{"id":"188d2a65-8f77-455a-9c1c-6355ba87648e","entityType":"agent","slug":"clawhub-deciqai-aarrr-pirate-metrics","name":"AARRR (Pirate Metrics)","canonicalUrl":"https://www.xpersona.co/agent/clawhub-deciqai-aarrr-pirate-metrics","canonicalPath":"/agent/clawhub-deciqai-aarrr-pirate-metrics","generatedAt":"2026-10-10T23:48:30.544Z","source":"CLAWHUB","claimStatus":"UNCLAIMED","verificationTier":"NONE","summary":{"evidence":{"source":"editorial-content","verified":true,"confidence":"high","updatedAt":"2026-10-10T19:44:42.349Z","emptyReason":null},"description":"Activate when: user says 'our growth is stalling and we don't know why,' 'product and marketing are arguing about whose fault it is,' 'where is our funnel br... Skill: AARRR (Pirate Metrics) Owner: deciqai Summary: Activate when: user says 'our growth is stalling and we don't know why,' 'product and marketing are arguing about whose fault it is,' 'where is our funnel br... Tags: latest:1.0.6 Version history: v1.0.6 | 2026-07-16T17:51:05.759Z | user Description tail link + agents machine-readable metadata line (deciqai.com/s/aarrr-pirate-metrics.json) v1.0.5 | 2026-07-09T11:1","descriptionLabel":"Technical summary","evidenceSummary":"Capability contract not published. No trust telemetry is available yet. 1.3K downloads reported by the source. Last updated 10/10/2026.","installCommand":"clawhub skill install s17a4mqcnk515kvaca5ze55d0x88pfpx:aarrr-pirate-metrics","sourceUrl":"https://clawhub.ai/deciqai/aarrr-pirate-metrics","homepage":"https://clawhub.ai/deciqai/skills/aarrr-pirate-metrics","primaryLinks":[{"label":"View on ClawHub","url":"https://clawhub.ai/deciqai/aarrr-pirate-metrics","kind":"source"},{"label":"Homepage","url":"https://clawhub.ai/deciqai/skills/aarrr-pirate-metrics","kind":"homepage"}],"safetyScore":84,"overallRank":62,"popularityScore":62,"trustScore":null,"claimedByName":null,"isOwner":false,"seoDescription":"Activate when: user says 'our growth is stalling and we don't know why,' 'product and marketing are arguing about whose fault it is,' 'where is our funnel br..."},"coverage":{"evidence":{"source":"public-profile","verified":false,"confidence":"medium","updatedAt":"2026-10-10T19:44:42.349Z","emptyReason":null},"protocols":[{"protocol":"OPENCLEW","label":"OpenClaw","status":"self-declared","notes":"Declared in the public agent profile."}],"capabilities":[],"verifiedCount":0,"selfDeclaredCount":1,"capabilityMatrix":{"rows":[{"key":"OPENCLEW","type":"protocol","support":"unknown","confidenceSource":"profile","notes":"Listed on profile"}],"flattenedTokens":"protocol:OPENCLEW|unknown|profile"}},"adoption":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-10T19:44:42.349Z","emptyReason":null},"stars":null,"forks":null,"downloads":1276,"packageName":null,"latestVersion":"1.0.6","tractionLabel":"1.3K downloads"},"release":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"medium","updatedAt":"2026-10-10T19:44:42.349Z","emptyReason":null},"lastUpdatedAt":"2026-10-10T19:44:42.349Z","lastCrawledAt":"2026-10-10T19:44:42.349Z","lastIndexedAt":null,"nextCrawlAt":"2026-10-11T19:44:42.349Z","lastVerifiedAt":null,"highlights":[{"version":"1.0.6","createdAt":"2026-07-16T17:51:05.759Z","changelog":"Description tail link + agents machine-readable metadata line (deciqai.com/s/aarrr-pirate-metrics.json)","fileCount":7,"zipByteSize":15742},{"version":"1.0.5","createdAt":"2026-07-09T11:14:51.577Z","changelog":"Refresh: 2024-2026 AI-era worked examples added (strategy/leadership + systems/game-theory batch)","fileCount":7,"zipByteSize":15590},{"version":"1.0.4","createdAt":"2026-07-08T10:52:52.785Z","changelog":"Footer now uses /c/ short link (fixes UTM truncation when SKILL.md is read in a terminal)","fileCount":6,"zipByteSize":11292},{"version":"1.0.3","createdAt":"2026-07-08T03:50:18.702Z","changelog":"Second primary-sourced worked example","fileCount":6,"zipByteSize":11478},{"version":"1.0.2","createdAt":"2026-07-08T00:37:20.500Z","changelog":"Refreshed content + GitHub star link in footer","fileCount":5,"zipByteSize":8849},{"version":"1.0.1","createdAt":"2026-07-07T20:29:04.812Z","changelog":"Add catalog categories and topics","fileCount":5,"zipByteSize":8754},{"version":"1.0.0","createdAt":"2026-06-26T03:51:50.969Z","changelog":"Initial publish","fileCount":5,"zipByteSize":8870}]},"execution":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No published capability contract is available yet."},"installCommand":"clawhub skill install s17a4mqcnk515kvaca5ze55d0x88pfpx:aarrr-pirate-metrics","setupComplexity":"low","setupSteps":["Setup complexity is classified as HIGH. You must provision dedicated cloud infrastructure or an isolated VM. Do not run this directly on your local workstation.","Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data."],"contract":{"contractStatus":"missing","authModes":[],"requires":[],"forbidden":[],"supportsMcp":false,"supportsA2a":false,"supportsStreaming":false,"inputSchemaRef":null,"outputSchemaRef":null,"dataRegion":null,"contractUpdatedAt":null,"sourceUpdatedAt":null,"freshnessSeconds":null},"invocationGuide":{"preferredApi":{"snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-aarrr-pirate-metrics/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-aarrr-pirate-metrics/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-aarrr-pirate-metrics/trust"},"curlExamples":["curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-aarrr-pirate-metrics/snapshot\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-aarrr-pirate-metrics/contract\"","curl -s \"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-aarrr-pirate-metrics/trust\""],"jsonRequestTemplate":{"query":"summarize this repo","constraints":{"maxLatencyMs":2000,"protocolPreference":["OPENCLEW"]}},"jsonResponseTemplate":{"ok":true,"result":{"summary":"...","confidence":0.9},"meta":{"source":"CLAWHUB","generatedAt":"2026-10-10T23:48:30.537Z"}},"retryPolicy":{"maxAttempts":3,"backoffMs":[500,1500,3500],"retryableConditions":["HTTP_429","HTTP_503","NETWORK_TIMEOUT"]}},"endpoints":{"dossierUrl":"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-aarrr-pirate-metrics/dossier","snapshotUrl":"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-aarrr-pirate-metrics/snapshot","contractUrl":"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-aarrr-pirate-metrics/contract","trustUrl":"https://www.xpersona.co/api/v1/agents/clawhub-deciqai-aarrr-pirate-metrics/trust"}},"reliability":{"evidence":{"source":"runtime-metrics","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No trust, reliability, or runtime telemetry is available."},"trust":{"status":"unavailable","handshakeStatus":"UNKNOWN","verificationFreshnessHours":null,"reputationScore":null,"p95LatencyMs":null,"successRate30d":null,"fallbackRate":null,"attempts30d":null,"trustUpdatedAt":null,"trustConfidence":"unknown","sourceUpdatedAt":null,"freshnessSeconds":null},"decisionGuardrails":{"doNotUseIf":["Contract metadata is missing or unavailable for deterministic execution."],"safeUseWhen":[],"riskFlags":["missing_or_unavailable_contract","trust_data_unavailable","schema_references_missing"],"operationalConfidence":"low"},"executionMetrics":{"observedLatencyMsP50":null,"observedLatencyMsP95":null,"estimatedCostUsd":null,"uptime30d":null,"rateLimitRpm":null,"rateLimitBurst":null,"lastVerifiedAt":null,"verificationSource":null},"runtimeMetrics":{"successRate":null,"avgLatencyMs":null,"avgCostUsd":null,"hallucinationRate":null,"retryRate":null,"disputeRate":null,"p50Latency":null,"p95Latency":null,"lastUpdated":null}},"benchmarks":{"evidence":{"source":"no-benchmark-data","verified":false,"confidence":"low","updatedAt":null,"emptyReason":"No benchmark suites or observed failure patterns are available."},"suites":[],"failurePatterns":[]},"artifacts":{"evidence":{"source":"CLAWHUB","verified":false,"confidence":"high","updatedAt":"2026-10-10T19:44:42.349Z","emptyReason":null},"readme":"Skill: AARRR (Pirate Metrics)\n\nOwner: deciqai\n\nSummary: Activate when: user says 'our growth is stalling and we don't know why,' 'product and marketing are arguing about whose fault it is,' 'where is our funnel br...\n\nTags: latest:1.0.6\n\nVersion history:\n\nv1.0.6 | 2026-07-16T17:51:05.759Z | user\n\nDescription tail link + agents machine-readable metadata line (deciqai.com/s/aarrr-pirate-metrics.json)\n\nv1.0.5 | 2026-07-09T11:14:51.577Z | user\n\nRefresh: 2024-2026 AI-era worked examples added (strategy/leadership + systems/game-theory batch)\n\nv1.0.4 | 2026-07-08T10:52:52.785Z | user\n\nFooter now uses /c/ short link (fixes UTM truncation when SKILL.md is read in a terminal)\n\nv1.0.3 | 2026-07-08T03:50:18.702Z | user\n\nSecond primary-sourced worked example\n\nv1.0.2 | 2026-07-08T00:37:20.500Z | user\n\nRefreshed content + GitHub star link in footer\n\nv1.0.1 | 2026-07-07T20:29:04.812Z | user\n\nAdd catalog categories and topics\n\nv1.0.0 | 2026-06-26T03:51:50.969Z | user\n\nInitial publish\n\nArchive index:\n\nArchive v1.0.6: 7 files, 15742 bytes\n\nFiles: examples/ai-saas-funnel-leaks-2023-2026.md (6925b), examples/dropbox-referral-program-2009.md (4627b), examples/facebook-7-friends-in-10-days-2008.md (4428b), references/sources.md (2384b), skill-card.md (2805b), SKILL.md (9193b), _meta.json (139b)\n\nFile v1.0.6:SKILL.md\n\n---\nname: aarrr-pirate-metrics\ndescription: \"Activate when: user says 'our growth is stalling and we don't know why,' 'product and marketing are arguing about whose fault it is,' 'where is our funnel breaking,' 'pirate metrics,' or 'acquisition vs activation vs retention.' Also activate for: designing end-to-end metrics for a new product, building a shared instrumentation framework across teams.\n Do NOT activate when: the product has no users yet (use lean-startup instead), or the bottleneck is already obvious and obvious to fix. More: deciqai.com/c/aarrr-pirate-metrics\"\n---\n\n# AARRR (Pirate Metrics)\n\n## Overview\n\nA startup's growth is a **sequential funnel** — each stage gates the next. Great Acquisition is worthless if Activation is broken; great Activation is worthless if Retention is zero. Optimizing the wrong stage produces work that looks like progress while the bottleneck stays.\n\nThe **AARRR framework** (Dave McClure, *Startup Metrics for Pirates*, 2007) names five stages: Acquisition (do they show up?), Activation (good first experience?), Retention (do they come back?), Referral (do they tell others?), Revenue (do they pay?). The bottleneck stage governs total growth — improving any other stage produces no system-level gain (Goldratt, *The Goal*, 1984).\n\n**Compose with:** first-principles to identify your specific Activation event; pmf-crossing-the-chasm to recognize when Retention will always be the bottleneck pre-PMF; probabilistic-thinking to set base rates per stage.\n\n## When to Use\n\n- Growth is **stalling** and the cause is not obvious — multiple teams have plausible explanations\n- Product, marketing, and sales are **arguing about whose problem it is**\n- You need a **shared instrumentation framework** across teams\n- Designing **end-to-end metrics for a new product**\n- Someone says: *\"AARRR,\" \"pirate metrics,\" \"growth funnel,\" \"where is our funnel breaking?\"*\n- An **AI-native product has strong signups but weak activation/retention** — cheap AI-hype acquisition masks empty-state and post-novelty leaks; or you're facing AI-native competition on a commoditized model and need to find where your funnel actually loses users\n\n**When NOT to use:** No users yet → use lean-startup. Signups still in the tens — AARRR rates need volume. Bottleneck already obvious → fix that first, then return.\n\n## Coaching Novices (Adaptive Front Door)\n\n- **Engine mode:** user has funnel data and wants the bottleneck named → run The Process directly.\n- **Coach mode:** vague situation or unfamiliarity → guide step by step.\n\nIn Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.\n\n1. **One-line what-it-is.** AARRR is a five-stage funnel (Acquisition / Activation / Retention / Referral / Revenue) — the point is to find which stage is broken so you don't waste effort on the wrong one.\n2. **Check fit.** No users / tiny base → redirect. Bottleneck already obvious → fix it first.\n3. **Elicit their real case.** Force the user to define their *specific* Activation event — not \"signed up\" but e.g. *\"completed onboarding and used the core feature once within day 1.\"* > **[WAIT — do not advance until user responds]**\n4. **One stage at a time.** Walk Acquisition → Activation → Retention → Referral → Revenue. Compute conversion rate stage-to-stage if possible. > **[WAIT — do not advance until user responds]**\n5. **Close by naming the bottleneck and the next experiment.** They leave with one stage identified and one specific experiment. > **[WAIT — do not advance until user responds]**\n\n## The Process\n\nRun the **Funnel Audit**: define stages, measure conversions, identify bottleneck.\n\n1. **Define each stage for this specific product.** Generic definitions hide the bottleneck. Activation = the *aha moment*, not signup. Name a measurable event per stage.\n2. **Measure conversion at each stage.** Acquisition→Activation %, Activation→Retention %, Retention→Referral %, Activation/Retention→Revenue %.\n3. **Compare against domain benchmarks.** A stage below benchmark is a bottleneck candidate (B2B SaaS day-30 retention 70%+; consumer freemium paid conversion 1–5% typical, 5–10% strong).\n4. **Identify the load-bearing bottleneck.** Worst conversion *relative to its benchmark* — fixing it produces the largest system-level gain.\n5. **Pre-commit one experiment.** Name the specific change, the metric, and the pre-committed threshold (per lean-startup).\n6. **Re-measure and re-identify.** After the experiment, the bottleneck moves. Repeat — the audit is iterative.\n\n### Output: Funnel Audit\n\n```\n# AARRR Funnel Audit: \nStage definitions: Acquisition | Activation | Retention | Referral | Revenue \nConversions: [Stage | Users | Conv.% from prior | vs. benchmark ↑↓]\nBottleneck: \nExperiment: Change · Metric · Threshold · Window \n```\n\n*→ Method in Action: [Dropbox's Referral Program (2009)](examples/dropbox-referral-program-2009.md) · [Facebook's \"7 Friends in 10 Days\" (2007–2009)](examples/facebook-7-friends-in-10-days-2008.md)*\n\n*→ 2026 lens: [Where AI SaaS Funnels Leak (2023–2026)](examples/ai-saas-funnel-leaks-2023-2026.md) — cheap AI-hype acquisition, but activation (empty-state / first-value) and retention (after novelty decays) are the real leaks.*\n\n## Funnel Packs\n\nStage definitions and benchmarks are domain-specific. **Consumer freemium:** Activation = first core value; paid conversion 1–5% typical, 5–10% strong. **B2B SaaS:** Activation = first non-trivial team use week 1; gross retention 90%+, net 100%+ mid-market. **Marketplaces:** Activation = first transaction; bottleneck = liquidity-thin side. **E-commerce:** Activation = first order; Retention = second order within 90 days.\n\nContribution: add a pack for your domain — one file with (a) stage definitions, (b) benchmarks, (c) typical bottleneck pattern, (d) canonical experiments.