Did that feature work? I don't think anyone checked.

Nobody checks whether it worked

Memory · Blindness · Learn and improve

WHAT’S GOING ON

Decisions get made and results come in, but nobody connects the two. Lessons from reviews get written down and never change anything. So mistakes repeat, and lucky successes teach the wrong lesson.

SIGNS YOU’D NOTICE
  • There's no step for looking back at whether a decision worked.
  • The same mistakes come up across projects.
  • Actions from past reviews sit open.
HOW TO OBSERVE IT

Who: whoever owns this workflow. Look at: three decisions from the last year that came with an expected result. Ask: "Did anyone compare what we expected with what happened?" Count: the actions from the last review, and how many were completed. You're probably looking at this pattern if nobody compared and most actions are still open. Probably not if results get checked and the actions close.

EASY TO CONFUSE WITH

The plan keeps going after the evidence turned. In that pattern, the evidence is in and being ignored. In this pattern, nobody is collecting it.

MOVES TO TRY
Start herePre-registration

Write one line before the pilot. What result would make this a success?

Notes

After the next cycle, spend a few minutes on what we keep and what we change.

Judge the Bet, Not the Result

Ask two separate questions. Was the process sound? Did it work out?

Blameless Review

Ask two questions. What did we know at each step? What would have caught this earlier? Skip the full timeline.

RESEARCH BEHIND THISEstablished

A combined analysis of 46 samples found that structured reviews after the work improved performance by about 25%, with larger gains when someone facilitated them (Tannenbaum and Cerasoli, 2013). Established.