What does it mean that error cascades?

Small uncorrected errors can trigger chains of increasingly large errors.

Example: You misread a quarterly revenue figure by one decimal place in a spreadsheet. You build a forecast on that number. Your team allocates budget based on the forecast. A hiring plan follows the budget. Three months later, the company has eight new employees it cannot afford — not because anyone made a reckless decision, but because a single transcription error propagated through four layers of dependent decisions without anyone checking the upstream data. Each decision was locally rational. The cascade was already in motion before anyone suspected a problem.

Try this: Pick one decision you made in the past month that led to further downstream decisions — a commitment, a purchase, a delegation, a plan. Trace the chain forward: what subsequent actions depended on that initial choice? Now ask one question about the original decision: what assumption did it rest on, and did you verify that assumption before the downstream actions began? Write down the chain and the unverified assumption. You have just mapped a potential error cascade in your own recent history. Time: 15 minutes.