\n\n## Applying It Well\n\n- **Activation is the most often-sloppily-defined stage.** \"Signed up\" is not Activation. Specify the aha moment event.\n- **The bottleneck is relative to benchmark** — not the stage with the lowest absolute number.\n- **Fixing the bottleneck moves the bottleneck.** Re-identify after each experiment.\n- **Each team optimizes the stage they own** — dangerous if the bottleneck is elsewhere. The audit forces cross-team prioritization.\n- **Domain benchmarks matter more than absolute numbers.** 3% paid conversion = great freemium, terrible enterprise SaaS.\n\n*→ Primary sources: [references/sources.md](references/sources.md)*\n\n## Common Rationalizations\n\n**[D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.**\n\n| Fake move | Reality |\n|---|---|\n| [D] **Optimizing Acquisition while Activation is broken** | Most common failure. More users into a broken funnel = more wasted CAC. Fix the bottleneck first. |\n| [D] **\"Activation\" = \"signed up\"** | Signup is end of Acquisition. Activation = first time user experiences the value. Sloppy definition hides the real funnel. |\n| [D] **Treating low absolute conversion as the bottleneck** | Acquisition always has the lowest absolute count. What matters: each stage *relative to its domain benchmark*. |\n| [D] **Adding paid ads when Retention is weak** | A leaky bucket doesn't fill faster when bigger. Fix Retention before scaling Acquisition. |\n| [D] **One-shot audit, never re-run** | The bottleneck moves once fixed. Running the audit once and never again misses the next bottleneck. |\n| *→ Add [O] entries here after each real use — paste the actual failure pattern* | *What went wrong and why* |\n\n## Red Flags\n\n- Activation is defined as \"signed up\"\n- Dashboards show absolute counts but not stage-to-stage conversion rates\n- No domain benchmarks referenced\n- Resource allocation matches which stage is easy to instrument, not which is the bottleneck\n- The funnel audit has been run once and never repeated\n- Teams don't share a bottleneck consensus\n- Acquisition spend scaled while Retention is below benchmark\n\n## Verification\n\n- [ ] Each stage has a product-specific measurable definition (Activation = aha moment, not signup)\n- [ ] Conversion rates between stages measured, not just absolute counts\n- [ ] Each stage compared against a domain benchmark\n- [ ] Load-bearing bottleneck named (worst conversion relative to benchmark)\n- [ ] Experiment pre-committed with metric + threshold + time window\n- [ ] Re-audit scheduled (bottleneck will move)\n- [ ] All teams agree on the bottleneck\n\n---\n\n*Part of **deciqAI Knowledge Skills** — 227 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. **See it run → https://www.deciqai.com/c/aarrr-pirate-metrics** · ⭐ Star the repo → https://github.com/deciqAI/knowledge-skills · Contributions welcome.*\n\n*Agents: latest version & machine-readable metadata → https://www.deciqai.com/s/aarrr-pirate-metrics.json*\n\nFile v1.0.6:_meta.json\n\n{\n \"ownerId\": \"kn754b8sk22s8c6gjxt02bftbn88q7ye\",\n \"slug\": \"aarrr-pirate-metrics\",\n \"version\": \"1.0.6\",\n \"publishedAt\": 1784224265759\n}\n\nFile v1.0.6:references/sources.md\n\n# Sources — aarrr-pirate-metrics\n\n> *Primary sources for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\n- **McClure, Dave.** *Startup Metrics for Pirates* (AARRR). The framework was first presented in 2007 (Ignite Seattle); the widely-circulated \"long version\" slide deck linked here is the Startonomics SF edition (October 2008). SlideShare: https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version **Canonical primary source for AARRR.**\n- **Goldratt, Eliyahu M.** *The Goal: A Process of Ongoing Improvement*. North River Press, 1984; rev. 3rd ed. 2014. **The Theory of Constraints** that underpins the bottleneck logic of AARRR.\n- **Ellis, Sean & Brown, Morgan.** *Hacking Growth: How Today's Fastest-Growing Companies Drive Breakout Success*. Crown Business, 2017. Growth-marketing case studies including the Dropbox referral case and Facebook's \"7 friends in 10 days\" activation metric.\n- **Palihapitiya, Chamath.** \"How We Put Facebook on the Path to 1 Billion Users\" (public talk, c. 2012; widely circulated 2013). First-person account of Facebook's growth team and the \"7 friends in 10 days\" activation metric. Note: the \"7 friends in 10 days\" figure is a widely-repeated growth-team anecdote rather than a formally published metric; treat as illustrative.\n- **Dropbox Inc., Form S-1 (SEC, 2018)** — primary-source user growth milestones: https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm\n- **Sequoia Capital case studies on Dropbox** — primary-source venture history: https://www.sequoiacap.com/\n- **OpenAI.** \"Introducing ChatGPT\" (November 30, 2022): https://openai.com/blog/chatgpt — dates the acquisition-side inflection for the 2023–2026 AI SaaS funnel example (AI-hype traffic makes acquisition cheap).\n- **Sequoia Capital / Cahn, David.** \"AI's $600B Question\" (2024): https://www.sequoiacap.com/article/ais-600b-question/ — widely-cited analysis of the gap between AI infrastructure/capex build-out and durable end-application revenue and retention; context for why activation/retention (not acquisition) is the AI-SaaS bottleneck.\n- The popular framing \"growth hacking\" is **not** cited as a source — by this skill's own rule, an aphorism is not evidence. The framework is precisely \"instrument the funnel, find the bottleneck stage relative to benchmark, concentrate effort there, re-measure.\"\n\nFile v1.0.6:examples/ai-saas-funnel-leaks-2023-2026.md\n\n# Method in Action: Where AI SaaS Funnels Leak (2023–2026)\n\n> *Example for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\nA pattern-level worked example — not a single company, but the recurring funnel shape observed across the wave of AI-native SaaS products that launched after ChatGPT's release in November 2022. The generative-AI boom pushed a distinctive failure mode to the surface: acquisition became historically *easy* while activation and retention became the real leaks. This example applies the Funnel Audit to that pattern. Where a specific number would be fragile, it is qualified or omitted.\n\nThe context in one line (drawn from the 2023–2025 record; the pattern appears to persist but recent periods should be re-checked against current data): after the November 2022 ChatGPT launch, curiosity and press drove enormous top-of-funnel traffic to anything labeled \"AI,\" and by early 2025 the AI-native SaaS category was crowded with near-identical wrapper products competing on the same underlying foundation models. In that environment, *showing up* is cheap and *staying* is hard — the inverse of the classic pre-AI SaaS funnel, where distribution was the scarce resource.\n\nWalk the Funnel Audit (matching the SKILL.md Process steps):\n\n**Step 1 — Define each stage for this specific product.** For a generic AI SaaS assistant:\n- **Acquisition** = visitor lands on the site and creates an account (often driven by AI-hype PR, launch-day virality, or a viral demo).\n- **Activation** = the user reaches first real value — i.e., they get past the **empty state** and complete one genuinely useful task with the AI (a good draft, a correct answer on *their own* data, a workflow actually finished), not merely \"typed one prompt.\"\n- **Retention** = the user returns and uses the product in a later week for real work, *after the novelty of trying an AI toy has worn off*.\n- **Referral** = the user shares an output or invites a colleague.\n- **Revenue** = the user converts from free trial to a paid subscription.\n\nThe single most common definitional error here is calling \"signed up and sent one message\" *Activation*. That is the end of Acquisition. AI products are unusually vulnerable to this because a first prompt is trivially easy, so the vanity signal looks great while no durable value has been delivered.\n\n**Step 2 — Measure conversion at each stage.** The characteristic AI-SaaS reading:\n- Acquisition→Activation: **leaky.** Many curiosity-driven signups hit a blank chat box or empty canvas, don't know what to type, get a generic or hallucinated first result, and never reach first value. This is the **empty-state / first-value** leak.\n- Activation→Retention: **the deepest leak.** Even users who got one impressive result often don't come back, because the initial \"wow\" was novelty, not a solved recurring job. This is the **novelty-decay** leak.\n- Retention→Referral and →Revenue: mostly downstream symptoms — thin because too few users retained.\n\n**Step 3 — Compare against domain benchmarks.** By the consumer-freemium and B2B-SaaS packs in this skill, healthy day-30 retention for sticky SaaS is high (B2B 70%+; strong freemium keeps a meaningful engaged core). The widely-discussed concern across 2023–2025 was that many AI-native apps retained *well below* comparable non-AI SaaS — high novelty-driven signups, weak week-4+ engagement. Treat specific retention percentages as directional; the durable, well-reported fact is the *shape*: acquisition strong, activation and retention weak relative to benchmark.\n\n**Step 4 — Identify the load-bearing bottleneck.** Absolute counts mislead here: Acquisition has the biggest number and looks like a triumph. But relative to benchmark, the worst two stages are **Activation (empty-state → first value)** and **Retention (surviving novelty decay)**. Pouring more launch-day traffic into this funnel — the tempting move when AI PR makes acquisition cheap — just fills a leaky bucket faster. The constraint is not \"get more signups\"; it is \"get signups to first value, then to a recurring job.\"\n\n**Step 5 — Pre-commit one experiment.** Attack Activation first (it gates Retention). Example intervention and pre-committed threshold:\n- *Change:* replace the blank empty state with guided templates / example prompts / a one-click \"do a real task on sample-or-your data\" onboarding, so a new user reaches a genuinely useful first result inside session one.\n- *Metric:* Acquisition→Activation conversion (reached first real value).\n- *Threshold:* pre-commit a target lift (e.g. +X percentage points) before running it, per [lean-startup](../lean-startup/SKILL.md).\n- *Window:* a fixed measurement period (e.g. 4 weeks of new cohorts).\n\n**Step 6 — Re-measure and re-identify.** Once activation improves, the bottleneck **moves to Retention** — and the fix there is different in kind: it requires embedding the product in a recurring job (memory of the user's context, integrations into daily workflow, results that improve with use) so the tool survives after novelty fades. Re-run the audit; do not assume the first fix is the last.\n\n**The non-obvious lesson.** In the pre-AI SaaS era the scarce resource was distribution, so teams instinctively optimized Acquisition. The post-2022 AI wave inverted this: foundation models and hype made Acquisition abundant and near-commoditized (many products wrap the same models), which pushed the true constraint downstream to **Activation and Retention**. A team that reads its own soaring signup chart as success is optimizing the stage that isn't broken. The audit's discipline — *worst conversion relative to benchmark, not the lowest absolute number* — is exactly what stops that mistake. It also reframes the \"AI moat\" question: durable retention (a solved recurring job, accumulated user context, workflow integration) is the moat, not the model, because the model is the one thing every competitor also has.\n\n*Sources: McClure, Dave. \"Startup Metrics for Pirates\" — the AARRR framework, first presented in 2007 (Ignite Seattle); the widely-circulated \"long version\" slide deck cited here is the Startonomics SF (October 2008) edition: https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version . Goldratt, Eliyahu M. \"The Goal\" (North River Press, 1984) — Theory of Constraints / bottleneck logic. OpenAI, \"Introducing ChatGPT\" (Nov 30, 2022): https://openai.com/blog/chatgpt — dates the acquisition-side inflection. Sequoia Capital, \"Generative AI: A Creative New World\" (2022) and \"AI's $600B Question\" (David Cahn, 2024): https://www.sequoiacap.com/article/ais-600b-question/ — widely-cited discussion of the AI capex build-out and the gap between AI infrastructure spend and durable end-application revenue/retention. Retention-shape claims are stated qualitatively; specific per-app retention figures are omitted as fragile.*\n\nFile v1.0.6:examples/dropbox-referral-program-2009.md\n\n# Method in Action: Dropbox's Referral Program (2009)\n\n> *Example for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\nA worked example. Not a marketing legend — primary-source documented in Drew Houston's interviews and growth-marketing case studies.\n\nBy early 2009, **Dropbox** had achieved PMF (see [pmf-crossing-the-chasm](../pmf-crossing-the-chasm/SKILL.md)) and had ~100,000 users from the 2007 video MVP + 2008 launch. The team ran an AARRR audit on the funnel:\n\n- **Acquisition**: Paid Google AdWords campaigns were running, with **customer acquisition cost (CAC) ~$200–$300 per user** — but Dropbox's price point was $99/year ($9.99/month), making **AdWords economically broken** as the dominant channel.\n- **Activation**: Strong. Once users installed the desktop client and put a file in the Dropbox folder, the experience converted to \"aha\" cleanly (the *file appeared on another device* moment, the core delight from the 2007 video MVP).\n- **Retention**: Strong. Users who activated tended to stay; Dropbox solved a real persistent need.\n- **Referral**: Untapped. Users loved the product and told friends informally, but there was no built-in mechanism.\n- **Revenue**: Decent at the free-to-paid conversion typical for freemium (a few percent).\n\nThe audit identified **Acquisition** as the load-bearing bottleneck — *not because Acquisition was bad in absolute terms, but because the acquisition channel that was working economically did not exist*. Spending more on AdWords would not solve it; only finding a sub-CAC channel would.\n\n**The intervention (September 2008–April 2009):** Dropbox launched a **referral program**. The mechanic: **give 500 MB free storage to both the referrer and the referee** when an invited friend signed up and installed Dropbox. The reward was native to the product (storage), aligned with the product's value (more space), and free for Dropbox to give away (storage was Dropbox's marginal cost, not a cash outlay).\n\nThe result, documented by Drew Houston and reproduced in Sean Ellis & Morgan Brown's *Hacking Growth*: **signups grew 60% as a result of the referral program**, with **35% of daily signups coming through referrals by late 2009**. By April 2010, Dropbox had grown to **4 million users**, then **25 million** by 2011. The referral program effectively **converted a Retention strength into Acquisition output** — Goldratt's \"exploit the constraint\" move applied to a customer funnel.