Maya revisits a decision that worked, finds the reasoning behind it was wrong, and realises nothing in her process would ever have told her
Right For The Wrong Reason — Volume I, Understanding Humans Before Products · Capstone · Historical · Practitioner. Maya revisits a decision that worked, finds the reasoning behind it was wrong, and realises nothing in her process would ever have told her
01 · Opening
Maya revisits a decision that worked, finds the reasoning behind it was wrong, and realises nothing in her process would ever have told her
02 · The setup
A year on from the survey that started all this. Maya is writing the annual review and it is going well.
Maya thinks: Killed the bulk export, built onboarding instead. Churn down four points. Best call I made.
03 · The setup
The year after the decision — Churn: down 4.1 pts; Onboarding completion: up 22 pts; Bulk export, still not built: correct call; Reasoning recorded at the time: none
04 · The evidence
Priya asks a straightforward question and Maya cannot answer it.
Priya: Why did you kill bulk export? I want to use the same reasoning on something.
05 · The evidence
Maya thinks: Because the survey was biased. That is what I remember. Is that what I thought at the time?
06 · The evidence
The thread from that week is still there. Maya reads her own messages and does not recognise the person writing them.
Maya thinks: I killed it because engineering said six weeks. I never mentioned the sample once.
07 · The evidence
What the thread actually says — Reason given at the time: cost, six weeks; Sample bias mentioned: no; Reason remembered a year later: sample bias; Outcome: good either way
08 · The evidence
Maya: I got the right answer by worrying about the wrong thing.
Dev: It still worked.
09 · The evidence
It worked, and the reasoning that produced it would fail the moment a bad idea happened to be cheap.
Maya thinks: Cost as a proxy for value. That rule kills good expensive work and ships bad cheap work.
10 · The evidence
Priya: So the story you have been telling for a year is one you invented afterwards.
Maya: I did not invent it on purpose. That is the part that bothers me.
11 · The evidence
Write the reason down before the outcome exists — One line at the time of the decision, and the prediction that goes with it. Memory rewrites reasoning to fit results, without asking permission and without feeling like invention.
Grade the two separately — A good outcome from bad reasoning is a warning, not a win. A bad outcome from sound reasoning is often the correct call. Scoring only outcomes teaches you nothing, because the loudest signal is luck.
Look hardest at the ones that worked — Failures get examined automatically. Successes never do, so a flawed rule survives inside them until it meets a case where it costs something.
12 · You make the call
The story does not tell you first.
A. It does not — the outcome is what counts
B. The rule that produced it will fail on a case where cost and value diverge
C. It shows she should be more humble about her decisions
13 · What happened
The template gains two lines. What we think will happen, and why we think so, written before anyone knows.
Priya: So next year I can actually reuse your reasoning.
Maya: Or find out it was never worth reusing.
14 · What happened
Maya thinks: A whole year of being right about something I never actually noticed.
15 · Complete
Outcomes are a poor signal for judgment, because luck is louder than reasoning and only failures get examined. Volume II starts where this leaves off, with products rather than the people using them.