Hypothesis-to-Debrief Learning Loop
Set an ambitious hypothesis, examine failure, and apply the lesson fast
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 5
- Confidence
- 96%
This loop begins with a clear hypothesis and an ambitious goal, accepting that uncertain innovation will sometimes fail. The team needs enough safety to commit to the attempt without believing that an honest miss automatically ends a career. After the attempt, it performs a direct debrief: what happened, why did the hypothesis hold or fail, and what should change next? The learning is then applied to subsequent work and communicated through stories so the organisation's behaviour changes rather than the insight remaining local. Nicola compared this muscle to military debriefs after both good and bad operations. The framework does not celebrate failure for its own sake; failure earns value only through self-reflection, evidence, and a changed next action.
Origin
Nicola Mendelsohn described Meta's approach to ambitious goals and debriefing on The Diary of a CEO.
Core principles
- 01Ambitious work creates some failure
- 02A failed hypothesis is useful only when examined
- 03Psychological safety enables honest experimentation
- 04Lessons should change the next attempt
- 05Leaders should retell failures without hiding them
How to run it
- 1
Frame the hypothesis
State what you believe will work and the reasoning behind it.
Pro tip Write the expected signal before running the attempt.
Watch out A vague ambition cannot produce a clear lesson.
- 2
Commit to the experiment
Give the team permission and resources to pursue the ambitious goal seriously.
Watch out Do not call under-resourced work an experiment and blame the team for the result.
- 3
Debrief promptly
Examine what worked, what did not, and why while evidence and memory are fresh.
Pro tip Debrief successful operations too; luck can hide weak reasoning.
Watch out Avoid turning the review into a search for a culprit.
- 4
Convert insight into change
Name the lesson and alter the next decision, process, or hypothesis accordingly.
Watch out A lesson that changes nothing is only commentary.
- 5
Tell the learning story
Share the attempt and its lesson so other teams can reuse it and see that honest reflection is expected.
Pro tip Leaders should include their own misses.
In the wild
A product team predicts that a new format will increase meaningful use, defines the expected behaviour, and runs a bounded release. The metric does not move. In the immediate debrief, the team finds that discovery—not the format itself—blocked adoption, then changes the next test and shares the learning with adjacent teams.
→ The failed result produces a better next hypothesis instead of blame or repetition.
Common mistakes
Celebrating failure without learning
Failure is not progress unless the debrief changes a later action.
Debriefing only bad outcomes
Successful results also need examination because luck can conceal a flawed process.
Is it for you?
Best for
Teams testing uncertain products, methods, or strategic shifts where learning speed matters.
Not ideal for
High-stakes actions that cannot safely fail or be reversed.
From the transcript
“if we're not setting ambitious goals then we then we we wouldn't not be failing”
“what's the debriefing that needs to happen”
From the episode
VP Of Meta: Running A Trillion Dollar Business With An Incurable Illness: Nicola Mendelsohn