\n\nWalk the audit on Dropbox 2009:\n\n- **Stage definitions (Step 1):** Acquisition = ad click → landing page → signup; Activation = client installed + first file synced; Retention = active in day 7; Referral = invite sent → friend installs + syncs file; Revenue = upgrade from free to paid tier.\n- **Conversions (Step 2):** Acquisition CAC vs LTV economically broken; Activation, Retention, Revenue all benchmarks-fine; Referral nominal.\n- **Bottleneck (Step 4):** Acquisition channel economics — *specifically*, the cost-effective channel did not exist.\n- **Experiment (Step 5):** Build referral program; pre-committed threshold = signup growth ≥ 30% from referral source within 6 months.\n- **Result:** 60% growth; threshold cleared by ~2×. *Persevere*; double down on the channel; iterate on referral mechanics.\n\n**The non-obvious lesson:** Dropbox's \"referral was the bottleneck\" framing would have been wrong. The bottleneck was Acquisition; the *intervention* was building a referral mechanism *because* Retention was strong enough that activated users would actually send referrals. **The referral program was the solution to the Acquisition bottleneck**, not a Referral-stage improvement in isolation.\n\n**Sources:** McClure, Dave. ***Startup Metrics for Pirates*** (slide deck, Startonomics SF, August 2007). Original SlideShare: https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version . **Canonical primary source for AARRR**. Goldratt, Eliyahu M. ***The Goal: A Process of Ongoing Improvement***. North River Press, 1984; rev. 3rd ed. 2014. **Theory of Constraints**, the bottleneck logic underpinning AARRR. Ellis, Sean & Brown, Morgan. ***Hacking Growth: How Today's Fastest-Growing Companies Drive Breakout Success***. Crown Business, 2017, ch. 5 — primary growth-marketing case account of the Dropbox referral program. Houston, Drew. Drew Houston's interviews on growth, including talks at Stanford and the Sequoia Capital growth case study at: https://www.sequoiacap.com/ Dropbox SEC S-1 filing (2018) provided official user-growth milestones: https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm\n\nFile v1.0.6:examples/facebook-7-friends-in-10-days-2008.md\n\n# Method in Action: Facebook's \"7 Friends in 10 Days\" Activation Metric (2007–2009)\n\n> *Example for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\nA worked example from a social network — a different domain and a different bottleneck stage than the Dropbox case. Documented in Chamath Palihapitiya's public 2013 talk on Facebook's growth team and in Sean Ellis & Morgan Brown's *Hacking Growth* (2017).\n\nIn 2007–2008, **Facebook's** growth flattened at roughly 90 million monthly users. Inside the company, some believed the product might simply be capped near 100 million — that social networks saturate. Instead of accepting the ceiling, Facebook formed a dedicated Growth team under Chamath Palihapitiya and ran what was, in effect, a funnel audit:\n\n- **Acquisition**: Healthy. Millions of people were still landing on Facebook and signing up every month. Registrations were not the problem.\n- **Activation**: Broken — and hidden by the sloppy definition. \"Signed up\" looked fine; but a large fraction of new registrants arrived to an empty feed, saw no friends, and never returned. Signup was the end of Acquisition, not Activation.\n- **Retention**: Weak *downstream of* Activation. Users who never connected to their real-world friends had nothing to come back for; users who did connect retained strongly.\n- **Referral**: Structurally strong — a social network's core loop is inviting and finding friends — but only activated users exercised it.\n- **Revenue**: Advertising-driven; entirely a function of retained, engaged users. Not the constraint.\n\nThe Growth team's decisive move was **Step 1 of the audit: define Activation as a measurable, product-specific event.** By correlating early behavior with long-term retention, they found the aha moment: **a new user who reached 7 friends within 10 days of signup retained; a user who didn't, churned.** \"7 friends in 10 days\" became the team's north-star activation metric — not a vanity registration count, but the event at which a new user first experienced the product's actual value (a feed full of people they knew).\n\n**The intervention:** With Activation named as the load-bearing bottleneck, the Growth team re-pointed the company's effort at one number. Onboarding was rebuilt around friend-finding — contact importers, the \"People You May Know\" recommendation engine, prompts that pushed every new user toward their first connections — and internationalization (a crowdsourced translation platform) removed the language barrier that kept non-English users from finding their friends at all. Every experiment was judged by whether it moved new users toward the 7-in-10 threshold, then re-measured — the audit run as a loop, not a one-shot.\n\n**The result:** Growth resumed through the supposed ceiling. Facebook passed 500 million users in 2010 and 1 billion monthly users in October 2012 — the trajectory Palihapitiya's talk describes as the direct product of instrumenting the funnel and concentrating on the activation constraint rather than buying more traffic into a leaky bucket.\n\n**The non-obvious lesson:** Facebook's bottleneck was invisible as long as Activation meant \"registered.\" The single highest-leverage act was definitional — replacing a generic stage label with a measured aha-moment event. Only then did the funnel show where growth was actually dying, and only then could every team optimize the same constraint instead of the stage each happened to own.\n\nThe mapped steps:\n1. Define stages product-specifically: Activation = 7 friends within 10 days, not signup\n2. Measure conversions: registrations healthy; signup→activated conversion the weak link; retention strong conditional on activation\n3. Benchmark: activated users retained at healthy social-network rates; non-activated users churned — the gap localized the bottleneck\n4. Identify the load-bearing bottleneck: Activation, not Acquisition\n5. Pre-commit experiments against the single metric: onboarding friend-finders, People You May Know, translations — each judged by movement toward 7-in-10\n6. Re-measure and repeat: continuous experiment loop; growth resumed to 1 billion users by 2012\n\nPrimary source: Palihapitiya, Chamath. \"How We Put Facebook on the Path to a Billion Users\" (public talk, 2013). Case account: Ellis, Sean & Brown, Morgan. *Hacking Growth: How Today's Fastest-Growing Companies Drive Breakout Success*. Crown Business, 2017.\n\nFile v1.0.6:skill-card.md\n\n## Description:\n\nAARRR (Pirate Metrics) helps agents diagnose growth funnel bottlenecks by defining acquisition, activation, retention, referral, and revenue stages, comparing conversions to benchmarks, and proposing the next experiment.\n\nThis skill is ready for commercial/non-commercial use.\n\n## Publisher:\n\n[deciqai](https://clawhub.ai/user/deciqai)\n\n### License/Terms of Use:\n\nMIT-0\n\n## Use Case:\n\nEmployees, external teams, founders, product leaders, growth teams, and developers use this skill to audit a product growth funnel, align teams on the load-bearing bottleneck, and choose a measurable experiment. It is useful when growth is stalling, teams disagree about the cause, or a shared instrumentation framework is needed.\n\n### Deployment Geography for Use:\n\nGlobal\n\n## Known Risks and Mitigations:\n\nRisk: Users may rely on stale or mismatched funnel benchmarks when interpreting acquisition, activation, retention, referral, or revenue performance.\n\nMitigation: Verify benchmarks against current domain data and treat skill-provided benchmark comparisons as advisory before making business decisions.\n\nRisk: Funnel diagnosis can involve sensitive business metrics such as revenue, retention, acquisition cost, or cohort performance.\n\nMitigation: Use aggregated or redacted metrics unless the user intends the assistant to analyze confidential business data.\n\n## Reference(s):\n\n- [ClawHub skill page](https://clawhub.ai/deciqai/skills/aarrr-pirate-metrics)\n- [deciqAI AARRR skill page](https://www.deciqai.com/c/aarrr-pirate-metrics)\n- [Machine-readable skill metadata](https://www.deciqai.com/s/aarrr-pirate-metrics.json)\n- [Sources - aarrr-pirate-metrics](references/sources.md)\n- [Startup Metrics for Pirates](https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version)\n- [Dropbox SEC Form S-1](https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm)\n- [Introducing ChatGPT](https://openai.com/blog/chatgpt)\n- [AI's $600B Question](https://www.sequoiacap.com/article/ais-600b-question/)\n\n## Skill Output:\n\n**Output Type(s):** [text, markdown, guidance]\n\n**Output Format:** [Markdown funnel audit with staged questions, stage definitions, conversion comparisons, bottleneck identification, and experiment guidance.]\n\n**Output Parameters:** [1D]\n\n**Other Properties Related to Output:** [May proceed one question at a time in coach mode; produces advisory analysis only and does not run code or request tool access.]\n\n## Skill Version(s):\n\n1.0.6 (source: server release metadata)\n\n## Ethical Considerations:\n\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.\n\nArchive v1.0.5: 7 files, 15590 bytes\n\nFiles: examples/ai-saas-funnel-leaks-2023-2026.md (6925b), examples/dropbox-referral-program-2009.md (4627b), examples/facebook-7-friends-in-10-days-2008.md (4428b), references/sources.md (2384b), skill-card.md (2634b), SKILL.md (9042b), _meta.json (139b)\n\nFile v1.0.5:SKILL.md\n\n---\nname: aarrr-pirate-metrics\ndescription: \"Activate when: user says 'our growth is stalling and we don't know why,' 'product and marketing are arguing about whose fault it is,' 'where is our funnel breaking,' 'pirate metrics,' or 'acquisition vs activation vs retention.' Also activate for: designing end-to-end metrics for a new product, building a shared instrumentation framework across teams.\n Do NOT activate when: the product has no users yet (use lean-startup instead), or the bottleneck is already obvious and obvious to fix.\"\n---\n\n# AARRR (Pirate Metrics)\n\n## Overview\n\nA startup's growth is a **sequential funnel** — each stage gates the next. Great Acquisition is worthless if Activation is broken; great Activation is worthless if Retention is zero. Optimizing the wrong stage produces work that looks like progress while the bottleneck stays.\n\nThe **AARRR framework** (Dave McClure, *Startup Metrics for Pirates*, 2007) names five stages: Acquisition (do they show up?), Activation (good first experience?), Retention (do they come back?), Referral (do they tell others?), Revenue (do they pay?). The bottleneck stage governs total growth — improving any other stage produces no system-level gain (Goldratt, *The Goal*, 1984).\n\n**Compose with:** first-principles to identify your specific Activation event; pmf-crossing-the-chasm to recognize when Retention will always be the bottleneck pre-PMF; probabilistic-thinking to set base rates per stage.\n\n## When to Use\n\n- Growth is **stalling** and the cause is not obvious — multiple teams have plausible explanations\n- Product, marketing, and sales are **arguing about whose problem it is**\n- You need a **shared instrumentation framework** across teams\n- Designing **end-to-end metrics for a new product**\n- Someone says: *\"AARRR,\" \"pirate metrics,\" \"growth funnel,\" \"where is our funnel breaking?\"*\n- An **AI-native product has strong signups but weak activation/retention** — cheap AI-hype acquisition masks empty-state and post-novelty leaks; or you're facing AI-native competition on a commoditized model and need to find where your funnel actually loses users\n\n**When NOT to use:** No users yet → use lean-startup. Signups still in the tens — AARRR rates need volume. Bottleneck already obvious → fix that first, then return.\n\n## Coaching Novices (Adaptive Front Door)\n\n- **Engine mode:** user has funnel data and wants the bottleneck named → run The Process directly.\n- **Coach mode:** vague situation or unfamiliarity → guide step by step.\n\nIn Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.\n\n1. **One-line what-it-is.** AARRR is a five-stage funnel (Acquisition / Activation / Retention / Referral / Revenue) — the point is to find which stage is broken so you don't waste effort on the wrong one.\n2. **Check fit.** No users / tiny base → redirect. Bottleneck already obvious → fix it first.\n3. **Elicit their real case.** Force the user to define their *specific* Activation event — not \"signed up\" but e.g. *\"completed onboarding and used the core feature once within day 1.\"* > **[WAIT — do not advance until user responds]**\n4. **One stage at a time.** Walk Acquisition → Activation → Retention → Referral → Revenue. Compute conversion rate stage-to-stage if possible. > **[WAIT — do not advance until user responds]**\n5. **Close by naming the bottleneck and the next experiment.** They leave with one stage identified and one specific experiment. > **[WAIT — do not advance until user responds]**\n\n## The Process\n\nRun the **Funnel Audit**: define stages, measure conversions, identify bottleneck.\n\n1. **Define each stage for this specific product.** Generic definitions hide the bottleneck. Activation = the *aha moment*, not signup. Name a measurable event per stage.\n2. **Measure conversion at each stage.** Acquisition→Activation %, Activation→Retention %, Retention→Referral %, Activation/Retention→Revenue %.\n3. **Compare against domain benchmarks.** A stage below benchmark is a bottleneck candidate (B2B SaaS day-30 retention 70%+; consumer freemium paid conversion 1–5% typical, 5–10% strong).\n4. **Identify the load-bearing bottleneck.** Worst conversion *relative to its benchmark* — fixing it produces the largest system-level gain.\n5. **Pre-commit one experiment.** Name the specific change, the metric, and the pre-committed threshold (per lean-startup).\n6. **Re-measure and re-identify.** After the experiment, the bottleneck moves. Repeat — the audit is iterative.\n\n### Output: Funnel Audit\n\n```\n# AARRR Funnel Audit: \nStage definitions: Acquisition | Activation | Retention | Referral | Revenue \nConversions: [Stage | Users | Conv.% from prior | vs. benchmark ↑↓]\nBottleneck: \nExperiment: Change · Metric · Threshold · Window \n```\n\n*→ Method in Action: [Dropbox's Referral Program (2009)](examples/dropbox-referral-program-2009.md) · [Facebook's \"7 Friends in 10 Days\" (2007–2009)](examples/facebook-7-friends-in-10-days-2008.md)*\n\n*→ 2026 lens: [Where AI SaaS Funnels Leak (2023–2026)](examples/ai-saas-funnel-leaks-2023-2026.md) — cheap AI-hype acquisition, but activation (empty-state / first-value) and retention (after novelty decays) are the real leaks.*\n\n## Funnel Packs\n\nStage definitions and benchmarks are domain-specific. **Consumer freemium:** Activation = first core value; paid conversion 1–5% typical, 5–10% strong. **B2B SaaS:** Activation = first non-trivial team use week 1; gross retention 90%+, net 100%+ mid-market. **Marketplaces:** Activation = first transaction; bottleneck = liquidity-thin side. **E-commerce:** Activation = first order; Retention = second order within 90 days.