Keep a written PDSA log so the cycle has memory

Record each cycle so patterns across attempts are visible and the same mistake is not repeated indefinitely.

Why it works

Without a written record, each iteration starts from a distorted memory of the last. Motivated forgetting selectively erases uncomfortable details, and serial repetition of the same error becomes possible. A written log externalizes the cycle’s memory, making the recurrence of a mistake visible and creating the multi-cycle data set that reveals real patterns rather than individual noise.

How to do it

1. Use a simple format: Prediction | What I tried | Result | Gap | One change.
2. Write it the same day as the result, while memory is specific.
3. Review the last three entries before planning the next cycle — look for recurrent gaps.

Evidence

External storage and writing-to-learn research consistently show that externalized records improve recall accuracy and support better self-assessment compared to purely mental review. (observational)

The benefit depends on reviewing the log, not just keeping it; a log that is never re-read provides little advantage over memory.

Common mistake

Keeping the loop "in your head" — which means each cycle depends on a reconstruction of the last one, and the same blind spot recurs across months without ever being named.