8/8/2019 Plenary Speaker Hau Lee Pres
1/51
Building Supply ChainExcellence in Emerging
Economies
Hau L. Lee
Stanford University
POMS-Vancouver 2010
8/8/2019 Plenary Speaker Hau Lee Pres
2/51
Source: WTO Report, 2008
Global Trade Growth
8/8/2019 Plenary Speaker Hau Lee Pres
3/51
Globalization Trends
Selling to and Sourcing fromOutside of Home Market
27%
35%
42%
22%
31%
38%
FY02
FY05
FY08
Source: Accenture Global Operations Survey, 2005
Sourcing
Revenue Generated
Growing middle class:
1.5 billion in 52
developing nations
300 million in China, with
1% annual growth Tripled in India since 1985
Emergence of BRICs (Brazil,Russia, India, China):
15% of G6 today Half of G6 by 2025
Surpass G6 in less than 40years
8/8/2019 Plenary Speaker Hau Lee Pres
4/51
Supply Chain Trends
Increasing trend of outsourcing and offshoring toemerging economies.
Challenges of underdeveloped infrastructures inemerging economies.
Complexity and diversity of cultures and needs. Emerging economies are not just a source of supply
and manufacturing, but could become markets andsources of innovation.
But trade barriers, tariffs, special agreements,regulations are also on the rise.
Rising importance of sustainability and socialresponsibility.
8/8/2019 Plenary Speaker Hau Lee Pres
5/51
Key Considerations in Supply Chain
Research with Emerging Economies
Data availability and accuracy
Rapid dynamic changes
Complexity of inter-country trade flows andcultural differences
Different incentives and objectives of different
stakeholders other than profit maximization Need to understand the details of all process
steps
8/8/2019 Plenary Speaker Hau Lee Pres
6/51
Examples of Research Topics
Supply chain design
Product design for diverse market needs
Process efficiency improvements
Business model re-engineering
Socially responsible supply chains
Environmentally sound businesses
Logistics coordination
Innovations and new product development
8/8/2019 Plenary Speaker Hau Lee Pres
7/51
Examples of Research Topics
Supply chain design
Product design for diverse market needs
Process efficiency improvements
Business model re-engineering
Socially responsible supply chains
Environmentally sound businesses
Logistics coordination
Innovations and new product development
8/8/2019 Plenary Speaker Hau Lee Pres
8/51
Network Design
Value chain
design
Location &
deployment
Product
configuration
Total Landed Cost Analysis
Variable cost Setup cost Safety stocks Transit stock
WIP Custom duties Taxes Freight
Market presence Sourcing Compliance Security
8/8/2019 Plenary Speaker Hau Lee Pres
9/51
Increasing Regional Trade Agreements
Source - WTO
8/8/2019 Plenary Speaker Hau Lee Pres
10/51
Renaults
Logan Car
0 20000 40000 60000 80000 100000
Romania
Russia
France
Morocco
Algeria
Turkey
Columbia
Spain
Germany
Ukraine
2006 Sales
8/8/2019 Plenary Speaker Hau Lee Pres
11/51
Network of Agreements
Romania
Turkey
E. Europe
EU
Russia
Brazil
Columbia
India
Morocco
Russia,
Ukraine
Columbia,
Ecuador,
Venezuela
Maghreb
region,
Egypt,
Tunisia,
Jordan
CKD
CBU
Suppliers
Parts
8/8/2019 Plenary Speaker Hau Lee Pres
12/51
Implications for Research
Supply chain design still a very hard problem, inaddition to the usual NP-hard nature of mixed-integerprogramming optimization.
Standard, linear cost of customs and duties may notbe sufficient for supply chain design.
Need the integrate the network of trade agreementswith the bill of materials (the DNA) of products in the
design problem. But trade agreements and government policies changerapidly, so need to develop dynamic models forsupply chain design.
