Product metrics

Product Metrics

Select metrics that reveal whether the current product decision is producing durable customer and business value.

Inputs

• Decision and owner
• Product and business model
• Target customer and lifecycle
• Current company or product stage
• Baselines, instrumentation, and known data quality issues
• Expected behavior change and possible harms

Workflow

1. State the decision the metrics must inform.
2. Map the value mechanism from customer action to customer outcome to business outcome.
3. Identify the current stage: problem empathy, stickiness, growth loop, revenue, or scale.
4. Choose one focus metric for the current constraint and time horizon.
5. Build a metric tree of controllable input metrics, intermediate behaviors, and lagging outcomes.
6. Add guardrails for customer harm, quality, reliability, cost, and segment inequity.
7. Define each metric precisely: event, population, numerator, denominator, window, exclusions, and owner.
8. Audit instrumentation, missingness, identity resolution, and susceptibility to gaming.
9. Set thresholds and the decision each result triggers.

Metric roles

• Focus metric: current constraint the team intends to move
• Input metric: behavior the team can influence directly
• Outcome metric: customer or business result
• Leading indicator: earlier signal with a validated relationship to the outcome
• Guardrail: unacceptable side effect or boundary
• Diagnostic: helps explain movement without defining success

Output contract

Return the decision, stage and business model, metric tree, focus metric, metric dictionary, guardrails, instrumentation gaps, thresholds, cadence, and owner.

Quality gate

• The focus metric changes a decision.
• A customer-value mechanism connects inputs to outcomes.
• Cohorts and segments are visible where aggregates can mislead.
• Guardrails protect against local optimization.
• Metric definitions are reproducible.

Avoid

• Total signups, page views, or cumulative volume without decision context
• A universal north star detached from stage
• Correlation presented as causal impact
• Metrics the team cannot instrument reliably
• Choosing many priorities to avoid choosing one

Source grounding

Operational synthesis informed by business-model, stage, cohort, vanity-metric, and One Metric That Matters concepts in Lean Analytics, plus outcome principles from Escaping the Build Trap.