\n\nContribution: add a pack for your domain — one file with (a) stage definitions, (b) benchmarks, (c) typical bottleneck pattern, (d) canonical experiments.\n\n## Applying It Well\n\n- **Activation is the most often-sloppily-defined stage.** \"Signed up\" is not Activation. Specify the aha moment event.\n- **The bottleneck is relative to benchmark** — not the stage with the lowest absolute number.\n- **Fixing the bottleneck moves the bottleneck.** Re-identify after each experiment.\n- **Each team optimizes the stage they own** — dangerous if the bottleneck is elsewhere. The audit forces cross-team prioritization.\n- **Domain benchmarks matter more than absolute numbers.** 3% paid conversion = great freemium, terrible enterprise SaaS.\n\n*→ Primary sources: [references/sources.md](references/sources.md)*\n\n## Common Rationalizations\n\n**[D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.**\n\n| Fake move | Reality |\n|---|---|\n| [D] **Optimizing Acquisition while Activation is broken** | Most common failure. More users into a broken funnel = more wasted CAC. Fix the bottleneck first. |\n| [D] **\"Activation\" = \"signed up\"** | Signup is end of Acquisition. Activation = first time user experiences the value. Sloppy definition hides the real funnel. |\n| [D] **Treating low absolute conversion as the bottleneck** | Acquisition always has the lowest absolute count. What matters: each stage *relative to its domain benchmark*. |\n| [D] **Adding paid ads when Retention is weak** | A leaky bucket doesn't fill faster when bigger. Fix Retention before scaling Acquisition. |\n| [D] **One-shot audit, never re-run** | The bottleneck moves once fixed. Running the audit once and never again misses the next bottleneck. |\n| *→ Add [O] entries here after each real use — paste the actual failure pattern* | *What went wrong and why* |\n\n## Red Flags\n\n- Activation is defined as \"signed up\"\n- Dashboards show absolute counts but not stage-to-stage conversion rates\n- No domain benchmarks referenced\n- Resource allocation matches which stage is easy to instrument, not which is the bottleneck\n- The funnel audit has been run once and never repeated\n- Teams don't share a bottleneck consensus\n- Acquisition spend scaled while Retention is below benchmark\n\n## Verification\n\n- [ ] Each stage has a product-specific measurable definition (Activation = aha moment, not signup)\n- [ ] Conversion rates between stages measured, not just absolute counts\n- [ ] Each stage compared against a domain benchmark\n- [ ] Load-bearing bottleneck named (worst conversion relative to benchmark)\n- [ ] Experiment pre-committed with metric + threshold + time window\n- [ ] Re-audit scheduled (bottleneck will move)\n- [ ] All teams agree on the bottleneck\n\n---\n\n*Part of **deciqAI Knowledge Skills** — 189 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. **See it run → https://www.deciqai.com/c/aarrr-pirate-metrics** · ⭐ Star the repo → https://github.com/deciqAI/knowledge-skills · Contributions welcome.*\n\nFile v1.0.5:_meta.json\n\n{\n \"ownerId\": \"kn754b8sk22s8c6gjxt02bftbn88q7ye\",\n \"slug\": \"aarrr-pirate-metrics\",\n \"version\": \"1.0.5\",\n \"publishedAt\": 1783595691577\n}\n\nFile v1.0.5:references/sources.md\n\n# Sources — aarrr-pirate-metrics\n\n> *Primary sources for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\n- **McClure, Dave.** *Startup Metrics for Pirates* (AARRR). The framework was first presented in 2007 (Ignite Seattle); the widely-circulated \"long version\" slide deck linked here is the Startonomics SF edition (October 2008). SlideShare: https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version **Canonical primary source for AARRR.**\n- **Goldratt, Eliyahu M.** *The Goal: A Process of Ongoing Improvement*. North River Press, 1984; rev. 3rd ed. 2014. **The Theory of Constraints** that underpins the bottleneck logic of AARRR.\n- **Ellis, Sean & Brown, Morgan.** *Hacking Growth: How Today's Fastest-Growing Companies Drive Breakout Success*. Crown Business, 2017. Growth-marketing case studies including the Dropbox referral case and Facebook's \"7 friends in 10 days\" activation metric.\n- **Palihapitiya, Chamath.** \"How We Put Facebook on the Path to 1 Billion Users\" (public talk, c. 2012; widely circulated 2013). First-person account of Facebook's growth team and the \"7 friends in 10 days\" activation metric. Note: the \"7 friends in 10 days\" figure is a widely-repeated growth-team anecdote rather than a formally published metric; treat as illustrative.\n- **Dropbox Inc., Form S-1 (SEC, 2018)** — primary-source user growth milestones: https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm\n- **Sequoia Capital case studies on Dropbox** — primary-source venture history: https://www.sequoiacap.com/\n- **OpenAI.** \"Introducing ChatGPT\" (November 30, 2022): https://openai.com/blog/chatgpt — dates the acquisition-side inflection for the 2023–2026 AI SaaS funnel example (AI-hype traffic makes acquisition cheap).\n- **Sequoia Capital / Cahn, David.** \"AI's $600B Question\" (2024): https://www.sequoiacap.com/article/ais-600b-question/ — widely-cited analysis of the gap between AI infrastructure/capex build-out and durable end-application revenue and retention; context for why activation/retention (not acquisition) is the AI-SaaS bottleneck.\n- The popular framing \"growth hacking\" is **not** cited as a source — by this skill's own rule, an aphorism is not evidence. The framework is precisely \"instrument the funnel, find the bottleneck stage relative to benchmark, concentrate effort there, re-measure.\"\n\nFile v1.0.5:examples/ai-saas-funnel-leaks-2023-2026.md\n\n# Method in Action: Where AI SaaS Funnels Leak (2023–2026)\n\n> *Example for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\nA pattern-level worked example — not a single company, but the recurring funnel shape observed across the wave of AI-native SaaS products that launched after ChatGPT's release in November 2022. The generative-AI boom pushed a distinctive failure mode to the surface: acquisition became historically *easy* while activation and retention became the real leaks. This example applies the Funnel Audit to that pattern. Where a specific number would be fragile, it is qualified or omitted.\n\nThe context in one line (drawn from the 2023–2025 record; the pattern appears to persist but recent periods should be re-checked against current data): after the November 2022 ChatGPT launch, curiosity and press drove enormous top-of-funnel traffic to anything labeled \"AI,\" and by early 2025 the AI-native SaaS category was crowded with near-identical wrapper products competing on the same underlying foundation models. In that environment, *showing up* is cheap and *staying* is hard — the inverse of the classic pre-AI SaaS funnel, where distribution was the scarce resource.\n\nWalk the Funnel Audit (matching the SKILL.md Process steps):\n\n**Step 1 — Define each stage for this specific product.** For a generic AI SaaS assistant:\n- **Acquisition** = visitor lands on the site and creates an account (often driven by AI-hype PR, launch-day virality, or a viral demo).\n- **Activation** = the user reaches first real value — i.e., they get past the **empty state** and complete one genuinely useful task with the AI (a good draft, a correct answer on *their own* data, a workflow actually finished), not merely \"typed one prompt.\"\n- **Retention** = the user returns and uses the product in a later week for real work, *after the novelty of trying an AI toy has worn off*.\n- **Referral** = the user shares an output or invites a colleague.\n- **Revenue** = the user converts from free trial to a paid subscription.\n\nThe single most common definitional error here is calling \"signed up and sent one message\" *Activation*. That is the end of Acquisition. AI products are unusually vulnerable to this because a first prompt is trivially easy, so the vanity signal looks great while no durable value has been delivered.\n\n**Step 2 — Measure conversion at each stage.** The characteristic AI-SaaS reading:\n- Acquisition→Activation: **leaky.** Many curiosity-driven signups hit a blank chat box or empty canvas, don't know what to type, get a generic or hallucinated first result, and never reach first value. This is the **empty-state / first-value** leak.\n- Activation→Retention: **the deepest leak.** Even users who got one impressive result often don't come back, because the initial \"wow\" was novelty, not a solved recurring job. This is the **novelty-decay** leak.\n- Retention→Referral and →Revenue: mostly downstream symptoms — thin because too few users retained.\n\n**Step 3 — Compare against domain benchmarks.** By the consumer-freemium and B2B-SaaS packs in this skill, healthy day-30 retention for sticky SaaS is high (B2B 70%+; strong freemium keeps a meaningful engaged core). The widely-discussed concern across 2023–2025 was that many AI-native apps retained *well below* comparable non-AI SaaS — high novelty-driven signups, weak week-4+ engagement. Treat specific retention percentages as directional; the durable, well-reported fact is the *shape*: acquisition strong, activation and retention weak relative to benchmark.\n\n**Step 4 — Identify the load-bearing bottleneck.** Absolute counts mislead here: Acquisition has the biggest number and looks like a triumph. But relative to benchmark, the worst two stages are **Activation (empty-state → first value)** and **Retention (surviving novelty decay)**. Pouring more launch-day traffic into this funnel — the tempting move when AI PR makes acquisition cheap — just fills a leaky bucket faster. The constraint is not \"get more signups\"; it is \"get signups to first value, then to a recurring job.\"\n\n**Step 5 — Pre-commit one experiment.** Attack Activation first (it gates Retention). Example intervention and pre-committed threshold:\n- *Change:* replace the blank empty state with guided templates / example prompts / a one-click \"do a real task on sample-or-your data\" onboarding, so a new user reaches a genuinely useful first result inside session one.\n- *Metric:* Acquisition→Activation conversion (reached first real value).\n- *Threshold:* pre-commit a target lift (e.g. +X percentage points) before running it, per [lean-startup](../lean-startup/SKILL.md).\n- *Window:* a fixed measurement period (e.g. 4 weeks of new cohorts).\n\n**Step 6 — Re-measure and re-identify.** Once activation improves, the bottleneck **moves to Retention** — and the fix there is different in kind: it requires embedding the product in a recurring job (memory of the user's context, integrations into daily workflow, results that improve with use) so the tool survives after novelty fades. Re-run the audit; do not assume the first fix is the last.\n\n**The non-obvious lesson.** In the pre-AI SaaS era the scarce resource was distribution, so teams instinctively optimized Acquisition. The post-2022 AI wave inverted this: foundation models and hype made Acquisition abundant and near-commoditized (many products wrap the same models), which pushed the true constraint downstream to **Activation and Retention**. A team that reads its own soaring signup chart as success is optimizing the stage that isn't broken. The audit's discipline — *worst conversion relative to benchmark, not the lowest absolute number* — is exactly what stops that mistake. It also reframes the \"AI moat\" question: durable retention (a solved recurring job, accumulated user context, workflow integration) is the moat, not the model, because the model is the one thing every competitor also has.\n\n*Sources: McClure, Dave. \"Startup Metrics for Pirates\" — the AARRR framework, first presented in 2007 (Ignite Seattle); the widely-circulated \"long version\" slide deck cited here is the Startonomics SF (October 2008) edition: https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version . Goldratt, Eliyahu M. \"The Goal\" (North River Press, 1984) — Theory of Constraints / bottleneck logic. OpenAI, \"Introducing ChatGPT\" (Nov 30, 2022): https://openai.com/blog/chatgpt — dates the acquisition-side inflection. Sequoia Capital, \"Generative AI: A Creative New World\" (2022) and \"AI's $600B Question\" (David Cahn, 2024): https://www.sequoiacap.com/article/ais-600b-question/ — widely-cited discussion of the AI capex build-out and the gap between AI infrastructure spend and durable end-application revenue/retention. Retention-shape claims are stated qualitatively; specific per-app retention figures are omitted as fragile.*\n\nFile v1.0.5:examples/dropbox-referral-program-2009.md\n\n# Method in Action: Dropbox's Referral Program (2009)\n\n> *Example for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\nA worked example. Not a marketing legend — primary-source documented in Drew Houston's interviews and growth-marketing case studies.\n\nBy early 2009, **Dropbox** had achieved PMF (see [pmf-crossing-the-chasm](../pmf-crossing-the-chasm/SKILL.md)) and had ~100,000 users from the 2007 video MVP + 2008 launch. The team ran an AARRR audit on the funnel:\n\n- **Acquisition**: Paid Google AdWords campaigns were running, with **customer acquisition cost (CAC) ~$200–$300 per user** — but Dropbox's price point was $99/year ($9.99/month), making **AdWords economically broken** as the dominant channel.\n- **Activation**: Strong. Once users installed the desktop client and put a file in the Dropbox folder, the experience converted to \"aha\" cleanly (the *file appeared on another device* moment, the core delight from the 2007 video MVP).\n- **Retention**: Strong. Users who activated tended to stay; Dropbox solved a real persistent need.\n- **Referral**: Untapped. Users loved the product and told friends informally, but there was no built-in mechanism.\n- **Revenue**: Decent at the free-to-paid conversion typical for freemium (a few percent).\n\nThe audit identified **Acquisition** as the load-bearing bottleneck — *not because Acquisition was bad in absolute terms, but because the acquisition channel that was working economically did not exist*. Spending more on AdWords would not solve it; only finding a sub-CAC channel would.\n\n**The intervention (September 2008–April 2009):** Dropbox launched a **referral program**. The mechanic: **give 500 MB free storage to both the referrer and the referee** when an invited friend signed up and installed Dropbox. The reward was native to the product (storage), aligned with the product's value (more space), and free for Dropbox to give away (storage was Dropbox's marginal cost, not a cash outlay).\n\nThe result, documented by Drew Houston and reproduced in Sean Ellis & Morgan Brown's *Hacking Growth*: **signups grew 60% as a result of the referral program**, with **35% of daily signups coming through referrals by late 2009**. By April 2010, Dropbox had grown to **4 million users**, then **25 million** by 2011. The referral program effectively **converted a Retention strength into Acquisition output** — Goldratt's \"exploit the constraint\" move applied to a customer funnel.