8/8/2019 Plenary Speaker Hau Lee Pres
13/51
Examples of Research Topics
Supply chain design
Product design for diverse market needs
Process efficiency improvements
Business model re-engineering
Socially responsible supply chains
Environmentally sound businesses
Logistics coordination Innovations and new product development
8/8/2019 Plenary Speaker Hau Lee Pres
14/51
Workstation Postponement Boundary
ChassisPower
Supply
Backplane
&I/O Card
Processor
Board Graphics
Board HardDisk
Memory
Postponement Boundary
Manufacturing Distribution
Build-to-Stock (Push) Build-to-Order (Pull)
8/8/2019 Plenary Speaker Hau Lee Pres
15/51
Model
One factory, multiple DCs
Factory builds intermediate products and
stocks at DCs
DCs configures products to order
Promised response time to customers
Imodules, fixed assembly sequence
Total CostMin:ObjectiveP
1
P
P +1
I
Factory
DCs
8/8/2019 Plenary Speaker Hau Lee Pres
16/51
Complications with Emerging Economies
Customs duty rates vary greatly.
Classification of product may result in vastlydifferent duty rates.
Availability of local component supply sourcesvary.
Labor cost and efficiency vary greatly for final
customization process steps. Additional testing sometimes necessary.
Quality and reliability concerns.
8/8/2019 Plenary Speaker Hau Lee Pres
17/51
Base Case Analysis
0
1
2
3
4
5
6
7
8
9
10
Alla
tDC
Backplane
Processor
Graphics
Store&Mem
Alla
tFactory
Postponement Boundary
Supply
Chain Cost
($10M perYear)
SourcingDifferentialProcessingDifferential
Test
Trans.
Duty
Inventory
8/8/2019 Plenary Speaker Hau Lee Pres
18/51
Evolution of Postponement
Factory:Core assembly,customization
Market DC:Distribution
Market DC:Customization,
distribution
DC MarketsSuppliers Factory
Factory:Core assembly
Pre-
Postponement
Market DCPostponement
Offshoring
Offshore
Postponement
Market DC:Customization,
distribution
Factory in LCC:Core assembly
Factory in LCC:Core assembly,customization,
Market DC:Distribution
New
Postponement
Factory in LCC:Core assembly
Near-Market DC:Customization,
distribution
8/8/2019 Plenary Speaker Hau Lee Pres
19/51
Supply Chain Evolution
70s
80s
90s
Today
Mfg Strategy
W. Europe
E. Europe
China
Dual-Response
Drivers
Capability & control
Cost reduction
Cost & supply network
Service pressure, China
cost, oil prices
8/8/2019 Plenary Speaker Hau Lee Pres
20/51
Implications for Research
Optimal product design and postponement strategiesincreasingly valuable as more emerging marketsbecome part of demand points.
Again, need to examine product classification andduty implications for partially assembled product.
Complete considerations of cost and risk componentsnecessary.
Need to consider dynamic, not static, strategies, asmarket conditions change more rapidly in emergingeconomies.
8/8/2019 Plenary Speaker Hau Lee Pres
21/51
Examples of Research Topics
Supply chain design
Product design for diverse market needs
Process efficiency improvements
Business model re-engineering
Socially responsible supply chains
Environmentally sound businesses
Logistics coordination Innovations and new product development
8/8/2019 Plenary Speaker Hau Lee Pres
22/51
Stanford Trade Process Model*
Pre-Export
TransportArrangement &
Export Filing
TransportExecution &CustomsClearance
Post ImportSettlement &Reconciliation
Order Collaboration
Classification
Compliance MgtTrade Finance Set Up
Manufacturing
Freight Booking
Goods Picked, Packed
Export DeclarationSecurity Filings
Shipping Doc Prep
Goods Transported
Goods Tracking
Customs ClearanceFinance Doc Prep
Goods Received
Data Reconciliation
Loan ProcessingFinancial Settlement
Duty Relief Mgt
* Joint research with Warren Hausman, sponsored by Tradebeam
8/8/2019 Plenary Speaker Hau Lee Pres
23/51
Information-Based Innovations
ProcessExcellence
Process
Redesign
Focus
Faster
More accurate
More reliable
Re-sequencing
Parallel processing
Elimination
Values
Shorter cycle time
Less delays & reworks Lower capital tied up
Faster cash cycle
Less penalties from
errors
Accurate duty payment
and refunds
8/8/2019 Plenary Speaker Hau Lee Pres
24/51
IT-Enabled Benefits
0
20
40
60
80
100
120
Manufacture
to Invoice
Cycle
Days
Days of
Sales
Outstanding
Order to
Receipt
Cycle
0%
5%
10%
15%
20%
25%
30%
35%
Benefit as % o f Sales Pro fit Increase (%) In termed iaries as % of their
sales
Base Case
IT-Enabled
Exporter
Importer
* Joint research with Warren Hausman, China-US apparel trade lane
8/8/2019 Plenary Speaker Hau Lee Pres
25/51
Implications for Research
Process improvement must be grounded on deepunderstanding of how the process works, but there aremuch greater complexity and diversity of processflows in emerging economies.
