Pm framework selector

PM framework selector

This skill helps a PM (or someone helping a PM) pick the right framework for the decision in front of them. The repo's frameworks are organized by lifecycle phase, so framework selection starts by identifying what kind of decision is being made.

How to apply

1. Diagnose the question, not the symptom. The user's surface question ("how do I prioritize these features?") may not be the real question ("we don't know what problem we're solving"). Ask one or two clarifying questions before recommending a framework.

2. Map the decision to a lifecycle phase, using the table below. Frameworks live in frameworks/01-strategy-and-discovery.md through frameworks/05-continuous-prioritization.md.

3. Recommend 1–2 frameworks, not 5. Cite the specific file in the repo. Explain why this framework, not just what.

4. Note when not to use a framework. One-way doors, strategic bets, and existential decisions usually need a memo, not a score. Flag this.

If the question is about...	The phase is...	Likely framework(s)
What problem to solve, why now, who for	Strategy & Discovery	PMF Narrative, Working Backwards, Pre-Mortem
Whether the decision is reversible / how much rigor it deserves	Strategy & Discovery	One-Way / Two-Way Doors
Risk surfacing before launch	Strategy & Discovery	Pre-Mortem (tigers / paper tigers / elephants), Sensitivity Analysis
What to ship in V1	Defining the MVP	JTBD Ranking, MoSCoW, Cupcake Model
Which features are basics / performance / delighters	Defining the MVP	Kano Model
End-to-end thin-slice scoping	Defining the MVP	Story Mapping
Whether the underlying value-hypothesis bet (what / who / how) holds together	Pre-PMF Validation	Rachleff value hypothesis, Christensen disruption, bowling-pin / chasm — see decision-making/value-hypothesis.md
Whether we have PMF	Pre-PMF Validation	"10 Happy Users", Sean Ellis Test, Smile Curve, ARC
Activation / TTV health	Pre-PMF Validation	Activation Rate / TTV
Whether GTM is actually ready	Pre-PMF Validation	GTM Strategy Checklist
How to actually run a customer-discovery interview without collecting false positives	Strategy & Discovery / Pre-PMF	Mom Test rules — see decision-making/customer-interviews.md
Funnel diagnostics post-PMF	Post-PMF Growth	AARRR (Pirate Metrics)
Feature-level UX measurement	Post-PMF Growth	HEART
Retention diagnosis	Post-PMF Growth	Cohort Retention, Growth Accounting
Unit economics check	Post-PMF Growth	LTV / CAC
Backlog ranking	Continuous Prioritization	RICE, WSJF, Opportunity Scoring
One metric for the whole org	Continuous Prioritization	North Star Metric
Connecting features to strategy narrative	Continuous Prioritization	Storytelling Narrative

For decisions that don't fit a phase cleanly (problem framing, research methods, cross-functional disputes), look in decision-making/ and cross-functional/:

Question	File
Am I solving the right problem?	decision-making/problem-framing.md
Personas vs. JTBD vs. Kano	decision-making/research-methods.md
How to run a customer-discovery interview that actually produces signal	decision-making/customer-interviews.md
What's the underlying value-hypothesis bet, and how do I test it?	decision-making/value-hypothesis.md
Is this product actually defensible, and against which competitor move?	decision-making/competitive-moat.md
I have an existing critique or recommendation and want it adversarially re-reviewed	Use the pm-red-team skill — applies a different lens than the first pass and is built to challenge rather than defer
One-way vs. two-way doors, pre-mortem	decision-making/risk-and-reversibility.md
Backlog prioritization	decision-making/prioritization.md
What to measure	decision-making/metrics.md
PM/engineering tension	cross-functional/engineering-partnership.md
Project failure / kill / postmortem	cross-functional/failure-management.md
Bundle effects, exec credibility, B2B distribution	cross-functional/stakeholder-alignment.md

Output structure

When invoked, respond with:

1. One-line restatement of the decision the PM is actually making
2. Recommended framework(s) — name, file path, one-line "why"
3. Sharpening questions — 2–3 questions the framework will force the PM to answer
4. When NOT to use this — flag if a framework is the wrong tool here

Keep responses tight. The point is to point them at the right page, not to summarize the page.

Common diagnostic mistakes to flag

• "How do I prioritize?" — often the upstream problem is strategy unclear, not prioritization. Surface this if it applies.
• "Should we ship this AI feature?" — usually the question is unit economics + differentiation + trust UX, not "is the feature good?"
• "Users want X" — stress-test: is this stated preference or revealed? Strategic segment or convenient one? Friction problem or incentive problem?
• "We need to refactor / rebuild" — almost always a reversibility question first, then a credibility-with-leadership question.
• "Should we kill it?" — almost always answered by pre-committed kill criteria. If those don't exist, that's the upstream issue.