{ "id": "metrics-frameworks", "name": "Product Metrics Frameworks", "category": "research-frameworks", "summary": "Canonical frameworks for measuring product success. Each framework scopes which metrics to track and how to organize them around a decision. Using the right framework matters more than perfect instrumentation of the wrong one.", "principles_referenced": ["start-with-the-decision", "match-method-to-question", "tie-findings-to-action"], "frameworks": [ { "id": "heart", "name": "HEART (Google)", "summary": "Five-category framework from Google's UX research team: Happiness, Engagement, Adoption, Retention, Task success. For each, pick a Goal → Signal → Metric. Forces breadth; avoids metric monoculture.", "categories": { "happiness": { "what": "Attitudinal — how users feel about the product", "signals": "Satisfaction survey score, NPS, CSAT, App Store rating, written feedback sentiment", "metrics": "Median satisfaction score, % of 5-star reviews, NPS over time" }, "engagement": { "what": "Depth of interaction per user per time window", "signals": "Sessions per user per week, actions per session, content views, time on core surfaces", "metrics": "Weekly active users (WAU) × median session count; share of users who hit core action ≥ N times / week" }, "adoption": { "what": "Share of users who adopt a new feature or capability", "signals": "New user signup rate, feature discovery rate, trial-to-paid conversion", "metrics": "% of eligible users who used feature X within 30 days of launch" }, "retention": { "what": "Users who return over time", "signals": "Week-N retention, churn rate, reactivation rate", "metrics": "Retention curves by cohort (D1/D7/D30 retention); churn rate for paying users" }, "task-success": { "what": "Can users complete their critical tasks effectively?", "signals": "Task completion rate, time on task, error rate, support tickets for task confusion", "metrics": "% of attempts to complete task X that succeed; median time on task; errors per 100 attempts" } }, "when_to_use": [ "You need a balanced scorecard across attitudinal + behavioral + task-level metrics", "The team is overindexed on engagement and needs to also watch satisfaction or task success", "You're setting up UX measurement from scratch and want a complete starting template" ], "pitfalls": [ "Filling all 5 categories with metrics you can instrument, not ones that matter — forces bloat", "Confusing Engagement with success — users can be highly engaged with a broken flow", "Ignoring the Happiness dimension because it's 'soft' — leading indicator of churn", "Stopping at metric definition; skipping the goal-signal-metric chain Google intended" ], "example": "For a calendar app: Happiness = in-app thumbs up/down, target 80% positive. Engagement = events-created per WAU, target 3+/week. Adoption = % of new users who create event in first session, target 70%. Retention = D28 retention, target 50%. Task Success = % of event creations completed without error, target 95%.", "source": "https://www.thinkwithgoogle.com/marketing-strategies/app-and-mobile/how-to-choose-the-right-ux-metrics-for-your-product/" }, { "id": "aarrr", "name": "AARRR / Pirate Metrics (Dave McClure)", "summary": "Funnel-shaped framework from Dave McClure for growth-stage products: Acquisition, Activation, Retention, Revenue, Referral. Simpler than HEART; tailored to product-led growth.", "categories": { "acquisition": { "what": "How users first arrive", "signals": "Channel-level visits, signup rate, landing-page conversion", "metrics": "CAC by channel, visitor-to-signup conversion rate" }, "activation": { "what": "Do new users reach the 'aha' moment?", "signals": "First successful core action, time-to-value, setup completion", "metrics": "% of new signups who complete activation event (must be defined explicitly)" }, "retention": { "what": "Do they keep coming back?", "signals": "Return rate, feature re-use, cohort retention curves", "metrics": "D7 / D30 / W12 retention; churn rate" }, "revenue": { "what": "Do they pay?", "signals": "Free-to-paid conversion, expansion, contraction", "metrics": "ARPU, MRR, net revenue retention" }, "referral": { "what": "Do they invite others?", "signals": "Invite sent, invite accepted, viral K-factor", "metrics": "K-factor (invites sent × acceptance rate); viral cycle time" } }, "when_to_use": [ "Early-stage or growth-stage PLG product", "You need a funnel view from acquisition through revenue", "The business case depends on viral/referral mechanics" ], "pitfalls": [ "Focusing on Acquisition before Retention — leaking bucket problem", "Ill-defined Activation event — usually vague, needs a crisp 'aha' milestone", "Ignoring Referral when there's no genuine viral loop — fake metric" ], "example": "For a note-taking app: Acquisition = landing visits → signups; Activation = created ≥3 notes in first session; Retention = W1 active; Revenue = upgraded to paid within 30 days; Referral = invited ≥1 collaborator in first 60 days.", "source": "https://500hats.typepad.com/500blogs/2007/09/startup-metrics.html" }, { "id": "north-star-metric", "name": "North Star Metric", "summary": "Single primary metric that captures the core value a product delivers to its users. Forcing function for alignment; best paired with a small set of supporting input metrics.", "definition": { "criteria": [ "Measures actual value delivered to users — not just revenue or pageviews", "Leading indicator of long-term business success", "Actionable — teams can directly influence it through their work", "Understandable — a single sentence tells the whole company what to optimize", "Measurable in a timeframe that allows iteration (daily/weekly, not quarterly)" ], "examples": [ "Airbnb: Nights booked", "Spotify: Time spent listening", "Netflix: Hours streamed per user per week", "Slack: Daily active users who have sent ≥2,000 messages", "Facebook: Daily active users", "WhatsApp: Messages sent", "Uber: Weekly completed rides", "Amplitude: