Resolution notes and lessons

The resolution screen has two extra fields below the claim cards. They're optional — but they're the difference between a number on a dashboard and a real lesson.

Learning note

A free-text field for the why. The claim marks tell you what happened; the learning note tells you why, and what you'd do differently.

Good learning notes are short, specific, and surface a transferable insight:

"Held on intervention and metric, broke on business outcome. Trial signups lifted but didn't convert because the activation bar in week 2 still required a manual integration. Next time: don't claim a downstream business outcome unless every blocker upstream is also addressed."

This text is what shows up in the Learning Library when other people search for past bets. Treat it as the artefact, not a chore.

"I disagree with the mechanical score"

Tick this if you think the system's mark on one or more claims is unfair — for example, the metric you predicted got contaminated by an unrelated launch, or the data wasn't comparable.

The check is recorded but does not change the score. Calibration scoring is mechanical and never overridable. Two reasons:

• If PMs could regrade their own bets, calibration would drift toward whatever felt good.
• The aggregate disagree-rate is itself a signal — high disagreement on one PM or template tells the Head of Product to investigate.

So tick it freely when you genuinely disagree, but don't expect it to move your headline number.

Previous: Resolving a hypothesis ←    Next: Extending a resolution date →