When Supply-Chain Numbers Lose Their History

Suppose customers normally buy 100 units of a product each week. This week, they buy 105.

The retailer does not simply pass along the five-unit increase. Expecting demand to keep rising, or protecting against a stockout, it orders 110 units for the next week. The distributor sees that larger order and raises its own order further. By the time the signal reaches the manufacturer, the request may be for 130 units.

Customer demand rose 5 percent. The signal reaching the manufacturer rose 30 percent.

Supply-chain planners call this the bullwhip effect. In the classic research, the variability of orders can exceed the variability of sales, and the distortion tends to grow as it moves upstream. Forecasting is only one possible cause. Order batching, price changes, and behavior during shortages can also amplify the movement.

The people in the chain do not have to be careless. They may understand every number they receive and make a sensible adjustment for the uncertainty at their own layer.

The distortion can arise from competence applied repeatedly.

The missing history

Imagine that someone estimates a piece of work will take eight weeks. The number moves to the next level of planning, but its history often does not.

Eight weeks might be the estimator’s best forecast. It might be six weeks of expected work plus two weeks of contingency. It might include protection against integration trouble, or have been enlarged because estimates in this organization are routinely cut before approval.

The recipient sees only “eight weeks.” They cannot tell which uncertainties it already includes. Because they remain accountable for the date, adding protection of their own can be perfectly reasonable.

This is different from the game of telephone, where a message mutates accidentally. Here, the information may be understood correctly at every step. The changes are deliberate, and they can all point in the same direction.

Finishing early is often easy to explain; missing a commitment is not. Excess inventory costs money, while a stockout can lose customers. These asymmetries give different people the same reason to lean toward caution. Their judgments become correlated without coordination.

The adjusted number survives. The uncertainty history that produced it disappears.

When caution anticipates caution

Project estimates make the human part of the mechanism especially visible.

A team may pad an estimate because it expects management to reduce it. Management may reduce the estimate because it assumes the team has padded it. Each side then compensates for the other’s expected move: the team adds more protection, and management becomes more skeptical.

The organization is no longer dealing only with unknowns in the project. It must also guess what other people have already done to the numbers describing them.

A base estimate and an explicit reserve communicate different judgments. When they are quietly fused, the next person cannot tell whether they are seeing the expected work, protection against risk, or both.

Trust is part of this loop, but it is not the whole explanation. Even people who trust one another may lack the reasoning behind an earlier adjustment. The problem begins whenever a number preserves the result of judgment while shedding its provenance.

Margins without memory

In aerospace engineering, the same pattern has a name: margin stacking.

Margins are indispensable in aerospace, where they support safety, reliability, and regulatory compliance. Some safety factors are deliberately independent and are supposed to remain layered. Conservatism itself is not the mistake.

The difficulty arises when teams cannot see protection already present elsewhere. A 2026 Design Science study of experienced engineers and managers at an aerospace component company found that margin rationales could be implicit or inconsistently documented. Legacy practices and tacit knowledge could contribute to stacking, while duplicated margins and opaque reasoning could produce excess capacity and design inefficiency.

The final design parameter may be clear while concealing several earlier judgments about uncertainty. A downstream engineer can see the requirement without seeing which risks shaped it. Adding another margin protects the work under their control, even when the same uncertainty may already have been insured upstream.

As in the supply chain, an adjusted signal passes between accountable specialists while the reasons for the adjustment travel less reliably. Each layer responds to what it can see. The result can be more conservative than anyone intended.

Once this pattern is visible, apparently precise numbers begin to look different. A budget request may contain both expected cost and protection against an anticipated cut. A delivery date may describe the work and the accumulated caution of everyone who handled it.

Neither protection is automatically irrational. What matters is what the number can no longer tell us about itself.

Every participant may have acted responsibly. Every adjustment may have made sense where it was made. Yet the final number can embody a level of caution—and a judgment about reality—that nobody actually chose.

Reviewed and edited by John Pile.

Sources and further reading

• Hau L. Lee, V. Padmanabhan, and Seungjin Whang, “Information Distortion in a Supply Chain: The Bullwhip Effect,” Management Science 43, no. 4 (1997).
• Arindam Brahma et al., “Managing uncertainty through margins: Insights from aerospace engineering practice,” Design Science 12, e24 (2026).
• Project Management Institute, “Contingency—Are You Covered?”

Featured image: “Snowball effect” by Tim Green via Flickr, licensed under CC BY 2.0.