PDCA / PDSA Cycle

Origin. Shewhart (1939); Deming, who later renamed it PDSA and objected to 'check'.

Mechanism. Iterative empirical control: a change is a hypothesis, its rollout is an experiment, and the loop closes only if the study step compares outcome against prediction. Deming's substitution of 'study' for 'check' was substantive — 'check' invites inspection for conformance, 'study' requires learning from the discrepancy.

Procedure. Plan: state the change and predict its effect quantitatively. Do: execute at small scale. Study: compare outcome to prediction, and attend specifically to where they differ. Act: adopt, adapt, or abandon. Repeat.

Applies to. Continuous process improvement with fast feedback.

Limitations. Without a prediction in the Plan step, the Study step has nothing to compare against and the cycle becomes 'do things and then look at them'. This is the ordinary state of PDCA in practice.