Decide-Calibrate-Retrospect Loop
Make the call, gather feedback, and correct direction quickly
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 5
- Confidence
- 96%
The Decide-Calibrate-Retrospect Loop combines open debate with decisive execution. A leader first invites conflicting views, requiring opinions to be grounded in evidence, historical examples, data, or clearly labeled estimates. The leader then sifts through the complexity and makes a call, because prolonged ambiguity deprives the team of direction. The decision is treated as a current best move rather than proof of infallibility. During execution, the team gathers feedback and compares results with expectations. If the direction is wrong, it calibrates instead of defending the original choice. A retrospective then identifies what the team misunderstood, which signals mattered, and what should change in the next cycle. This mechanism makes decisiveness compatible with humility: clarity accelerates action, while feedback and retrospection keep an imperfect call from becoming a permanent error.
Origin
Mustafa Suleyman describes his leadership style as open, collaborative, evidence-based, and humble, while emphasizing that his job is ultimately to make a call and rely on feedback, iteration, and retrospectives to calibrate.
Core principles
- 01Leaders should hear diverse evidence before deciding
- 02Team clarity is more useful than indefinite deliberation
- 03A wrong direction can be corrected through calibration
- 04Retrospectives turn mistakes into better future decisions
How to run it
- 1
Invite grounded disagreement
Ask for strong, conflicting views and require each view to cite evidence or disclose that it is an estimate. Keep the discussion open and collaborative.
Pro tip Seek disagreement before the call, when it can still improve the decision.
Watch out Volume and confidence should not substitute for evidence.
- 2
Sift and decide
Integrate the competing inputs and make a clear call once further debate has diminishing value. Give the team an actionable direction.
Pro tip State what the team will do now, not only what it believes.
Watch out Do not preserve ambiguity merely to avoid ownership of the decision.
- 3
Observe execution feedback
Track what happens after the decision and create a process for people to surface evidence that the direction is working or failing.
Pro tip Choose observable signals before execution so calibration is not driven by mood.
Watch out Silence from the team is not evidence that the decision is working.
- 4
Calibrate the direction
When evidence contradicts the plan, correct course without treating the initial error as a leadership failure.
Pro tip Explain what changed in the evidence so the correction builds trust.
Watch out Do not confuse consistency with defending a choice after its assumptions fail.
- 5
Retrospect and encode learning
Review the decision, execution, and correction. Identify which assumptions failed and carry the lesson into the next decision cycle.
Pro tip Focus the retrospective on improving the process, not assigning shame.
Watch out A retrospective without a changed future behavior is only a discussion.
In the wild
A team disagrees about releasing a new assistant feature. The leader hears evidence from safety, design, and growth, then approves a bounded launch with explicit usage and complaint signals. Early feedback shows confusion in one workflow, so the team narrows the feature and reviews why its initial user assumption failed.
→ The team moves without waiting for certainty and limits the cost of being wrong.
Common mistakes
Debating without a decision point
Open collaboration becomes paralysis when no one is responsible for converting the evidence into direction.
Treating correction as weakness
Defending a failing direction sacrifices learning and turns a reversible mistake into a larger one.
Skipping the retrospective
The team may correct today's problem but repeat the same decision failure in the next cycle.
Is it for you?
Best for
It is best for teams that need a clear direction now but can observe results and adjust after acting.
Not ideal for
It is not ideal for irreversible decisions where a wrong initial move cannot be safely corrected.
From the transcript
“fundamentally my job is to sift through all the complexity and make a call because there's really, you know, nothing worse than not having clarity…”
“Even if we end up going in the wrong direction, that's totally fine because we'll calibrate.”
“we have a process you know for feedback and iteration and retrospective and that will really really help.”
From the episode
Mustafa Suleyman: Harnessing AI to Transform Work, Business, and Innovation
Mustafa Suleyman