Product Management

Product Management

Use this skill for product decisions that need evidence, trade-offs, and a concrete artifact. It owns discovery framing, PMF measurement, outcome roadmaps, prioritization, and stakeholder decision support. It is not a general PM theory skill.

Quick Reference

When to Use This Skill

• Turn founder notes, customer inputs, or market signals into a roadmap, PMF plan, or decision brief.
• Define activation, retention, guardrails, and business metrics for a product area.
• Prioritize a backlog, set kill criteria, or cut low-value work.
• Build a quarterly product review, opportunity assessment, or strategy narrative.
• Write a product-facing artifact that needs clear trade-offs and measurable outcomes.

Route Elsewhere

• PRDs and implementation-ready specs: use docs-ai-prd.
• GTM motion, ICP choice, or channel strategy: use startup-gtm-strategy.
• Growth experiments and acquisition loops: use startup-growth-execution.
• Product analytics instrumentation and event design: use marketing-product-analytics.
• Architecture or technical target-state design: use software-architecture-design.

Defaults

• Start from the decision, not the document.
• Define metrics with formula, timeframe, and data source.
• Use evidence labels such as strong, medium, and weak when confidence matters.
• Prefer outcome roadmaps over feature lists.
• Require kill criteria or rollback conditions for material bets.
• Measure PMF by segment, not as one blended company-wide score.

Workflow

1. Clarify the decision, horizon, owner, and what would change the recommendation.
2. Choose the artifact type: discovery plan, roadmap, PMF assessment, prioritization, strategy note, or stakeholder brief.
3. Gather only the evidence needed to support that decision.
4. Define success metrics, guardrails, and explicit non-goals.
5. Rank options with one consistent method and document the trade-offs.
6. Produce the artifact plus the next review trigger, not just a static document.

ASCII Flow

Product decision or planning request
  -> Clarify decision, horizon, owner, and review trigger
  -> Select artifact type
     +-- discovery plan -> assumptions, interviews, opportunity map
     +-- PMF assessment -> segment signals, retention, activation, scorecard
     +-- roadmap -> outcomes, bets, guardrails, non-goals
     +-- prioritization -> scoring method, rank, kill criteria
     +-- strategy brief -> recommendation, evidence, trade-offs
  -> Gather only decision-relevant evidence
  -> Define success metrics, data source, timeframe, and guardrails
  -> Rank options and document what will not be done
  -> Return artifact plus next decision or experiment checkpoint

Core Decisions

Discovery and Evidence

Use discovery to de-risk value before building:

• customer interviews for pain, switching behavior, and decision criteria
• assumption tests for risky beliefs
• opportunity mapping when multiple problems compete for attention

If the evidence is thin, say so and define what would increase confidence.

Running the discovery cadence is not the same as learning. Check for discovery theatre — interviews that only confirm, an opportunity tree that hasn't changed shape in a quarter, experiments with no real fail condition — before trusting the artifact.

Prioritization and Saying No

Use one framework consistently:

• RICE or ICE for ranked backlogs
• opportunity scoring for discovery-heavy work
• cost-of-delay or WSJF for time-sensitive flow problems

Minimum control set:

• a scorecard
• kill criteria
• one sentence explaining why lower-ranked work is not being done now

Do not allow stakeholder pressure to replace trade-off documentation.

A scored ranking is not a substitute for judgment. RICE and similar formulas produce false precision from point-estimate guesses — see RICE Precision Theatre for the tells (rankings that never change, ties broken by seniority instead of evidence, zero-to-one bets scored against tactical work on the same stack). Use the framework to force an explicit trade-off conversation, not to end one.

PMF and Retention

PMF is not one survey result. Check:

• Sean Ellis style disappointment or must-have signal
• retention curve shape
• activation that predicts durable retention
• segment-specific PMF rather than blended averages

If the product is liked but not indispensable, tighten the must-have path before adding breadth.

For data-rich products, run the PMF Insight Engine in marketing-product-analytics (assets/pmf-insight-engine.md + 10 blind-spot detectors) to surface signals the team cannot see by intuition. Then score against the appropriate path:

• Path A (B2B / SaaS) — assets/pmf-scorecard-b2b.yaml. Heavier weights on retention curve, NRR, CAC payback (top-quartile is <=6 months; 2025 median is ~16 months; <12 months is a strong/goal benchmark but not top-quartile — do not report it as such at board level), ICP concentration, and value-metric alignment. For usage-based or AI-native products, seat-based PMF assumptions misdiagnose consumption products — use the UBP signals in references/pmf-measurement.md alongside this scorecard.