\n\nWalk the audit on Dropbox 2009:\n\n- **Stage definitions (Step 1):** Acquisition = ad click → landing page → signup; Activation = client installed + first file synced; Retention = active in day 7; Referral = invite sent → friend installs + syncs file; Revenue = upgrade from free to paid tier.\n- **Conversions (Step 2):** Acquisition CAC vs LTV economically broken; Activation, Retention, Revenue all benchmarks-fine; Referral nominal.\n- **Bottleneck (Step 4):** Acquisition channel economics — *specifically*, the cost-effective channel did not exist.\n- **Experiment (Step 5):** Build referral program; pre-committed threshold = signup growth ≥ 30% from referral source within 6 months.\n- **Result:** 60% growth; threshold cleared by ~2×. *Persevere*; double down on the channel; iterate on referral mechanics.\n\n**The non-obvious lesson:** Dropbox's \"referral was the bottleneck\" framing would have been wrong. The bottleneck was Acquisition; the *intervention* was building a referral mechanism *because* Retention was strong enough that activated users would actually send referrals. **The referral program was the solution to the Acquisition bottleneck**, not a Referral-stage improvement in isolation.\n\n**Sources:** McClure, Dave. ***Startup Metrics for Pirates*** (slide deck, Startonomics SF, August 2007). Original SlideShare: https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version . **Canonical primary source for AARRR**. Goldratt, Eliyahu M. ***The Goal: A Process of Ongoing Improvement***. North River Press, 1984; rev. 3rd ed. 2014. **Theory of Constraints**, the bottleneck logic underpinning AARRR. Ellis, Sean & Brown, Morgan. ***Hacking Growth: How Today's Fastest-Growing Companies Drive Breakout Success***. Crown Business, 2017, ch. 5 — primary growth-marketing case account of the Dropbox referral program. Houston, Drew. Drew Houston's interviews on growth, including talks at Stanford and the Sequoia Capital growth case study at: https://www.sequoiacap.com/ Dropbox SEC S-1 filing (2018) provided official user-growth milestones: https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm\n\nFile v1.0.5:examples/facebook-7-friends-in-10-days-2008.md\n\n# Method in Action: Facebook's \"7 Friends in 10 Days\" Activation Metric (2007–2009)\n\n> *Example for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\nA worked example from a social network — a different domain and a different bottleneck stage than the Dropbox case. Documented in Chamath Palihapitiya's public 2013 talk on Facebook's growth team and in Sean Ellis & Morgan Brown's *Hacking Growth* (2017).\n\nIn 2007–2008, **Facebook's** growth flattened at roughly 90 million monthly users. Inside the company, some believed the product might simply be capped near 100 million — that social networks saturate. Instead of accepting the ceiling, Facebook formed a dedicated Growth team under Chamath Palihapitiya and ran what was, in effect, a funnel audit:\n\n- **Acquisition**: Healthy. Millions of people were still landing on Facebook and signing up every month. Registrations were not the problem.\n- **Activation**: Broken — and hidden by the sloppy definition. \"Signed up\" looked fine; but a large fraction of new registrants arrived to an empty feed, saw no friends, and never returned. Signup was the end of Acquisition, not Activation.\n- **Retention**: Weak *downstream of* Activation. Users who never connected to their real-world friends had nothing to come back for; users who did connect retained strongly.\n- **Referral**: Structurally strong — a social network's core loop is inviting and finding friends — but only activated users exercised it.\n- **Revenue**: Advertising-driven; entirely a function of retained, engaged users. Not the constraint.\n\nThe Growth team's decisive move was **Step 1 of the audit: define Activation as a measurable, product-specific event.** By correlating early behavior with long-term retention, they found the aha moment: **a new user who reached 7 friends within 10 days of signup retained; a user who didn't, churned.** \"7 friends in 10 days\" became the team's north-star activation metric — not a vanity registration count, but the event at which a new user first experienced the product's actual value (a feed full of people they knew).\n\n**The intervention:** With Activation named as the load-bearing bottleneck, the Growth team re-pointed the company's effort at one number. Onboarding was rebuilt around friend-finding — contact importers, the \"People You May Know\" recommendation engine, prompts that pushed every new user toward their first connections — and internationalization (a crowdsourced translation platform) removed the language barrier that kept non-English users from finding their friends at all. Every experiment was judged by whether it moved new users toward the 7-in-10 threshold, then re-measured — the audit run as a loop, not a one-shot.\n\n**The result:** Growth resumed through the supposed ceiling. Facebook passed 500 million users in 2010 and 1 billion monthly users in October 2012 — the trajectory Palihapitiya's talk describes as the direct product of instrumenting the funnel and concentrating on the activation constraint rather than buying more traffic into a leaky bucket.\n\n**The non-obvious lesson:** Facebook's bottleneck was invisible as long as Activation meant \"registered.\" The single highest-leverage act was definitional — replacing a generic stage label with a measured aha-moment event. Only then did the funnel show where growth was actually dying, and only then could every team optimize the same constraint instead of the stage each happened to own.\n\nThe mapped steps:\n1. Define stages product-specifically: Activation = 7 friends within 10 days, not signup\n2. Measure conversions: registrations healthy; signup→activated conversion the weak link; retention strong conditional on activation\n3. Benchmark: activated users retained at healthy social-network rates; non-activated users churned — the gap localized the bottleneck\n4. Identify the load-bearing bottleneck: Activation, not Acquisition\n5. Pre-commit experiments against the single metric: onboarding friend-finders, People You May Know, translations — each judged by movement toward 7-in-10\n6. Re-measure and repeat: continuous experiment loop; growth resumed to 1 billion users by 2012\n\nPrimary source: Palihapitiya, Chamath. \"How We Put Facebook on the Path to a Billion Users\" (public talk, 2013). Case account: Ellis, Sean & Brown, Morgan. *Hacking Growth: How Today's Fastest-Growing Companies Drive Breakout Success*. Crown Business, 2017.\n\nFile v1.0.5:skill-card.md\n\n## Description:
\nGuides agents through an AARRR funnel audit to define acquisition, activation, retention, referral, and revenue stages, identify the load-bearing bottleneck, and choose a measurable growth experiment.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[deciqai](https://clawhub.ai/user/deciqai)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nProduct, growth, marketing, sales, and startup leadership teams use this skill to audit a product funnel, align teams on the stage limiting growth, and define a measurable experiment.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: The funnel audit can produce misleading recommendations when teams provide vague stage definitions, vanity metrics, or benchmark-free conversion data.
\nMitigation: Require product-specific measurable events for each AARRR stage, compare conversion rates against relevant domain benchmarks, and review the proposed bottleneck before acting.
\nRisk: Growth experiments may over-optimize one funnel stage while creating downstream retention, revenue, or customer-experience tradeoffs.
\nMitigation: Pre-commit the experiment metric, threshold, and measurement window, then re-audit the full funnel after the experiment because the bottleneck can move.
\n\n\n## Reference(s):
\n- [Primary Sources for AARRR Pirate Metrics](references/sources.md)
\n- [ClawHub Skill Page](https://clawhub.ai/deciqai/skills/aarrr-pirate-metrics)
\n- [Startup Metrics for Pirates](https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version)
\n- [Dropbox Form S-1](https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm)
\n- [Introducing ChatGPT](https://openai.com/blog/chatgpt)
\n- [AI's $600B Question](https://www.sequoiacap.com/article/ais-600b-question/)
\n\n\n## Skill Output:
\n**Output Type(s):** [text, markdown, guidance]
\n**Output Format:** [Markdown guidance with a structured funnel audit template]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Produces staged questions, stage definitions, conversion comparisons, bottleneck identification, and a proposed experiment.]
\n\n## Skill Version(s):
\n1.0.5 (source: ClawHub release evidence)
\n\n## Ethical Considerations:
\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.
\n\nArchive v1.0.4: 6 files, 11292 bytes\n\nFiles: examples/dropbox-referral-program-2009.md (4627b), examples/facebook-7-friends-in-10-days-2008.md (4428b), references/sources.md (1557b), skill-card.md (2353b), SKILL.md (8539b), _meta.json (139b)\n\nFile v1.0.4:SKILL.md\n\n---\nname: aarrr-pirate-metrics\ndescription: \"Activate when: user says 'our growth is stalling and we don't know why,' 'product and marketing are arguing about whose fault it is,' 'where is our funnel breaking,' 'pirate metrics,' or 'acquisition vs activation vs retention.' Also activate for: designing end-to-end metrics for a new product, building a shared instrumentation framework across teams.\n Do NOT activate when: the product has no users yet (use lean-startup instead), or the bottleneck is already obvious and obvious to fix.\"\n---\n\n# AARRR (Pirate Metrics)\n\n## Overview\n\nA startup's growth is a **sequential funnel** — each stage gates the next. Great Acquisition is worthless if Activation is broken; great Activation is worthless if Retention is zero. Optimizing the wrong stage produces work that looks like progress while the bottleneck stays.\n\nThe **AARRR framework** (Dave McClure, *Startup Metrics for Pirates*, 2007) names five stages: Acquisition (do they show up?), Activation (good first experience?), Retention (do they come back?), Referral (do they tell others?), Revenue (do they pay?). The bottleneck stage governs total growth — improving any other stage produces no system-level gain (Goldratt, *The Goal*, 1984).\n\n**Compose with:** first-principles to identify your specific Activation event; pmf-crossing-the-chasm to recognize when Retention will always be the bottleneck pre-PMF; probabilistic-thinking to set base rates per stage.\n\n## When to Use\n\n- Growth is **stalling** and the cause is not obvious — multiple teams have plausible explanations\n- Product, marketing, and sales are **arguing about whose problem it is**\n- You need a **shared instrumentation framework** across teams\n- Designing **end-to-end metrics for a new product**\n- Someone says: *\"AARRR,\" \"pirate metrics,\" \"growth funnel,\" \"where is our funnel breaking?\"*\n\n**When NOT to use:** No users yet → use lean-startup. Signups still in the tens — AARRR rates need volume. Bottleneck already obvious → fix that first, then return.\n\n## Coaching Novices (Adaptive Front Door)\n\n- **Engine mode:** user has funnel data and wants the bottleneck named → run The Process directly.\n- **Coach mode:** vague situation or unfamiliarity → guide step by step.\n\nIn Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.\n\n1. **One-line what-it-is.** AARRR is a five-stage funnel (Acquisition / Activation / Retention / Referral / Revenue) — the point is to find which stage is broken so you don't waste effort on the wrong one.\n2. **Check fit.** No users / tiny base → redirect. Bottleneck already obvious → fix it first.\n3. **Elicit their real case.** Force the user to define their *specific* Activation event — not \"signed up\" but e.g. *\"completed onboarding and used the core feature once within day 1.\"* > **[WAIT — do not advance until user responds]**\n4. **One stage at a time.** Walk Acquisition → Activation → Retention → Referral → Revenue. Compute conversion rate stage-to-stage if possible. > **[WAIT — do not advance until user responds]**\n5. **Close by naming the bottleneck and the next experiment.** They leave with one stage identified and one specific experiment. > **[WAIT — do not advance until user responds]**\n\n## The Process\n\nRun the **Funnel Audit**: define stages, measure conversions, identify bottleneck.\n\n1. **Define each stage for this specific product.** Generic definitions hide the bottleneck. Activation = the *aha moment*, not signup. Name a measurable event per stage.\n2. **Measure conversion at each stage.** Acquisition→Activation %, Activation→Retention %, Retention→Referral %, Activation/Retention→Revenue %.\n3. **Compare against domain benchmarks.** A stage below benchmark is a bottleneck candidate (B2B SaaS day-30 retention 70%+; consumer freemium paid conversion 1–5% typical, 5–10% strong).\n4. **Identify the load-bearing bottleneck.** Worst conversion *relative to its benchmark* — fixing it produces the largest system-level gain.\n5. **Pre-commit one experiment.** Name the specific change, the metric, and the pre-committed threshold (per lean-startup).\n6. **Re-measure and re-identify.** After the experiment, the bottleneck moves. Repeat — the audit is iterative.\n\n### Output: Funnel Audit\n\n```\n# AARRR Funnel Audit: \nStage definitions: Acquisition | Activation | Retention | Referral | Revenue \nConversions: [Stage | Users | Conv.% from prior | vs. benchmark ↑↓]\nBottleneck: \nExperiment: Change · Metric · Threshold · Window \n```\n\n*→ Method in Action: [Dropbox's Referral Program (2009)](examples/dropbox-referral-program-2009.md) · [Facebook's \"7 Friends in 10 Days\" (2007–2009)](examples/facebook-7-friends-in-10-days-2008.md)*\n\n## Funnel Packs\n\nStage definitions and benchmarks are domain-specific. **Consumer freemium:** Activation = first core value; paid conversion 1–5% typical, 5–10% strong. **B2B SaaS:** Activation = first non-trivial team use week 1; gross retention 90%+, net 100%+ mid-market. **Marketplaces:** Activation = first transaction; bottleneck = liquidity-thin side. **E-commerce:** Activation = first order; Retention = second order within 90 days.\n\nContribution: add a pack for your domain — one file with (a) stage definitions, (b) benchmarks, (c) typical bottleneck pattern, (d) canonical experiments.\n\n## Applying It Well\n\n- **Activation is the most often-sloppily-defined stage.** \"Signed up\" is not Activation. Specify the aha moment event.\n- **The bottleneck is relative to benchmark** — not the stage with the lowest absolute number.\n- **Fixing the bottleneck moves the bottleneck.** Re-identify after each experiment.\n- **Each team optimizes the stage they own** — dangerous if the bottleneck is elsewhere. The audit forces cross-team prioritization.\n- **Domain benchmarks matter more than absolute numbers.** 3% paid conversion = great freemium, terrible enterprise SaaS.