Post-Launch Review Template

RFC: PRJ-XXX Project Name Shipped: [Date] Review type: 2-week / 30-day / 90-day Reviewer: [Reviewer name]

What We Shipped

[1-2 sentences. What went live? Link to changelog, announcement, or PR.]

Persona(s) served: User / Admin / Sponsor

What We Predicted vs What Happened

Pull metrics from the RFC Success section. Fill in actuals.

Metric	Predicted	Actual	Delta

Guardrails: Did anything degrade that shouldn’t have?

Vision Check

Did this move a named outcome and its Signal from the product vision? (Where one number captures the whole product, that Signal is the single headline metric.)

How the reading is built: see the Product Analytics guide.

Direct movement? [Evidence: the outcome’s Signal itself shifted, in the right direction, by a measurable amount]
Leading indicator? [Evidence: a known leading indicator of the Signal moved]
Persona-specific check? [Each affected persona’s vision-lens question improved?]

If none of these improved, why not? What would need to change?

Mechanism Check (90-day review)

The RFC made a commercial mechanism prediction. Score it against what actually happened.

Predicted mechanism (from RFC): [paste verbatim]

Score	Pick one
Fired	Mechanism happened, predicted direction, attributably
Partial	Moved materially less than predicted, or via a different path
Missed	Didn’t fire, or fired the wrong way
Unscoreable	Too vague to score, or attribution impossible. Worse than Missed: it means the bet was never testable.

Score: [Fired / Partial / Missed / Unscoreable]

Which load-bearing assumption broke? (Partial/Missed only. This line is where the learning is; feed it back into the customer model in the Product Spec.)

Attribution note: (external factors, seasonality, confounders that qualify the read)

Prediction scored. The review is not complete until the score is written.

Adoption

Who is using it?

Number of accounts / users actively using the feature
Adoption curve: growing, flat, or declining?
Which persona(s) adopted? Which didn’t?

Who isn’t using it, and why?

Discovery problem? (They don’t know it exists)
Onboarding problem? (They tried and got stuck)
Value problem? (They tried and it didn’t help)
Trust problem? (They don’t believe the signal)

What We Heard

Customer feedback, support tickets, sales conversations since launch.

Source	Signal	Persona

Recommendation

Option	Criteria	When to choose
Accelerate	Exceeding targets + positive feedback	Double down: more investment, broader rollout
Iterate	On track + minor friction	Continue with targeted adjustments
Pivot	Below targets + feedback explains why	Change approach based on what we learned
Investigate	Below targets + unclear why	Dig deeper before deciding; more data needed
Stop	Flat adoption + no pull	Wind down. Redirect engineering time.

Our recommendation: [Accelerate / Iterate / Pivot / Investigate / Stop]

Reasoning: [2-3 sentences. What evidence drives this recommendation?]

Next actions:

[Action 1]
[Action 2]
[Action 3]

Review Cadence

Review	When	Focus
2-week	2 weeks post-launch	Early signal. Is it being used? Any red flags? Quick fixes needed?
30-day	30 days post-launch	Trend check. Adoption trajectory. Feedback patterns emerging.
90-day	60-90 days post-launch	Full review. Data meets decision. Accelerate/Iterate/Pivot/Investigate/Stop.

Schedule all three reviews when you ship, not after. Add calendar invites at launch time.

Stopping is not failure. Failure is continuing to invest in something the data says isn’t working.

Craft — the build loop the Mechanism Check closes
Product Spec: the customer model — where broken assumptions get fed back