Decisions

Measurements that support a decision and expose uncertainty.

Compare an intervention with a meaningful reference

A change in a number does not automatically show an improvement caused by an intervention. The reference period may differ in workload, season or measurement method. A baseline makes the comparison explicit enough to examine those differences.

Decisions
Pair a target metric with protective measures

Optimising one measure can move harm elsewhere. A faster process may create more errors; a higher response rate may increase unwanted messages. Guardrails make those side effects part of the decision rather than an afterthought.

Decisions
Choose an observation window before the result

An experiment needs enough time to observe the behaviour it intends to change. A short window may miss delayed effects or ordinary weekly variation. Repeatedly changing the end date after looking at results makes interpretation harder.

Decisions
Design an alert around an action

An alert consumes attention and should lead to a useful response. A threshold crossing alone does not explain who should act, what changed or whether waiting is reasonable. Connect the signal to an operational decision.

Decisions
Keep the reason for a decision with its evidence

A dashboard shows measurements, but it rarely records why a team chose one option over another. A decision record preserves the context, alternatives and limits of the evidence. It helps future readers distinguish a deliberate tradeoff from an accidental default.

Decisions

Search the library