\n\n*→ Primary sources: [references/sources.md](references/sources.md)*\n\n## Common Rationalizations\n\n**[D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.**\n\n| Fake move | Reality |\n|---|---|\n| [D] **Optimizing Acquisition while Activation is broken** | Most common failure. More users into a broken funnel = more wasted CAC. Fix the bottleneck first. |\n| [D] **\"Activation\" = \"signed up\"** | Signup is end of Acquisition. Activation = first time user experiences the value. Sloppy definition hides the real funnel. |\n| [D] **Treating low absolute conversion as the bottleneck** | Acquisition always has the lowest absolute count. What matters: each stage *relative to its domain benchmark*. |\n| [D] **Adding paid ads when Retention is weak** | A leaky bucket doesn't fill faster when bigger. Fix Retention before scaling Acquisition. |\n| [D] **One-shot audit, never re-run** | The bottleneck moves once fixed. Running the audit once and never again misses the next bottleneck. |\n| *→ Add [O] entries here after each real use — paste the actual failure pattern* | *What went wrong and why* |\n\n## Red Flags\n\n- Activation is defined as \"signed up\"\n- Dashboards show absolute counts but not stage-to-stage conversion rates\n- No domain benchmarks referenced\n- Resource allocation matches which stage is easy to instrument, not which is the bottleneck\n- The funnel audit has been run once and never repeated\n- Teams don't share a bottleneck consensus\n- Acquisition spend scaled while Retention is below benchmark\n\n## Verification\n\n- [ ] Each stage has a product-specific measurable definition (Activation = aha moment, not signup)\n- [ ] Conversion rates between stages measured, not just absolute counts\n- [ ] Each stage compared against a domain benchmark\n- [ ] Load-bearing bottleneck named (worst conversion relative to benchmark)\n- [ ] Experiment pre-committed with metric + threshold + time window\n- [ ] Re-audit scheduled (bottleneck will move)\n- [ ] All teams agree on the bottleneck\n\n---\n\n*Part of **deciqAI Knowledge Skills** — 164 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. **See it run → https://www.deciqai.com/c/aarrr-pirate-metrics** · ⭐ Star the repo → https://github.com/deciqAI/knowledge-skills · Contributions welcome.*\n\nFile v1.0.4:_meta.json\n\n{\n \"ownerId\": \"kn754b8sk22s8c6gjxt02bftbn88q7ye\",\n \"slug\": \"aarrr-pirate-metrics\",\n \"version\": \"1.0.4\",\n \"publishedAt\": 1783507972785\n}\n\nFile v1.0.4:references/sources.md\n\n# Sources — aarrr-pirate-metrics\n\n> *Primary sources for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\n- **McClure, Dave.** *Startup Metrics for Pirates* (slide deck, Startonomics SF, August 2007). Original SlideShare: https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version **Canonical primary source for AARRR.**\n- **Goldratt, Eliyahu M.** *The Goal: A Process of Ongoing Improvement*. North River Press, 1984; rev. 3rd ed. 2014. **The Theory of Constraints** that underpins the bottleneck logic of AARRR.\n- **Ellis, Sean & Brown, Morgan.** *Hacking Growth: How Today's Fastest-Growing Companies Drive Breakout Success*. Crown Business, 2017. Growth-marketing case studies including the Dropbox referral case and Facebook's \"7 friends in 10 days\" activation metric.\n- **Palihapitiya, Chamath.** \"How We Put Facebook on the Path to a Billion Users\" (public talk, 2013). Primary first-person account of Facebook's growth team and the \"7 friends in 10 days\" north-star activation metric.\n- **Dropbox Inc., Form S-1 (SEC, 2018)** — primary-source user growth milestones: https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm\n- **Sequoia Capital case studies on Dropbox** — primary-source venture history: https://www.sequoiacap.com/\n- The popular framing \"growth hacking\" is **not** cited as a source — by this skill's own rule, an aphorism is not evidence. The framework is precisely \"instrument the funnel, find the bottleneck stage relative to benchmark, concentrate effort there, re-measure.\"\n\nFile v1.0.4:examples/dropbox-referral-program-2009.md\n\n# Method in Action: Dropbox's Referral Program (2009)\n\n> *Example for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\nA worked example. Not a marketing legend — primary-source documented in Drew Houston's interviews and growth-marketing case studies.\n\nBy early 2009, **Dropbox** had achieved PMF (see [pmf-crossing-the-chasm](../pmf-crossing-the-chasm/SKILL.md)) and had ~100,000 users from the 2007 video MVP + 2008 launch. The team ran an AARRR audit on the funnel:\n\n- **Acquisition**: Paid Google AdWords campaigns were running, with **customer acquisition cost (CAC) ~$200–$300 per user** — but Dropbox's price point was $99/year ($9.99/month), making **AdWords economically broken** as the dominant channel.\n- **Activation**: Strong. Once users installed the desktop client and put a file in the Dropbox folder, the experience converted to \"aha\" cleanly (the *file appeared on another device* moment, the core delight from the 2007 video MVP).\n- **Retention**: Strong. Users who activated tended to stay; Dropbox solved a real persistent need.\n- **Referral**: Untapped. Users loved the product and told friends informally, but there was no built-in mechanism.\n- **Revenue**: Decent at the free-to-paid conversion typical for freemium (a few percent).\n\nThe audit identified **Acquisition** as the load-bearing bottleneck — *not because Acquisition was bad in absolute terms, but because the acquisition channel that was working economically did not exist*. Spending more on AdWords would not solve it; only finding a sub-CAC channel would.\n\n**The intervention (September 2008–April 2009):** Dropbox launched a **referral program**. The mechanic: **give 500 MB free storage to both the referrer and the referee** when an invited friend signed up and installed Dropbox. The reward was native to the product (storage), aligned with the product's value (more space), and free for Dropbox to give away (storage was Dropbox's marginal cost, not a cash outlay).\n\nThe result, documented by Drew Houston and reproduced in Sean Ellis & Morgan Brown's *Hacking Growth*: **signups grew 60% as a result of the referral program**, with **35% of daily signups coming through referrals by late 2009**. By April 2010, Dropbox had grown to **4 million users**, then **25 million** by 2011. The referral program effectively **converted a Retention strength into Acquisition output** — Goldratt's \"exploit the constraint\" move applied to a customer funnel.\n\nWalk the audit on Dropbox 2009:\n\n- **Stage definitions (Step 1):** Acquisition = ad click → landing page → signup; Activation = client installed + first file synced; Retention = active in day 7; Referral = invite sent → friend installs + syncs file; Revenue = upgrade from free to paid tier.\n- **Conversions (Step 2):** Acquisition CAC vs LTV economically broken; Activation, Retention, Revenue all benchmarks-fine; Referral nominal.\n- **Bottleneck (Step 4):** Acquisition channel economics — *specifically*, the cost-effective channel did not exist.\n- **Experiment (Step 5):** Build referral program; pre-committed threshold = signup growth ≥ 30% from referral source within 6 months.\n- **Result:** 60% growth; threshold cleared by ~2×. *Persevere*; double down on the channel; iterate on referral mechanics.\n\n**The non-obvious lesson:** Dropbox's \"referral was the bottleneck\" framing would have been wrong. The bottleneck was Acquisition; the *intervention* was building a referral mechanism *because* Retention was strong enough that activated users would actually send referrals. **The referral program was the solution to the Acquisition bottleneck**, not a Referral-stage improvement in isolation.\n\n**Sources:** McClure, Dave. ***Startup Metrics for Pirates*** (slide deck, Startonomics SF, August 2007). Original SlideShare: https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version . **Canonical primary source for AARRR**. Goldratt, Eliyahu M. ***The Goal: A Process of Ongoing Improvement***. North River Press, 1984; rev. 3rd ed. 2014. **Theory of Constraints**, the bottleneck logic underpinning AARRR. Ellis, Sean & Brown, Morgan. ***Hacking Growth: How Today's Fastest-Growing Companies Drive Breakout Success***. Crown Business, 2017, ch. 5 — primary growth-marketing case account of the Dropbox referral program. Houston, Drew. Drew Houston's interviews on growth, including talks at Stanford and the Sequoia Capital growth case study at: https://www.sequoiacap.com/ Dropbox SEC S-1 filing (2018) provided official user-growth milestones: https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm\n\nFile v1.0.4:examples/facebook-7-friends-in-10-days-2008.md\n\n# Method in Action: Facebook's \"7 Friends in 10 Days\" Activation Metric (2007–2009)\n\n> *Example for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\nA worked example from a social network — a different domain and a different bottleneck stage than the Dropbox case. Documented in Chamath Palihapitiya's public 2013 talk on Facebook's growth team and in Sean Ellis & Morgan Brown's *Hacking Growth* (2017).\n\nIn 2007–2008, **Facebook's** growth flattened at roughly 90 million monthly users. Inside the company, some believed the product might simply be capped near 100 million — that social networks saturate. Instead of accepting the ceiling, Facebook formed a dedicated Growth team under Chamath Palihapitiya and ran what was, in effect, a funnel audit:\n\n- **Acquisition**: Healthy. Millions of people were still landing on Facebook and signing up every month. Registrations were not the problem.\n- **Activation**: Broken — and hidden by the sloppy definition. \"Signed up\" looked fine; but a large fraction of new registrants arrived to an empty feed, saw no friends, and never returned. Signup was the end of Acquisition, not Activation.\n- **Retention**: Weak *downstream of* Activation. Users who never connected to their real-world friends had nothing to come back for; users who did connect retained strongly.\n- **Referral**: Structurally strong — a social network's core loop is inviting and finding friends — but only activated users exercised it.\n- **Revenue**: Advertising-driven; entirely a function of retained, engaged users. Not the constraint.\n\nThe Growth team's decisive move was **Step 1 of the audit: define Activation as a measurable, product-specific event.** By correlating early behavior with long-term retention, they found the aha moment: **a new user who reached 7 friends within 10 days of signup retained; a user who didn't, churned.** \"7 friends in 10 days\" became the team's north-star activation metric — not a vanity registration count, but the event at which a new user first experienced the product's actual value (a feed full of people they knew).\n\n**The intervention:** With Activation named as the load-bearing bottleneck, the Growth team re-pointed the company's effort at one number. Onboarding was rebuilt around friend-finding — contact importers, the \"People You May Know\" recommendation engine, prompts that pushed every new user toward their first connections — and internationalization (a crowdsourced translation platform) removed the language barrier that kept non-English users from finding their friends at all. Every experiment was judged by whether it moved new users toward the 7-in-10 threshold, then re-measured — the audit run as a loop, not a one-shot.\n\n**The result:** Growth resumed through the supposed ceiling. Facebook passed 500 million users in 2010 and 1 billion monthly users in October 2012 — the trajectory Palihapitiya's talk describes as the direct product of instrumenting the funnel and concentrating on the activation constraint rather than buying more traffic into a leaky bucket.\n\n**The non-obvious lesson:** Facebook's bottleneck was invisible as long as Activation meant \"registered.\" The single highest-leverage act was definitional — replacing a generic stage label with a measured aha-moment event. Only then did the funnel show where growth was actually dying, and only then could every team optimize the same constraint instead of the stage each happened to own.\n\nThe mapped steps:\n1. Define stages product-specifically: Activation = 7 friends within 10 days, not signup\n2. Measure conversions: registrations healthy; signup→activated conversion the weak link; retention strong conditional on activation\n3. Benchmark: activated users retained at healthy social-network rates; non-activated users churned — the gap localized the bottleneck\n4. Identify the load-bearing bottleneck: Activation, not Acquisition\n5. Pre-commit experiments against the single metric: onboarding friend-finders, People You May Know, translations — each judged by movement toward 7-in-10\n6. Re-measure and repeat: continuous experiment loop; growth resumed to 1 billion users by 2012\n\nPrimary source: Palihapitiya, Chamath. \"How We Put Facebook on the Path to a Billion Users\" (public talk, 2013). Case account: Ellis, Sean & Brown, Morgan. *Hacking Growth: How Today's Fastest-Growing Companies Drive Breakout Success*. Crown Business, 2017.\n\nFile v1.0.4:skill-card.md\n\n## Description:
\nGuides agents through an AARRR funnel audit to define product-specific growth stages, measure stage-to-stage conversion, identify the bottleneck relative to benchmarks, and propose a focused experiment.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[deciqai](https://clawhub.ai/user/deciqai)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nProduct, growth, marketing, and sales teams use this skill to diagnose where a live product's growth funnel is breaking. It helps an agent define Acquisition, Activation, Retention, Referral, and Revenue events, compare conversion rates to domain benchmarks, and recommend one bottleneck-focused experiment.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: Benchmark numbers and historical examples may be outdated or unsuitable for a specific market.
\nMitigation: Treat benchmarks and examples as guidance, and verify current domain benchmarks before major product, marketing, or budget decisions.
\n\n\n## Reference(s):
\n- [Sources - aarrr-pirate-metrics](references/sources.md)
\n- [Dropbox Referral Program example](examples/dropbox-referral-program-2009.md)
\n- [Facebook 7 Friends in 10 Days example](examples/facebook-7-friends-in-10-days-2008.md)
\n- [Startup Metrics for Pirates](https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version)
\n- [Dropbox Form S-1](https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm)
\n- [Sequoia Capital](https://www.sequoiacap.com/)
\n\n\n## Skill Output:
\n**Output Type(s):** [Text, Markdown, Guidance]
\n**Output Format:** [Markdown funnel audit with stage definitions, conversion table, bottleneck, and experiment recommendation]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Interactive coaching mode may pause for user input before completing the audit.]
\n\n## Skill Version(s):
\n1.0.4 (source: server-resolved release evidence)
\n\n## Ethical Considerations:
\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.