Weekly learning queries per user" ] }, "input_metrics": { "what": "The leading indicators that move the North Star", "examples": [ "For Airbnb's 'nights booked': search-to-booking conversion rate, listing supply, host response rate, trust signals", "For Spotify's 'time listening': sessions per user, skip rate, personalized playlist engagement", "For Slack's 'DAU who sent 2k messages': team size at signup, channels created, integrations installed" ] }, "when_to_use": [ "The team needs alignment on one primary growth metric across product/design/eng/marketing", "Revenue is a lagging indicator; you need a leading proxy for long-term value", "The company is overindexed on vanity metrics (downloads, signups) that don't predict success" ], "pitfalls": [ "Choosing a metric that's really revenue in disguise — loses the 'value to user' framing", "Single metric gaming — always pair with guardrail metrics (quality, satisfaction, retention)", "Never revisiting — the right metric at Series A is usually wrong at Series C", "Treating it as the only metric — it's the primary, not the only" ], "source": "https://amplitude.com/blog/north-star-metric" }, { "id": "conversion-funnel", "name": "Conversion Funnel", "summary": "Sequential model of user progression from entry to a valued outcome, with drop-off measured at each step. The foundation of most growth and onboarding work.", "structure": [ "Define the endpoints: entry event (e.g., landing page visit) and success event (e.g., paid subscription)", "List the sequential steps between them", "Measure conversion rate at each step", "Multiply step rates for overall funnel conversion", "Identify the step(s) with the largest drop-off (in absolute users, not just percent)" ], "how_to_use": [ "Start with the largest absolute loss — fixing a 50→30 drop matters more than a 10→7", "Run qualitative research on the worst step to understand why", "A/B test fixes against the worst step", "Re-measure after each fix; declare win only with statistical significance", "Watch for 'conservation of drop-off' — sometimes fixing one step just shifts friction to the next" ], "when_to_use": [ "Any sequential flow: signup, onboarding, checkout, upgrade, content consumption", "You need to identify where users leak before intervening", "Comparing variant flows (old signup vs. new signup) end-to-end" ], "pitfalls": [ "Funnel includes steps users don't actually traverse — breaks the comparison", "Confusing funnel drop-off with churn — funnel is within a single sequence, churn is across time", "Optimizing one step without watching the next — novelty effect or selection effect can fool you", "Measuring funnel conversion on signed-in users when the loss is pre-signup" ], "source": "https://www.nngroup.com/articles/funnel-analysis/" }, { "id": "rice-scoring", "name": "RICE Scoring (Intercom)", "summary": "Prioritization framework: Reach × Impact × Confidence ÷ Effort. Produces a single score per candidate, enabling ranked debate on relative priority.", "formula": "Score = (Reach × Impact × Confidence) / Effort", "variables": { "reach": "How many users this affects per time period (e.g., users per quarter)", "impact": "How much it moves the needle for those users. Use 0.25 / 0.5 / 1 / 2 / 3 scale (minimal / low / medium / high / massive)", "confidence": "How confident are we in the estimates? 50% / 80% / 100%", "effort": "Person-months to ship" }, "when_to_use": [ "Prioritizing across candidate features or initiatives with comparable metrics", "Making tradeoffs explicit in roadmap debates", "Structuring disagreements — 'we disagree on impact, let's talk'" ], "pitfalls": [ "False precision — scores are directional, not exact", "Reach, impact, and confidence are each estimates; compounding them hides uncertainty", "Ignoring strategic bets that score low because reach is small today — some things need qualitative override", "Using RICE as the only input — it's a tool for prioritization debates, not a substitute for them" ], "source": "https://www.intercom.com/blog/rice-simple-prioritization-for-product-managers/" }, { "id": "okrs", "name": "OKRs (Objectives and Key Results)", "summary": "Goal-setting framework popularized by Intel and Google. Objectives are qualitative ambitions; Key Results are quantitative measures of progress. Pair with metrics frameworks — OKRs set direction; metrics track delivery.", "structure": { "objective": "A short, aspirational qualitative statement. 'Make new-user activation a standout moment.'", "key_results": "3–5 measurable outcomes. 'Raise 7-day activation rate from 35% to 50%. Reduce time-to-first-value from 8 min to under 3. Double the rate of new users who share the product.'" }, "best_practices": [ "Objectives should be ambitious — 60–70% achievement is healthy, 100% means targets were too safe", "Key Results must be measurable — vague KRs undermine the system", "Cascade but don't waterfall — team OKRs should reference company OKRs without being strict subsets", "Review quarterly; don't churn objectives mid-quarter without real reason", "Decouple OKRs from performance reviews — otherwise KRs get sandbagged" ], "pitfalls": [ "Every metric is a KR — destroys focus", "Sandbagging KRs to guarantee green — defeats the ambition", "Treating KRs as contracts — they're targets, not SLAs", "Ignoring OKR progress mid-quarter — no learning loop" ], "source": "https://www.whatmatters.com/" } ], "checklist": [ "Have you picked a framework whose structure matches your business stage?", "Does each metric tie to a decision someone on the team can make?", "Are you pairing lagging metrics (revenue, retention) with leading indicators?", "Do you have guardrail metrics alongside primary ones?", "Is the metric definition precise enough that two people would compute the same number?", "Are you reviewing metrics on a cadence that supports action (weekly/monthly)?", "Are you triangulating behavioral metrics with attitudinal signals (surveys, interviews)?" ] }