FNDFoundationsStop 3 of 7
The bullwhip effect
Why small wobbles in consumer demand become big swings in orders upstream, its four causes, and the fixes.
Read atChange any time
Parents buy nappies at a fairly steady rate; babies don’t suddenly need twice as many. Yet when Procter & Gamble studied its Pampers business, it found orders from distributors swinging much more than retail sales, and its own orders to suppliers swinging more still. Variability grows as you move upstream, like the crack of a whip. That’s the bullwhip effect.
Hau Lee and colleagues named four causes:
- Forecasting from orders, not real demand: each stage sees the orders of the stage below, adds safety stock and reacts to every bump, so a 10% rise at the shelf becomes a 20% rise in the retailer’s order and more at the factory.
- Order batching: stages order in big lumps, such as weekly or monthly or full trucks, so steady demand turns into spiky orders.
- Price fluctuations: promotions and discounts make customers buy ahead (forward buying), then stop.
- Rationing and shortage gaming: when supply is short and allocated in proportion to orders, customers inflate orders, then cancel them when supply recovers.
The result: excess stock and shortages at the same time, idle capacity followed by overtime, and wasted transport. The MIT “beer game” shows that even smart people create bullwhip when each player sees only their own orders.
The fixes match the causes: share real end-customer data (POS data, vendor-managed inventory, CPFR), cut lead times, order in smaller, more frequent batches, stabilize prices (everyday low pricing instead of constant promotions), and allocate shortages on past sales instead of current orders.
Words to use
- Bullwhip effect
- Order variability increasing as you move upstream in a supply chain.
- VMI
- Vendor-managed inventory: the supplier sees the customer’s stock and sales and manages replenishment.
- CPFR
- Collaborative planning, forecasting and replenishment between trading partners.
Watching it happen
Each stage uses a simple, common rule: forecast next week’s demand as last week’s, and keep stock (on hand plus on order) equal to three weeks of that forecast, because the lead time is two weeks plus one week of review.
Consumer demand is 100 a week, then rises to 110 from week 5 and stays there.
| Week | 4 | 5 | 6 | 7 |
|---|---|---|---|---|
| Consumer demand | 100 | 110 | 110 | 110 |
| Retailer’s order to the manufacturer | 100 | 140 | 110 | 110 |
| Manufacturer’s order to its supplier | 100 | 260 | 20 | 110 |
In week 5 the retailer must cover the extra 10 sold and raise its target stock from 300 to 330, so it orders 140, a 40% jump. The manufacturer sees orders jump from 100 to 140 and does the same thing: it orders 140 plus an increase in its target from 300 to 420, a total of 260. Next week the retailer’s orders drop back to 110, the manufacturer’s target falls to 330, and it orders only 20.
A permanent 10% change at the shelf has become a 160% spike followed by an 80% collapse two stages upstream, with every stage following a sensible-looking rule.
Measuring it
The bullwhip ratio of a stage compares the variability of the orders it places with the variability of the demand it receives:
A ratio above 1 means the stage amplifies variability. For an order-up-to policy using a -period moving-average forecast and a lead time of periods, Chen and colleagues (2000) showed the ratio is at least:
With = 2 and = 5, that is at least 2.12. Longer lead times and more reactive forecasts (smaller ) both increase it.
Sharing information helps, but not completely
Chen and colleagues also showed what happens across several stages. If each stage forecasts from the orders it receives, the amplifications roughly multiply: with three stages like the one above, the bound is about . If every stage sees real consumer demand, the bound depends on the total lead time instead, about 6.3 here. Sharing demand data cuts the amplification by about a third in this case, but it doesn’t remove it, because forecasting and lead times still create it. Shorter lead times and smoother policies are needed too.
People make it worse
Sterman’s (1989) experiments with the “beer game” found that players consistently ignore orders they have already placed but not yet received (the supply line). When stock is low they order again and again, then drown in deliveries when everything arrives. Most bullwhip in the game comes from this behaviour, not from the game’s structure, which is why policies and systems that account for the supply line help so much.
What real industries show
Cachon, Randall and Schmidt (2007) looked for the bullwhip effect in US industry data. They found it clearly in many wholesale industries, but less often at retail and in manufacturing, partly because seasonality lets companies smooth production on purpose. The effect is real, but its size depends on the industry, the policies used and how data are measured.
Policies that dampen it
- Share point-of-sale data and plan from consumer demand, through vendor-managed inventory or CPFR.
- Shorten lead times and review periods.
- Smooth replenishment: order-up-to rules that correct only part of each gap per period (proportional rules) reduce variability at a small cost in stock.
- Account for the supply line in every ordering decision.
- Stable prices: everyday low pricing instead of frequent deals that invite forward buying.
- Allocate shortages on history, not on current orders.
The cost of bullwhip
Amplified orders force upstream stages to hold more stock, keep spare capacity, pay for overtime and expediting, and then idle, costs that ultimately reach the consumer price. Dampening the bullwhip is one of the clearest cases where coordinating across companies pays.
Interview kit
Practise it out loud
Answer each question yourself before opening the model answer.
The bullwhip effect is the amplification of order variability as you move upstream. Lee, Padmanabhan and Whang identified four causes: forecasting from orders instead of real demand, order batching, price fluctuations that cause forward buying, and rationing that triggers shortage gaming. The remedies are sharing point-of-sale data through VMI or CPFR, shorter lead times, smaller and more frequent orders, stable pricing, and allocation based on history rather than current orders.
Consumer demand is stable, but your factory sees wild swings in orders. What's going on?
Classic bullwhip. Check whether distributors are forecasting from their own orders, batching orders, buying ahead of promotions, or inflating orders because of past shortages. Then get access to their sell-through or POS data, smooth promotions and order cycles, and agree allocation rules that don’t reward over-ordering.
How does vendor-managed inventory reduce the bullwhip effect?
The supplier sees actual sales and stock at the customer and plans replenishment itself, so it responds to real demand rather than to the customer’s distorted orders. It removes one layer of forecasting and batching.
Connected stops
Where this idea feeds in, or gets its inputs from.
Sources and further reading
- Lee, H. L., Padmanabhan, V. & Whang, S. (1997). The bullwhip effect in supply chains. Sloan Management Review, 38(3), 93–102.
- Sterman, J. D. (1989). Modeling managerial behavior: Misperceptions of feedback in a dynamic decision making experiment. Management Science, 35(3), 321–339.
- Chen, F., Drezner, Z., Ryan, J. K., & Simchi-Levi, D. (2000). Quantifying the bullwhip effect in a simple supply chain: The impact of forecasting, lead times, and information. Management Science, 46(3), 436–443.
- Cachon, G. P., Randall, T., & Schmidt, G. M. (2007). In search of the bullwhip effect. Manufacturing & Service Operations Management, 9(4), 457–479.