\n\nArchive v1.0.3: 6 files, 11478 bytes\n\nFiles: examples/dropbox-referral-program-2009.md (4627b), examples/facebook-7-friends-in-10-days-2008.md (4428b), references/sources.md (1557b), skill-card.md (2684b), SKILL.md (8649b), _meta.json (139b)\n\nFile v1.0.3:SKILL.md\n\n---\nname: aarrr-pirate-metrics\ndescription: \"Activate when: user says 'our growth is stalling and we don't know why,' 'product and marketing are arguing about whose fault it is,' 'where is our funnel breaking,' 'pirate metrics,' or 'acquisition vs activation vs retention.' Also activate for: designing end-to-end metrics for a new product, building a shared instrumentation framework across teams.\n Do NOT activate when: the product has no users yet (use lean-startup instead), or the bottleneck is already obvious and obvious to fix.\"\n---\n\n# AARRR (Pirate Metrics)\n\n## Overview\n\nA startup's growth is a **sequential funnel** — each stage gates the next. Great Acquisition is worthless if Activation is broken; great Activation is worthless if Retention is zero. Optimizing the wrong stage produces work that looks like progress while the bottleneck stays.\n\nThe **AARRR framework** (Dave McClure, *Startup Metrics for Pirates*, 2007) names five stages: Acquisition (do they show up?), Activation (good first experience?), Retention (do they come back?), Referral (do they tell others?), Revenue (do they pay?). The bottleneck stage governs total growth — improving any other stage produces no system-level gain (Goldratt, *The Goal*, 1984).\n\n**Compose with:** first-principles to identify your specific Activation event; pmf-crossing-the-chasm to recognize when Retention will always be the bottleneck pre-PMF; probabilistic-thinking to set base rates per stage.\n\n## When to Use\n\n- Growth is **stalling** and the cause is not obvious — multiple teams have plausible explanations\n- Product, marketing, and sales are **arguing about whose problem it is**\n- You need a **shared instrumentation framework** across teams\n- Designing **end-to-end metrics for a new product**\n- Someone says: *\"AARRR,\" \"pirate metrics,\" \"growth funnel,\" \"where is our funnel breaking?\"*\n\n**When NOT to use:** No users yet → use lean-startup. Signups still in the tens — AARRR rates need volume. Bottleneck already obvious → fix that first, then return.\n\n## Coaching Novices (Adaptive Front Door)\n\n- **Engine mode:** user has funnel data and wants the bottleneck named → run The Process directly.\n- **Coach mode:** vague situation or unfamiliarity → guide step by step.\n\nIn Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.\n\n1. **One-line what-it-is.** AARRR is a five-stage funnel (Acquisition / Activation / Retention / Referral / Revenue) — the point is to find which stage is broken so you don't waste effort on the wrong one.\n2. **Check fit.** No users / tiny base → redirect. Bottleneck already obvious → fix it first.\n3. **Elicit their real case.** Force the user to define their *specific* Activation event — not \"signed up\" but e.g. *\"completed onboarding and used the core feature once within day 1.\"* > **[WAIT — do not advance until user responds]**\n4. **One stage at a time.** Walk Acquisition → Activation → Retention → Referral → Revenue. Compute conversion rate stage-to-stage if possible. > **[WAIT — do not advance until user responds]**\n5. **Close by naming the bottleneck and the next experiment.** They leave with one stage identified and one specific experiment. > **[WAIT — do not advance until user responds]**\n\n## The Process\n\nRun the **Funnel Audit**: define stages, measure conversions, identify bottleneck.\n\n1. **Define each stage for this specific product.** Generic definitions hide the bottleneck. Activation = the *aha moment*, not signup. Name a measurable event per stage.\n2. **Measure conversion at each stage.** Acquisition→Activation %, Activation→Retention %, Retention→Referral %, Activation/Retention→Revenue %.\n3. **Compare against domain benchmarks.** A stage below benchmark is a bottleneck candidate (B2B SaaS day-30 retention 70%+; consumer freemium paid conversion 1–5% typical, 5–10% strong).\n4. **Identify the load-bearing bottleneck.** Worst conversion *relative to its benchmark* — fixing it produces the largest system-level gain.\n5. **Pre-commit one experiment.** Name the specific change, the metric, and the pre-committed threshold (per lean-startup).\n6. **Re-measure and re-identify.** After the experiment, the bottleneck moves. Repeat — the audit is iterative.\n\n### Output: Funnel Audit\n\n```\n# AARRR Funnel Audit: \nStage definitions: Acquisition | Activation | Retention | Referral | Revenue \nConversions: [Stage | Users | Conv.% from prior | vs. benchmark ↑↓]\nBottleneck: \nExperiment: Change · Metric · Threshold · Window \n```\n\n*→ Method in Action: [Dropbox's Referral Program (2009)](examples/dropbox-referral-program-2009.md) · [Facebook's \"7 Friends in 10 Days\" (2007–2009)](examples/facebook-7-friends-in-10-days-2008.md)*\n\n## Funnel Packs\n\nStage definitions and benchmarks are domain-specific. **Consumer freemium:** Activation = first core value; paid conversion 1–5% typical, 5–10% strong. **B2B SaaS:** Activation = first non-trivial team use week 1; gross retention 90%+, net 100%+ mid-market. **Marketplaces:** Activation = first transaction; bottleneck = liquidity-thin side. **E-commerce:** Activation = first order; Retention = second order within 90 days.\n\nContribution: add a pack for your domain — one file with (a) stage definitions, (b) benchmarks, (c) typical bottleneck pattern, (d) canonical experiments.\n\n## Applying It Well\n\n- **Activation is the most often-sloppily-defined stage.** \"Signed up\" is not Activation. Specify the aha moment event.\n- **The bottleneck is relative to benchmark** — not the stage with the lowest absolute number.\n- **Fixing the bottleneck moves the bottleneck.** Re-identify after each experiment.\n- **Each team optimizes the stage they own** — dangerous if the bottleneck is elsewhere. The audit forces cross-team prioritization.\n- **Domain benchmarks matter more than absolute numbers.** 3% paid conversion = great freemium, terrible enterprise SaaS.\n\n*→ Primary sources: [references/sources.md](references/sources.md)*\n\n## Common Rationalizations\n\n**[D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.**\n\n| Fake move | Reality |\n|---|---|\n| [D] **Optimizing Acquisition while Activation is broken** | Most common failure. More users into a broken funnel = more wasted CAC. Fix the bottleneck first. |\n| [D] **\"Activation\" = \"signed up\"** | Signup is end of Acquisition. Activation = first time user experiences the value. Sloppy definition hides the real funnel. |\n| [D] **Treating low absolute conversion as the bottleneck** | Acquisition always has the lowest absolute count. What matters: each stage *relative to its domain benchmark*. |\n| [D] **Adding paid ads when Retention is weak** | A leaky bucket doesn't fill faster when bigger. Fix Retention before scaling Acquisition. |\n| [D] **One-shot audit, never re-run** | The bottleneck moves once fixed. Running the audit once and never again misses the next bottleneck. |\n| *→ Add [O] entries here after each real use — paste the actual failure pattern* | *What went wrong and why* |\n\n## Red Flags\n\n- Activation is defined as \"signed up\"\n- Dashboards show absolute counts but not stage-to-stage conversion rates\n- No domain benchmarks referenced\n- Resource allocation matches which stage is easy to instrument, not which is the bottleneck\n- The funnel audit has been run once and never repeated\n- Teams don't share a bottleneck consensus\n- Acquisition spend scaled while Retention is below benchmark\n\n## Verification\n\n- [ ] Each stage has a product-specific measurable definition (Activation = aha moment, not signup)\n- [ ] Conversion rates between stages measured, not just absolute counts\n- [ ] Each stage compared against a domain benchmark\n- [ ] Load-bearing bottleneck named (worst conversion relative to benchmark)\n- [ ] Experiment pre-committed with metric + threshold + time window\n- [ ] Re-audit scheduled (bottleneck will move)\n- [ ] All teams agree on the bottleneck\n\n---\n\n*Part of **deciqAI Knowledge Skills** — 163 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. **See it run → https://www.deciqai.com/skills/aarrr-pirate-metrics?utm_source=clawhub&utm_medium=marketplace&utm_campaign=knowledge-skills&utm_content=aarrr-pirate-metrics** · ⭐ Star the repo → https://github.com/deciqAI/knowledge-skills · Contributions welcome.*\n\nFile v1.0.3:_meta.json\n\n{\n \"ownerId\": \"kn754b8sk22s8c6gjxt02bftbn88q7ye\",\n \"slug\": \"aarrr-pirate-metrics\",\n \"version\": \"1.0.3\",\n \"publishedAt\": 1783482618702\n}\n\nFile v1.0.3:references/sources.md\n\n# Sources — aarrr-pirate-metrics\n\n> *Primary sources for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\n- **McClure, Dave.** *Startup Metrics for Pirates* (slide deck, Startonomics SF, August 2007). Original SlideShare: https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version **Canonical primary source for AARRR.**\n- **Goldratt, Eliyahu M.** *The Goal: A Process of Ongoing Improvement*. North River Press, 1984; rev. 3rd ed. 2014. **The Theory of Constraints** that underpins the bottleneck logic of AARRR.\n- **Ellis, Sean & Brown, Morgan.** *Hacking Growth: How Today's Fastest-Growing Companies Drive Breakout Success*. Crown Business, 2017. Growth-marketing case studies including the Dropbox referral case and Facebook's \"7 friends in 10 days\" activation metric.\n- **Palihapitiya, Chamath.** \"How We Put Facebook on the Path to a Billion Users\" (public talk, 2013). Primary first-person account of Facebook's growth team and the \"7 friends in 10 days\" north-star activation metric.\n- **Dropbox Inc., Form S-1 (SEC, 2018)** — primary-source user growth milestones: https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm\n- **Sequoia Capital case studies on Dropbox** — primary-source venture history: https://www.sequoiacap.com/\n- The popular framing \"growth hacking\" is **not** cited as a source — by this skill's own rule, an aphorism is not evidence. The framework is precisely \"instrument the funnel, find the bottleneck stage relative to benchmark, concentrate effort there, re-measure.\"\n\nFile v1.0.3:examples/dropbox-referral-program-2009.md\n\n# Method in Action: Dropbox's Referral Program (2009)\n\n> *Example for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\nA worked example. Not a marketing legend — primary-source documented in Drew Houston's interviews and growth-marketing case studies.\n\nBy early 2009, **Dropbox** had achieved PMF (see [pmf-crossing-the-chasm](../pmf-crossing-the-chasm/SKILL.md)) and had ~100,000 users from the 2007 video MVP + 2008 launch. The team ran an AARRR audit on the funnel:\n\n- **Acquisition**: Paid Google AdWords campaigns were running, with **customer acquisition cost (CAC) ~$200–$300 per user** — but Dropbox's price point was $99/year ($9.99/month), making **AdWords economically broken** as the dominant channel.\n- **Activation**: Strong. Once users installed the desktop client and put a file in the Dropbox folder, the experience converted to \"aha\" cleanly (the *file appeared on another device* moment, the core delight from the 2007 video MVP).\n- **Retention**: Strong. Users who activated tended to stay; Dropbox solved a real persistent need.\n- **Referral**: Untapped. Users loved the product and told friends informally, but there was no built-in mechanism.\n- **Revenue**: Decent at the free-to-paid conversion typical for freemium (a few percent).\n\nThe audit identified **Acquisition** as the load-bearing bottleneck — *not because Acquisition was bad in absolute terms, but because the acquisition channel that was working economically did not exist*. Spending more on AdWords would not solve it; only finding a sub-CAC channel would.\n\n**The intervention (September 2008–April 2009):** Dropbox launched a **referral program**. The mechanic: **give 500 MB free storage to both the referrer and the referee** when an invited friend signed up and installed Dropbox. The reward was native to the product (storage), aligned with the product's value (more space), and free for Dropbox to give away (storage was Dropbox's marginal cost, not a cash outlay).\n\nThe result, documented by Drew Houston and reproduced in Sean Ellis & Morgan Brown's *Hacking Growth*: **signups grew 60% as a result of the referral program**, with **35% of daily signups coming through referrals by late 2009**. By April 2010, Dropbox had grown to **4 million users**, then **25 million** by 2011. The referral program effectively **converted a Retention strength into Acquisition output** — Goldratt's \"exploit the constraint\" move applied to a customer funnel.\n\nWalk the audit on Dropbox 2009:\n\n- **Stage definitions (Step 1):** Acquisition = ad click → landing page → signup; Activation = client installed + first file synced; Retention = active in day 7; Referral = invite sent → friend installs + syncs file; Revenue = upgrade from free to paid tier.\n- **Conversions (Step 2):** Acquisition CAC vs LTV economically broken; Activation, Retention, Revenue all benchmarks-fine; Referral nominal.\n- **Bottleneck (Step 4):** Acquisition channel economics — *specifically*, the cost-effective channel did not exist.\n- **Experiment (Step 5):** Build referral program; pre-committed threshold = signup growth ≥ 30% from referral source within 6 months.\n- **Result:** 60% growth; threshold cleared by ~2×. *Persevere*; double down on the channel; iterate on referral mechanics.\n\n**The non-obvious lesson:** Dropbox's \"referral was the bottleneck\" framing would have been wrong. The bottleneck was Acquisition; the *intervention* was building a referral mechanism *because* Retention was strong enough that activated users would actually send referrals. **The referral program was the solution to the Acquisition bottleneck**, not a Referral-stage improvement in isolation.\n\n**Sources:** McClure, Dave. ***Startup Metrics for Pirates*** (slide deck, Startonomics SF, August 2007). Original SlideShare: https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version . **Canonical primary source for AARRR**. Goldratt, Eliyahu M. ***The Goal: A Process of Ongoing Improvement***. North River Press, 1984; rev. 3rd ed. 2014. **Theory of Constraints**, the bottleneck logic underpinning AARRR. Ellis, Sean & Brown, Morgan. ***Hacking Growth: How Today's Fastest-Growing Companies Drive Breakout Success***. Crown Business, 2017, ch. 5 — primary growth-marketing case account of the Dropbox referral program. Houston, Drew. Drew Houston's interviews on growth, including talks at Stanford and the Sequoia Capital growth case study at: https://www.sequoiacap.com/ Dropbox SEC S-1 filing (2018) provided official user-growth milestones: https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm\n\nFile v1.0.3:examples/facebook-7-friends-in-10-days-2008.md\n\n# Method in Action: Facebook's \"7 Friends in 10 Days\" Activation Metric (2007–2009)\n\n> *Example for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\nA worked example from a social network — a different domain and a different bottleneck stage than the Dropbox case. Documented in Chamath Palihapitiya's public 2013 talk on Facebook's growth team and in Sean Ellis & Morgan Brown's *Hacking Growth* (2017).\n\nIn 2007–2008, **Facebook's** growth flattened at roughly 90 million monthly users. Inside the company, some believed the product might simply be capped near 100 million — that social networks saturate. Instead of accepting the ceiling, Facebook formed a dedicated Growth team under Chamath Palihapitiya and ran what was, in effect, a funnel audit:\n\n- **Acquisition**: Healthy. Millions of people were still landing on Facebook and signing up every month. Registrations were not the problem.\n- **Activation**: Broken — and hidden by the sloppy definition. \"Signed up\" looked fine; but a large fraction of new registrants arrived to an empty feed, saw no friends, and never returned. Signup was the end of Acquisition, not Activation.\n- **Retention**: Weak *downstream of* Activation. Users who never connected to their real-world friends had nothing to come back for; users who did connect retained strongly.\n- **Referral**: Structurally strong — a social network's core loop is inviting and finding friends — but only activated users exercised it.\n- **Revenue**: Advertising-driven; entirely a function of retained, engaged users. Not the constraint.\n\nThe Growth team's decisive move was **Step 1 of the audit: define Activation as a measurable, product-specific event.** By correlating early behavior with long-term retention, they found the aha moment: **a new user who reached 7 friends within 10 days of signup retained; a user who didn't, churned.** \"7 friends in 10 days\" became the team's north-star activation metric — not a vanity registration count, but the event at which a new user first experienced the product's actual value (a feed full of people they knew).\n\n**The intervention:** With Activation named as the load-bearing bottleneck, the Growth team re-pointed the company's effort at one number. Onboarding was rebuilt around friend-finding — contact importers, the \"People You May Know\" recommendation engine, prompts that pushed every new user toward their first connections — and internationalization (a crowdsourced translation platform) removed the language barrier that kept non-English users from finding their friends at all. Every experiment was judged by whether it moved new users toward the 7-in-10 threshold, then re-measured — the audit run as a loop, not a one-shot.\n\n**The result:** Growth resumed through the supposed ceiling. Facebook passed 500 million users in 2010 and 1 billion monthly users in October 2012 — the trajectory Palihapitiya's talk describes as the direct product of instrumenting the funnel and concentrating on the activation constraint rather than buying more traffic into a leaky bucket.\n\n**The non-obvious lesson:** Facebook's bottleneck was invisible as long as Activation meant \"registered.\" The single highest-leverage act was definitional — replacing a generic stage label with a measured aha-moment event. Only then did the funnel show where growth was actually dying, and only then could every team optimize the same constraint instead of the stage each happened to own.\n\nThe mapped steps:\n1. Define stages product-specifically: Activation = 7 friends within 10 days, not signup\n2. Measure conversions: registrations healthy; signup→activated conversion the weak link; retention strong conditional on activation\n3. Benchmark: activated users retained at healthy social-network rates; non-activated users churned — the gap localized the bottleneck\n4. Identify the load-bearing bottleneck: Activation, not Acquisition\n5. Pre-commit experiments against the single metric: onboarding friend-finders, People You May Know, translations — each judged by movement toward 7-in-10\n6. Re-measure and repeat: continuous experiment loop; growth resumed to 1 billion users by 2012\n\nPrimary source: Palihapitiya, Chamath. \"How We Put Facebook on the Path to a Billion Users\" (public talk, 2013). Case account: Ellis, Sean & Brown, Morgan. *Hacking Growth: How Today's Fastest-Growing Companies Drive Breakout Success*. Crown Business, 2017.\n\nFile v1.0.3:skill-card.md\n\n## Description:
\nGuides agents through an AARRR funnel audit to define growth stages, measure conversions against benchmarks, identify the load-bearing bottleneck, and propose a focused experiment.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[deciqai](https://clawhub.ai/user/deciqai)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nExternal product, growth, marketing, and sales teams use this skill to build a shared funnel instrumentation framework, identify where growth is breaking, and choose the next experiment. It is most useful when a product has enough users to compare Acquisition, Activation, Retention, Referral, and Revenue stages.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: The skill's benchmarks and case studies may not match a user's market, product stage, or available data.
