Week 4: The Bullwhip Effect

transcript

Week 4:Week 4:The Bullwhip EffectThe Bullwhip Effect

MIS 3537: Internet & Supply Chains

Prof. Sunil Wattal

Learning ObjectivesLearning ObjectivesUnderstand the “bullwhip effect”

Learn what causes the effect

Learn ways to counteract the bullwhip

Have some fun with the beer game

The “Pampers” problemThe “Pampers” problem Pampers – a P&G products

Mild fluctuations in retail sales; excessive fluctuations at distributor level

The orders of materials to the Pamper’s suppliers, eg: 3M, fluctuated even more

Yet, Pampers were “consumed” at the same steady rate

What explains the variability?

At Hewlett-PackardAt Hewlett-PackardSimilar problem at HP – major

printer manufacturer

Fluctuations at HPFluctuations at HPSome fluctuations at retailer level

– understandable

More fluctuations at reseller level

Much greater fluctuations in the manufacturing division’s orders to the IC division

Fluctuations along the Fluctuations along the chainchain

Fluctuations along the Fluctuations along the chainchainConsumer sales are relatively stable

The retailer’s orders show more variability

The wholesaler’s orders show even more variability

The manufacturer’s orders show much greater fluctuations

The bullwhip effectThe bullwhip effectDistorted information from one

end of the supply chain to another creates a fluctuation in the way the various entities behave

The resulting variability in forecasts, orders and inventory levels is called the “bullwhip effect”

What causes the bullwhip What causes the bullwhip effect?effect?Demand forecast updating

Order batching

Price fluctuation

Rationing and shortage gaming

Demand forecast updatingDemand forecast updating Forecasts are generally made using the

exponential smoothing technique

At each level in the supply chain, the input to the demand forecast is the orders from one level below in the supply chain

As orders from the level below keep changing, so does your forecast

Your forecast is the input to the forecast of the entity one level above

As your forecasts vary, their forecasts vary even more

Order batchingOrder batchingA retailer orders products in batches

When demand comes in, the retailer does not order immediately, but accumulates demand and then orders again in batches

This causes a constant ebb and flow

Price fluctuationPrice fluctuation On average, 80% of transactions in the grocery

industry is “forward buy”

Forward buying results in price fluctuations

Also, there are price discounts, quantity discounts, coupons, rebates etc

Hence customers buy in quantities that doesn’t reflect immediate needs

How often have you bought an extra box of corn flakes or an extra bottle of juice because of a temporary price reduction?

Rationing and Shortage Rationing and Shortage GamingGamingScenario: Demand exceeds supply

Manufacturer can ration product supplyIf total supply is only 50% of total demand,

customers will receive only 50% of their order

Knowing this, customers exaggerate their real needs when they order

When demand cools, orders get cancelled

Examples: Sales of DRAM chips in the 1980s; disappearance of HP Laserjet orders

Counteracting the Counteracting the bullwhipbullwhipAvoid multiple demand update

forecasts

Break order batches

Stabilize prices

Eliminate gaming in shortage situations

Avoid multiple demand order Avoid multiple demand order forecastsforecastsWhat causes multiple demand order

forecasts?◦Forecast from one entity becomes input

for higher-level entity◦Simply put, different entities in the

supply chain work with different demand data

To counteract◦Share data◦Create demand forecasts using same

raw data

Avoid multiple forecasts Avoid multiple forecasts (contd.)(contd.)Tools & techniques

Use point-of-sale dataThe actual sale data becomes the raw

data for forecast updates along the system

Electronic Data Interchange / InternetEDI or Internet web services ensures that

the same data is shared across multiple entities at frequent intervals

Computer-assisted ordering

Break order batchesBreak order batchesOrders involve paperwork and red

Also companies offer differential pricing between full-truckload and less-than-truckload transportation

How to counteract?EDITruckloads with different products

Stabilize pricesStabilize pricesProblem

Forward buying leads to price fluctuations

Price discounting leads to uneven demand patterns

How to counteract?Reduce frequency and level of

wholesale price discountingEveryday Low Price / Value pricing

strategy

Eliminate shortage Eliminate shortage gaminggamingProblem

Demand exceeds supplySuppliers order more to counteract

lower supply

How to counteract?Do no allocate products on the basis

of orders aloneAllocate in proportion to past sales

records

Eliminate shortage gaming Eliminate shortage gaming (contd.)(contd.)Information sharing

◦“Shortage gaming” arises due to lack of trust

◦Sharing information can help overcome this

Stop generous return policies!◦Penalties for returns to manufacturers;

this ensures that retailers will not exaggerate needs, and later cancel orders

SummarySummary

The beer gameThe beer gameOriginally conceived at MIT

A very good exercise in understanding the bullwhip effect

The rules…

Setting the contextSetting the contextThe picture shows a simplified beer

supply chain

The arrows show the direction in which the beer flows

Demand information flows in the opposite direction

The rules of the gameThe rules of the gameYou, the player, can play one of the

four roles; you are the manager at theRetailerWholesalerDistributorFactory

The factory has access to unlimited amounts of raw materials, labor etc.

The rules (contd.)The rules (contd.)Each player has to work with the

following parametersOrder: this is the order that you have

received from the next level down your supply chainA retailer’s orders are dependent on the

perceived demand; a wholesaler’s order number is that which has been demanded by the retailer and so on.

Inventory: the numbers in stockBacklog: unfulfilled orders from past

week(s)

The rules (contd.)The rules (contd.)Inventory costs: For every item in

the inventory, the holding entity (retailer etc) is charged $ 0.50

Backorder costs: For every item that is unfulfilled, the entity unable to fulfill the order is charged $ 1.00

It takes two weeks for an order information to move one level up the supply chain

The rules (contd.)The rules (contd.)An example:

◦An order from the retailer reaches the wholesaler in week 2, the distributor in week 3, and the factory in week 4.

◦Each entity tries to fulfill the order with the inventory on hand

◦If not, it becomes a backlog◦It can take as many as 12 weeks to

fulfill an order (retailer wholesaler distributor factory distributor wholesaler retailer)

The rules (contd.)The rules (contd.)The objective of the game

◦To minimize the total supply chain cost (i.e. inventory costs + backorder costs)

Let’s play the game!

Your thoughtsYour thoughtsWhich role did you play?

What were your individual costs? What was the total supply chain cost?

Share your thoughts about the game

What could have helped you bring down the costs?

Next week…Next week…In class Beer Game

Week 4: The Bullwhip Effect

Documents