 15% while introducing two new
platforms
 In contrast to expected increase in inventory, if nothing is
done, by 40% and reduce turns to below three!
• Complete visibility to direct and indirect cost of
complexity
• Established a process to reduce complexity for every
business unit
 Now part of any new product introduction
Case Study: Shifting the Push-Pull
Boundary
• Manufacturer of circuit boards and other high-tech
products
• Sells customized products with high value and short
life cycles
• Multi-stage BOM
– e.g., copper & fiberglass ⇒ circuit board ⇒ enclosure ⇒
processor
• Case study concerns one of 27,000 SKUs
• The case study employed InventoryAnalystTM from
LogicTools (www.logic-tools.com)

How to Read the Diagrams
A Gray Box is a processing stage Number on the lane is
the transit time
Number in the white
box is the commitment
PART 2
0
time to the next stage
DALLAS ($0.50)
5 0
PART 1
30
DALLAS ($260)
2 15
Bins indicate safety
PART 3
88 stock levels- more Red
MONTGOMERY ($220)
15
means more safety
Cost in the box is the stock, empty means no
value of the product safety stock
Number under the box

is the processing time

Push-Pull Boundary
PART 2
Safety Stock Cost = $74,100/yr DALLAS ($0.50)
0
5 0
PART 1
30
PART 4 DALLAS ($260)
7
MALAYSIA ($180) 28 2 15
PART 5 3 PART 3
37 88
CHARLESTON ($12) MONTGOMERY ($220)
3 15
PART 7 4 PART 6
58 70
DENVER ($2.50) RALEIGH ($3)
8
x2
Safety Stock Cost = $45,400/yr PART 2
5
DALLAS ($0.50)
(39% savings)
5 0
PART 1
30
7
MALAYSIA ($180) 28 2 15
PART 5 3 PART 3
37 13
3
15
PART 7 4 PART 6
58 32
©Copyright 2021 D. Simchi-Levi Study used Inventory Analyst™
8
Push-Pull Boundary (cont.)
PART 2
0
5 0
PART 1
30
7
MALAYSIA ($180) 28 2 15
PART 5 3 PART 3
37 88
PART 7 4 PART 6
3 15
Make-to-stock
58 70
RALEIGH ($3)
(push)
DENVER ($2.50)
8
x2
5
DALLAS ($0.50)
(39% savings)
5 0
PART 1
30
7
MALAYSIA ($180) 28 2 15
PART 5 3 PART 3
37 13
3
15
PART 7 4 PART 6
58 32
©Copyright 2021 D. Simchi-Levi Study used Inventory Analyst™
8
PART 2
0
5 0
PART 1
30
7
MALAYSIA ($180) 28 2 15
PART 5 3 PART 3
37 88
PART 7 4 PART 6
3 15
Make-to-stock
58 70
8
(push)
x2
5
DALLAS ($0.50)
(39% savings)
5 0
PART 1
30
7
MALAYSIA ($180) 28 2 15
PART 5 3 PART 3
37 13
3
15
PART 7 4 PART 6
58 32
8
PART 2
0
5 0
PART 1
30
7
MALAYSIA ($180) 28 2 15
PART 5 3 PART 3
Make-to-stock
37 88
3 15
PART 7
DENVER ($2.50)
58
4 PART 6
RALEIGH ($3)
70 (push)
8
x2
5
DALLAS ($0.50)
(39% savings)
5 0
PART 1
30
7
MALAYSIA ($180) 28
Make-to-Stock 2 15
(Push) PART 5
37
3 PART 3
13
Make-to-order
3
15
PART 7 PART 6
(pull)
58 4 32
8
PART 31 6 PART 23 PART 18 PART 11 PART 2
40 50 51 54 55
SEA ($20) DAL ($30) DAL ($35) DAL ($40) DAL ($55)
4 1 3 1
PART 38 PART 32 PART 24 PART 3

8 10 16 50
NJ ($8) NJ ($22) NJ ($30) DAL ($6)
2 50
8 2 6
PART 39 28 PART 33 PART 25 3 PART 19 PART 12 PART 4

5 42 52 61 62 65
TAI ($15) WAS ($30) WAS ($75) DAL ($210) DAL ($260) DAL ($285)
9 3 6 1 3
PART 40 35 PART 34 PART 26 PART 5