Data is a big challenge: needs to do more empiricalwork (interviews, ethnographical methods, secondarydata, etc.).
Powerful link of IT and process improvements.
Since some investments are necessary, it isworthwhile to do research on the values of processimprovements.
8/8/2019 Plenary Speaker Hau Lee Pres
26/51
Examples of Research Topics
Supply chain design
Product design for diverse market needs
Process efficiency improvements
Business model re-engineering
Socially responsible supply chains
Environmentally sound businesses
Logistics coordination Innovations and new product development
8/8/2019 Plenary Speaker Hau Lee Pres
27/51
Soy Bean Supply Chain in India
Farmers Mandis
ITC
Other Buyers
Inbound
logisticsin bulk
Display & visual
inspection
Auction (spot market)
Outbound logistics
& payment
Congestion at Mandis can hold up farmers for days.
Price uncertainty faced by farmersmaterial flow must precede price
information discovery.
Farmers rely on word of mouth or localized previous prices.
Farmers vulnerable to delayed payments.
Source: Anupindi & Suvakumar, 2007
8/8/2019 Plenary Speaker Hau Lee Pres
28/51
ITC e-Choupal
Farmers ITC Processing
Other Buyers
Selected farmers (Sanchalak) maintain computers and inspect products. Price transparency provided to farmers.
Direct sales to ITC enabled.
Material flows occur only after sales to ITC or price discovery to Mandis.
Mandis
e-Choupal
Internet KioskMarket conditions,prices, weather,
knowledge
Mandi
Prices
Authorized
Prices
Source: Anupindi & Suvakumar, 2007
8/8/2019 Plenary Speaker Hau Lee Pres
29/51
Win-Win Value Proposition
370 375
120 215
0 200 400 600 800
Before
After
Farmer
ITC
Transaction Costs in Rupees per metric ton
Source: Anupindi & Suvakumar, 2007
8/8/2019 Plenary Speaker Hau Lee Pres
30/51
ITC e-Choupal
6,000 e-Choupals & more than120 hubs in 9 states, 36,000villages; empowering 3.5million e-farmers. 6,000 e-Choupal installations
Started with soya bean, nowcovering wheat, coffee, shrimpand other aqua products.
Reverse flow of fertilizers,household goods, seeds, etc.
back to farmers. Aims at extending to 100,000
villages and 10 millione-farmers in next decade.
8/8/2019 Plenary Speaker Hau Lee Pres
31/51
Modeling ITC and Farmers Incentives*
ITC e-Choupal investments helped reduced
production costs of farmers
Farmers in network benefit directly, and can sell
directly to ITC (implicit agreement) or to mandis. Farmers outside of network can also benefit, based on
e-Choupals investment, but have to sell to mandis.
Would farmers honor implicit agreement, and is itworthwhile for ITC to provide training to non-
network farmers?
* Chen et al., Working Paper, 2008.
