Back to the Playbook
DefineDD-01
How to define success metrics
Define success before requirements by combining one primary outcome, diagnostic usage and a counter-metric.
Product statusAvailable now
Input3 required inputsFeature outcome and eligible population · Baseline behaviour and available telemetry · Initiative focus and approved Product Brain guidance
SkillsDefine product success measures · Prepare task context · Review an artifact against agreed criteriaProduct Manager + Product Manager Agent
OutcomeSuccess metrics: one primary outcome, usage diagnostics and a counter-metric, each with baseline, window, source, target and owner.Artifact: Success metrics
Owner
Product Manager
Human + Agent
Participants · Agents
Data AnalystEngineerProduct Manager Agent
Trigger
A feature outcome and the population expected to benefit are named.
Required context
Product BrainFeature outcome and eligible populationAnalyticsBaseline behaviour and available telemetryProduct BrainInitiative focus and approved Product Brain guidance
Optional context
Customer evidencePossible adverse effectsConfluenceFeature brief
Skills
Define product success measuresPrepare task contextReview an artifact against agreed criteria
Suggested conversation arc
- 1Agree feature successDescribe the user and business change that would justify the feature. Select one primary outcome and clarify its eligible population.Skills: Prepare task context
- 2Define outcome and guardrailsSpecify the primary measure, usage diagnostics and a counter-metric with baselines, windows, sources and time-bound targets. Flag instrumentation gaps.Skills: Define product success measures
- 3Challenge misleading successAsk how a metric could improve while user value worsens. Check denominator shifts, missing baselines and whether targets are justified by evidence.Skills: Review an artifact against agreed criteria
- 4Confirm the measurement agreementAgree targets and when results will be judged; record unresolved instrumentation work as a requirement, without inventing collected data.
Output
Success metrics: one primary outcome, usage diagnostics and a counter-metric, each with baseline, window, source, target and owner.
Artifacts
Creates: Success metrics
What good looks like
- Metrics can be calculated consistently from their definitions.
- Targets and observations are not conflated.
- Instrumentation gaps are requirements, not invented values.
Quality gate
The Product Manager agrees the targets and the date on which results are judged.
Destination
Product BrainConfluence
Usually next
People and agents work from the same Product Brain. The owner stays accountable. Assigned agents prepare and check. A named person approves at the gate.
Deep dives from the AI PM Lab
Articles that explain the thinking behind this Activity.
Choose the path that matches your role.
Set the Product AI direction with us, or test the shared Product Brain on real work with Jira and code.