12 49 25 4
NZ ($22) WAS ($25) DAL ($80) 2 DAL ($3)
2 3 4
2
PART 35 PART 27 PART 13 8 PART 6

3 2 1 24 46 PART 1
NJ ($35) NJ ($4) MEX ($11) DAL ($18) 30
DAL ($535)
3 1 14
4
PART 41 1 PART 36 PART 28 PART 14 8 PART 7

6 20 17 10 21
PHI ($32) NJ ($40) DAL ($12) MEX ($4) DAL ($9)
13 7 3
PART 42 3 PART 37 PART 29 PART 20 3 PART 15 PART 8

3 10 12 18 26 56
PHI ($2) DAL ($8) WAS ($40) WAS ($42) DAL ($60) DAL ($65)
4 12 6 5 30
PART 21 35 PART 16 PART 9

Safety Stock Cost = $95,000/yr NZ ($18)
41
DAL ($21)
81
DAL ($30)
82
5 1

4 23 26 38
PHI ($6) DAL ($28) DAL ($30) DAL ($35)
3
16 2021 D. Simchi-Levi ©Copyright
12 2007 D. Simchi-Levi
©Copyright
PART 31 6 PART 23 PART 18 PART 11 PART 2
40 21 22 25 26
SEA ($20) DAL ($30) DAL ($35) DAL ($40) DAL ($55)
4 1 3 1
PART 38 PART 32 PART 24 PART 3

6 8 14 26
NJ ($8) NJ ($22) NJ ($30) DAL ($6)
2 50
8 2 6
PART 39 28 PART 33 PART 25 PART 19 PART 12 PART 4

5 10 13 22 23 26
TAI ($15) WAS ($30) WAS ($75) 3 DAL ($210) DAL ($260) DAL ($285)
9 3 6 1 3

12 10 16 4
NZ ($22) WAS ($25) DAL ($80) 2 DAL ($3)
2 3 4
2
PART 35 PART 27 PART 13 8 PART 6

3 2 1 24 26 PART 1
NJ ($35) NJ ($4) MEX ($11) DAL ($18) 30
DAL ($535)
3 1 14
4
PART 41 1 PART 36 PART 28 PART 14 8 PART 7

6 11 16 10 21
PHI ($32) NJ ($40) DAL ($12) MEX ($4) DAL ($9)
13 7 3
PART 42 3 PART 37 PART 29 PART 20 3 PART 15 PART 8

3 9 12 18 26 26
PHI ($2) DAL ($8) WAS ($40) WAS ($42) DAL ($60) DAL ($65)
4 12 6 5 30

41
35 PART 16
25
PART 9
26
NZ ($18) DAL ($21) DAL ($30)
(62% savings) 5 1

4 11 14 26
PHI ($6) DAL ($28) DAL ($30) DAL ($35)
3
16 2021 D. Simchi-Levi ©Copyright
12 2007 D. Simchi-Levi
©Copyright
From Local to Global Optimization
Safety Stock Cost vs. Quoted Lead Time
$100,000
$90,000
Safety Stock Cost ($/year)
$80,000
$70,000
$60,000
Local Optimization
$50,000
Global Optimization
$40,000
$30,000
$20,000
$10,000
$0
0 20 40 60 80 100
Lead Time Quoted to Customer (days)

Study used Inventory Analyst™
$100,000
$90,000 For a given lead-time, the

optimized supply chain provides
$80,000
reduced costs
$70,000
$60,000
Local Optimization
$50,000
Global Optimization
$40,000
$30,000
$20,000
$10,000
$0
0 20 40 60 80 100

$100,000
$90,000 For a given lead-time, the

optimized supply chain provides
$80,000
reduced costs
$70,000
For a given cost, the
$60,000 optimized supply chain
provides better lead-times Local Optimization
$50,000
Global Optimization
$40,000
$30,000
$20,000
$10,000
$0
0 20 40 60 80 100

with Products
Demand
uncertainty
(C.V.)
Pull H
Inventory
Positioning
Pull
A D
Continuous
Replenishment Push
Push L C B
Short Long Lead Time
Pull Push
Matching Demand and Supply
Characteristics
Supply
Uncertainty
•Steal •Cell phone
•Titanium •Flash memory
High
•Inventory •Collaboration
•Information Sharing •Integration