8/8/2019 Plenary Speaker Hau Lee Pres
32/51
Decisions by Players
Training Investmentsfor Non-Network Farmers
Quantity Producedfor ITC and Mandis
Quantity Producedfor Mandis
Also, need to model prices (endogenous or
exogenous) and capacity of ITC.
ITCIn-NetworkFarmers Out-of-NetworkFarmers
* Chen et al., Working Paper, 2008.
8/8/2019 Plenary Speaker Hau Lee Pres
33/51
Some Results
Implicit agreement between farmers and ITC
behaves like a formal agreementfarmers
always give priority to sell directly to ITC.
It is also optimal for ITC to provide the best
training investment to out-of-network farmers,
since ITC also needs to purchase from open
market (mandis).
* Chen et al., Working Paper, 2008.
8/8/2019 Plenary Speaker Hau Lee Pres
34/51
Implications for Research
Emerging economies offer great opportunity forbusiness model re-engineering that can directlyimprove business results and serving the mass.
Analyzing differential values to different direct andindirect stakeholders of such changes important.
Fertile ground for incentive alignment research.
Parallel research on cost efficiency & agility of ultra-
low cost products for extreme affordability (e.g., onelap-top per child, Tata Nano cars, D.Light, andShanzhai cell phones, etc.)
8/8/2019 Plenary Speaker Hau Lee Pres
35/51
Examples of Research Topics
Supply chain design
Product design for diverse market needs
Process efficiency improvements
Business model re-engineering
Socially responsible supply chains
Environmentally sound businesses
Logistics coordination Innovations and new product development
8/8/2019 Plenary Speaker Hau Lee Pres
36/51
Causes of Deaths in 2002
0 1000 2000 3000 4000 5000 6000
Infectiousdiseases
Respiratoryinfections
Maternal
Preinatal
Nutritionaldeficiencies
Otherdiseases*
Injuries
* Cancer, diabetes, neuropsychiatric disorders, cardiovascular diseases and respiratory diseases, etc.
US, Canada,Cuba
Africa
8/8/2019 Plenary Speaker Hau Lee Pres
37/51
Maternal Mortality Ratio(per 100,000 livebirths, 2008)
Hogan et al.,Lancelot, 2010
8/8/2019 Plenary Speaker Hau Lee Pres
38/51
Health Care in Africa
Short life expectancy: 49 years vs 77 in the US.
6% of population, or 25 million, had HIV/AIDS.
Infrastructure and logistics challenges: No integrated road, transportation and logisticsnetwork
Rural communities (with 62% of population)accessible only by single-lane sand or dirt paths.
Only 20-30% lived within 2 km of a road.
Big problem of the last mile to deliver medicineand health care to the bulk of the population.
8/8/2019 Plenary Speaker Hau Lee Pres
39/51
Motorcycle as Solution
Ability to access terraininaccessible by four-wheelvehicles.
Lower cost to purchase andoperate.
Fewer and less complicatedtools to maintain.
Lower carbon emissions than
four-wheel vehicles. Less subject to abuse by
corrupt officials.
8/8/2019 Plenary Speaker Hau Lee Pres
40/51
Transport Resource Management
Standardization of vehicle fleets. Preventive maintenance practices:
Self pre-ride checks by health workers
Monthly service by Riders-trained technician.
Extend useful life of motorcycles from 20,000 km to80,000 km.
Hub-and-spokes model for service support.
Spare parts inventory management.
Fixed service cost per km model to cover localhub, service parts, gas, protective gear, training and
logistics management.