•Portfolio Sourcing
•Dual Sourcing C D
•Efficiency •Computer direct
•Responsiveness
•Cost •Toys
•Lean •Flexibility
Low B A •Fashion
Low High Demand

•Grocery Uncertainty
•Computer retail
•Non fashion apparel ©Copyright 2021 D. Simchi-Levi
Matching Demand and Supply
Characteristics
Demand
Uncertainty
Online CM & OEM

High
Pull Risk-Sharing
C D
Push Build-to-Order Corporate Clients
Low B A
Loose Tight Customer
Relationships
Retail

Supply Chain Integration and Push-Pull Strategies

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Supply Chain Integration and Push-Pull Strategies

Uploaded by

Copyright:

Available Formats

Supply Chain Integration and

Professor of Engineering Systems

©Copyright 2021 D. Simchi-Levi

©Copyright 2021 D. Simchi-Levi

• On-line Grocery Industry

©Copyright 2021 D. Simchi-Levi

©Copyright 2021 D. Simchi-Levi

©Copyright 2021 D. Simchi-Levi

Forecast Accuracy High Low

©Copyright 2021 D. Simchi-Levi

©Copyright 2021 D. Simchi-Levi

©Copyright 2021 D. Simchi-Levi

©Copyright 2021 D. Simchi-Levi

Low Uncertainty High Uncertainty

©Copyright 2021 D. Simchi-Levi

©Copyright 2021 D. Simchi-Levi

©Copyright 2021 D. Simchi-Levi

©Copyright 2021 D. Simchi-Levi

$14 Total Cost

©Copyright 2021 D. Simchi-Levi

©Copyright 2021 D. Simchi-Levi

HomeGrocer.com 50000 $110 $9.95 < $75 free

Streamline.com 3400 $100 $30

Source: D. Ratliff ©Copyright 2021 D. Simchi-Levi

©Copyright 2021 D. Simchi-Levi

• Large and complex network

• The case study employed InventoryAnalystTM from

©Copyright 2021 D. Simchi-Levi

Sample Plot of Each SKU by Volume and By Demand Variability

Sample Plot of Each SKU by Volume and By Demand Variability

Sample Plot of Each SKU by Volume and By Demand Variability

Sample Plot of Each SKU by Volume and By Demand Variability

This is for a product family of SKU’s within the 40,000 SKU’s.

Sample Plot of Each SKU by Volume and By Demand Variability

SKU’s are stored at the hubs

This is for a product family of SKU’s within the 40,000 SKU’s.

•Is there an opportunity to simplify by managing the tail?

©Copyright 2021 D. Simchi-Levi

R&D Procurement Manufacturing Supply Chain Sales, Mktg,

Lack of visibility to Higher expediting, Broad portfolio is

©Copyright 2021 D. Simchi-Levi

R&D Procurement Manufacturing Supply Chain Sales, Mktg,

Lack of visibility to Higher expediting, Broad portfolio is

©Copyright 2021 D. Simchi-Levi

Case Study: High-Tech Company

Tail accounts for

©Copyright 2021 D. Simchi-Levi

Rule: Products in the Long Tail are associated with

©Copyright 2021 D. Simchi-Levi

• Retirement of one platform (out of 4)

©Copyright 2021 D. Simchi-Levi

Number under the box

©Copyright 2021 D. Simchi-Levi

PART 38 PART 32 PART 24 PART 3

PART 39 28 PART 33 PART 25 3 PART 19 PART 12 PART 4

PART 40 35 PART 34 PART 26 PART 5

PART 35 PART 27 PART 13 8 PART 6

PART 41 1 PART 36 PART 28 PART 14 8 PART 7

PART 42 3 PART 37 PART 29 PART 20 3 PART 15 PART 8

PART 21 35 PART 16 PART 9

PART 30 3 PART 22 PART 17 PART 10

PART 38 PART 32 PART 24 PART 3

PART 39 28 PART 33 PART 25 PART 19 PART 12 PART 4