Based on Stanford case, 2008
8/8/2019 Plenary Speaker Hau Lee Pres
41/51
Impact (2001-2002)
% of population indicating
increased visits100%
8%
Gokwe Binga
% change in malarial deaths
-0.21
0.44
Gokwe Binga
Frequency of health worker visits
49%
22%
35%
5%
13%
68%
3%5%
Gokwe Binga
Never / do notremember
Less thanmonthly
Monthly
Weekly
Based on Stanford case, 2008
8/8/2019 Plenary Speaker Hau Lee Pres
42/51
Research Problems*
How to measure effectiveness of programLogistics improvements
Health workers productivity improvements
Public health improvements Incentive issues and resulting performance
under different business models
Predictive models of performancesDifferent geographies, needs and terrainsDifferent business models
* Research grant to Stanford University from Gates Foundation
8/8/2019 Plenary Speaker Hau Lee Pres
43/51
Business Models with Governments
Business models used and proposed:
On-demand service
Managed care based on mileage
Total ownership model.
Incentives create different behaviors on:
Maintenance and service schedule
Inventory and stocking depot control
Vehicle retirement decisions
8/8/2019 Plenary Speaker Hau Lee Pres
44/51
A Vehicle Replacement Model*
Used field data to 44 vehicle replacements atIFRC relative to stated IFRC policy (5 year or
150,000 km).
Developed optimal replacement policy andanalyzed current policys effectiveness.
Data from 2002-6 for Afghanistan, Ethiopia,
Georgia and Sudan, covering 449 vehicles. Significance of problem due to IFRC policy
being used by many other NGOs.
* Martinez and Van Wassenhove, working paper, 2009
8/8/2019 Plenary Speaker Hau Lee Pres
45/51
Data is the Beast
Field data includes: procurement cost, vehicle
operation mileage, accident records,
maintenance costs (preventive & others), and
salvage values.
Data in records not totally reliable.
Internal data migration errors.
Personal usage not captured.
* Martinez and Van Wassenhove, working paper, 2009
8/8/2019 Plenary Speaker Hau Lee Pres
46/51
Stated Policy Not Followed
Odo
meter(000
km)
Below 5 Above 5
> 150
< 150
Age in Years
16
110
84
5
No. of Retired Vehicles and
Their Ages and Odometers
HQ purchased vehicles,
National Delegation
(ND) responsible for
operational &
maintenance costs, andpaid depreciation cost to
HQ up to 5 years.
Revenue from sale of
retired vehicle went toHQ.
Interesting incentive
problem arose.
* Martinez and Van Wassenhove, working paper, 2009
8/8/2019 Plenary Speaker Hau Lee Pres
47/51
Optimal Replacement Policy
Develop analytical models of operational cost
and salvage value as a function of age and
odometer.
Formulate dynamic program for optimal
replacement policy.
Optimal replacement odometer found at
100,000 km.
Average savings of 8.7% can be achieved.
* Martinez and Van Wassenhove, working paper, 2009
8/8/2019 Plenary Speaker Hau Lee Pres
48/51
Riders for Health Project*
Pilot to start in 20 counties of Gambia (withvariations of business models), eventually
rolling out to whole country & other countries.
Measurement systems are to be put in place:daily log records, GPS devices on vehicles,
and on-site manual audits, etc.
Predictive model building. Vehicle operational and replacement modeling
based on different business models.
* Research grant to Stanford University from Gates Foundation
8/8/2019 Plenary Speaker Hau Lee Pres
49/51
Implications for Research
Data is still the beastneed more creative methods tocollect data, or ways to use coarse data in research.
Performance measurement researchwhat is theright measure that would link to efficient use ofresources for social good?
Big incentive alignment problem, but stakeholderscan include operators, government, NGOs, donors,and the general mass.
Implications for regular commercial supply chainmanagement.
8/8/2019 Plenary Speaker Hau Lee Pres
50/51
Key Issues of Research Topics
Supply chain networkdesign
Product design fordiverse market needs
Process efficiencyimprovements
Business model re-engineering
Socially responsiblesupply chains
8/8/2019 Plenary Speaker Hau Lee Pres
51/51
Final Remarks
Great opportunities exist for OM research involvingemerging economies.
Use detailed field/teaching cases to build right
models. Data is the challenge, but sometimes support for
collaborations is there.
Beware of the dynamicsboth complicates but also
enriches your problem.
Incentive issues are BIG but exciting research
problems.