\nMitigation: Validate the funnel definitions, conversion rates, and benchmark comparisons against the user's own analytics before acting on recommendations.
\nRisk: Reference and promotional links may lead outside ClawHub if opened manually.
\nMitigation: Review external links before opening them and rely on trusted primary sources when validating case-study claims.
\n\n\n## Reference(s):
\n- [ClawHub Skill Page](https://clawhub.ai/deciqai/skills/aarrr-pirate-metrics)
\n- [Sources - aarrr-pirate-metrics](references/sources.md)
\n- [Dropbox Referral Program Example](examples/dropbox-referral-program-2009.md)
\n- [Facebook 7 Friends in 10 Days Example](examples/facebook-7-friends-in-10-days-2008.md)
\n- [Startup Metrics for Pirates](https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version)
\n- [Dropbox Inc. Form S-1](https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm)
\n- [Sequoia Capital Case Studies](https://www.sequoiacap.com/)
\n\n\n## Skill Output:
\n**Output Type(s):** [text, markdown, guidance]
\n**Output Format:** [Markdown funnel audit with stage definitions, conversion comparison, bottleneck, and experiment recommendation]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [No executable output; guidance depends on user-provided funnel data and domain benchmarks.]
\n\n## Skill Version(s):
\n1.0.3 (source: server release evidence)
\n\n## Ethical Considerations:
\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.
\n\nArchive v1.0.2: 5 files, 8849 bytes\n\nFiles: examples/dropbox-referral-program-2009.md (4627b), references/sources.md (1282b), skill-card.md (2492b), SKILL.md (8548b), _meta.json (139b)\n\nFile v1.0.2:SKILL.md\n\n---\nname: aarrr-pirate-metrics\ndescription: \"Activate when: user says 'our growth is stalling and we don't know why,' 'product and marketing are arguing about whose fault it is,' 'where is our funnel breaking,' 'pirate metrics,' or 'acquisition vs activation vs retention.' Also activate for: designing end-to-end metrics for a new product, building a shared instrumentation framework across teams.\n Do NOT activate when: the product has no users yet (use lean-startup instead), or the bottleneck is already obvious and obvious to fix.\"\n---\n\n# AARRR (Pirate Metrics)\n\n## Overview\n\nA startup's growth is a **sequential funnel** — each stage gates the next. Great Acquisition is worthless if Activation is broken; great Activation is worthless if Retention is zero. Optimizing the wrong stage produces work that looks like progress while the bottleneck stays.\n\nThe **AARRR framework** (Dave McClure, *Startup Metrics for Pirates*, 2007) names five stages: Acquisition (do they show up?), Activation (good first experience?), Retention (do they come back?), Referral (do they tell others?), Revenue (do they pay?). The bottleneck stage governs total growth — improving any other stage produces no system-level gain (Goldratt, *The Goal*, 1984).\n\n**Compose with:** first-principles to identify your specific Activation event; pmf-crossing-the-chasm to recognize when Retention will always be the bottleneck pre-PMF; probabilistic-thinking to set base rates per stage.\n\n## When to Use\n\n- Growth is **stalling** and the cause is not obvious — multiple teams have plausible explanations\n- Product, marketing, and sales are **arguing about whose problem it is**\n- You need a **shared instrumentation framework** across teams\n- Designing **end-to-end metrics for a new product**\n- Someone says: *\"AARRR,\" \"pirate metrics,\" \"growth funnel,\" \"where is our funnel breaking?\"*\n\n**When NOT to use:** No users yet → use lean-startup. Signups still in the tens — AARRR rates need volume. Bottleneck already obvious → fix that first, then return.\n\n## Coaching Novices (Adaptive Front Door)\n\n- **Engine mode:** user has funnel data and wants the bottleneck named → run The Process directly.\n- **Coach mode:** vague situation or unfamiliarity → guide step by step.\n\nIn Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.\n\n1. **One-line what-it-is.** AARRR is a five-stage funnel (Acquisition / Activation / Retention / Referral / Revenue) — the point is to find which stage is broken so you don't waste effort on the wrong one.\n2. **Check fit.** No users / tiny base → redirect. Bottleneck already obvious → fix it first.\n3. **Elicit their real case.** Force the user to define their *specific* Activation event — not \"signed up\" but e.g. *\"completed onboarding and used the core feature once within day 1.\"* > **[WAIT — do not advance until user responds]**\n4. **One stage at a time.** Walk Acquisition → Activation → Retention → Referral → Revenue. Compute conversion rate stage-to-stage if possible. > **[WAIT — do not advance until user responds]**\n5. **Close by naming the bottleneck and the next experiment.** They leave with one stage identified and one specific experiment. > **[WAIT — do not advance until user responds]**\n\n## The Process\n\nRun the **Funnel Audit**: define stages, measure conversions, identify bottleneck.\n\n1. **Define each stage for this specific product.** Generic definitions hide the bottleneck. Activation = the *aha moment*, not signup. Name a measurable event per stage.\n2. **Measure conversion at each stage.** Acquisition→Activation %, Activation→Retention %, Retention→Referral %, Activation/Retention→Revenue %.\n3. **Compare against domain benchmarks.** A stage below benchmark is a bottleneck candidate (B2B SaaS day-30 retention 70%+; consumer freemium paid conversion 1–5% typical, 5–10% strong).\n4. **Identify the load-bearing bottleneck.** Worst conversion *relative to its benchmark* — fixing it produces the largest system-level gain.\n5. **Pre-commit one experiment.** Name the specific change, the metric, and the pre-committed threshold (per lean-startup).\n6. **Re-measure and re-identify.** After the experiment, the bottleneck moves. Repeat — the audit is iterative.\n\n### Output: Funnel Audit\n\n```\n# AARRR Funnel Audit: \nStage definitions: Acquisition | Activation | Retention | Referral | Revenue \nConversions: [Stage | Users | Conv.% from prior | vs. benchmark ↑↓]\nBottleneck: \nExperiment: Change · Metric · Threshold · Window \n```\n\n*→ Method in Action: [Dropbox's Referral Program (2009)](examples/dropbox-referral-program-2009.md)*\n\n## Funnel Packs\n\nStage definitions and benchmarks are domain-specific. **Consumer freemium:** Activation = first core value; paid conversion 1–5% typical, 5–10% strong. **B2B SaaS:** Activation = first non-trivial team use week 1; gross retention 90%+, net 100%+ mid-market. **Marketplaces:** Activation = first transaction; bottleneck = liquidity-thin side. **E-commerce:** Activation = first order; Retention = second order within 90 days.\n\nContribution: add a pack for your domain — one file with (a) stage definitions, (b) benchmarks, (c) typical bottleneck pattern, (d) canonical experiments.\n\n## Applying It Well\n\n- **Activation is the most often-sloppily-defined stage.** \"Signed up\" is not Activation. Specify the aha moment event.\n- **The bottleneck is relative to benchmark** — not the stage with the lowest absolute number.\n- **Fixing the bottleneck moves the bottleneck.** Re-identify after each experiment.\n- **Each team optimizes the stage they own** — dangerous if the bottleneck is elsewhere. The audit forces cross-team prioritization.\n- **Domain benchmarks matter more than absolute numbers.** 3% paid conversion = great freemium, terrible enterprise SaaS.\n\n*→ Primary sources: [references/sources.md](references/sources.md)*\n\n## Common Rationalizations\n\n**[D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.**\n\n| Fake move | Reality |\n|---|---|\n| [D] **Optimizing Acquisition while Activation is broken** | Most common failure. More users into a broken funnel = more wasted CAC. Fix the bottleneck first. |\n| [D] **\"Activation\" = \"signed up\"** | Signup is end of Acquisition. Activation = first time user experiences the value. Sloppy definition hides the real funnel. |\n| [D] **Treating low absolute conversion as the bottleneck** | Acquisition always has the lowest absolute count. What matters: each stage *relative to its domain benchmark*. |\n| [D] **Adding paid ads when Retention is weak** | A leaky bucket doesn't fill faster when bigger. Fix Retention before scaling Acquisition. |\n| [D] **One-shot audit, never re-run** | The bottleneck moves once fixed. Running the audit once and never again misses the next bottleneck. |\n| *→ Add [O] entries here after each real use — paste the actual failure pattern* | *What went wrong and why* |\n\n## Red Flags\n\n- Activation is defined as \"signed up\"\n- Dashboards show absolute counts but not stage-to-stage conversion rates\n- No domain benchmarks referenced\n- Resource allocation matches which stage is easy to instrument, not which is the bottleneck\n- The funnel audit has been run once and never repeated\n- Teams don't share a bottleneck consensus\n- Acquisition spend scaled while Retention is below benchmark\n\n## Verification\n\n- [ ] Each stage has a product-specific measurable definition (Activation = aha moment, not signup)\n- [ ] Conversion rates between stages measured, not just absolute counts\n- [ ] Each stage compared against a domain benchmark\n- [ ] Load-bearing bottleneck named (worst conversion relative to benchmark)\n- [ ] Experiment pre-committed with metric + threshold + time window\n- [ ] Re-audit scheduled (bottleneck will move)\n- [ ] All teams agree on the bottleneck\n\n---\n\n*Part of **deciqAI Knowledge Skills** — 163 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. **See it run → https://www.deciqai.com/skills/aarrr-pirate-metrics?utm_source=clawhub&utm_medium=marketplace&utm_campaign=knowledge-skills&utm_content=aarrr-pirate-metrics** · ⭐ Star the repo → https://github.com/deciqAI/knowledge-skills · Contributions welcome.*\n\nFile v1.0.2:_meta.json\n\n{\n \"ownerId\": \"kn754b8sk22s8c6gjxt02bftbn88q7ye\",\n \"slug\": \"aarrr-pirate-metrics\",\n \"version\": \"1.0.2\",\n \"publishedAt\": 1783471040500\n}\n\nFile v1.0.2:references/sources.md\n\n# Sources — aarrr-pirate-metrics\n\n> *Primary sources for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\n- **McClure, Dave.** *Startup Metrics for Pirates* (slide deck, Startonomics SF, August 2007). Original SlideShare: https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version **Canonical primary source for AARRR.**\n- **Goldratt, Eliyahu M.** *The Goal: A Process of Ongoing Improvement*. North River Press, 1984; rev. 3rd ed. 2014. **The Theory of Constraints** that underpins the bottleneck logic of AARRR.\n- **Ellis, Sean & Brown, Morgan.** *Hacking Growth: How Today's Fastest-Growing Companies Drive Breakout Success*. Crown Business, 2017. Growth-marketing case studies including the Dropbox referral case.\n- **Dropbox Inc., Form S-1 (SEC, 2018)** — primary-source user growth milestones: https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm\n- **Sequoia Capital case studies on Dropbox** — primary-source venture history: https://www.sequoiacap.com/\n- The popular framing \"growth hacking\" is **not** cited as a source — by this skill's own rule, an aphorism is not evidence. The framework is precisely \"instrument the funnel, find the bottleneck stage relative to benchmark, concentrate effort there, re-measure.\"\n\nFile v1.0.2:examples/dropbox-referral-program-2009.md\n\n# Method in Action: Dropbox's Referral Program (2009)\n\n> *Example for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\nA worked example. Not a marketing legend — primary-source documented in Drew Houston's interviews and growth-marketing case studies.\n\nBy early 2009, **Dropbox** had achieved PMF (see [pmf-crossing-the-chasm](../pmf-crossing-the-chasm/SKILL.md)) and had ~100,000 users from the 2007 video MVP + 2008 launch. The team ran an AARRR audit on the funnel:\n\n- **Acquisition**: Paid Google AdWords campaigns were running, with **customer acquisition cost (CAC) ~$200–$300 per user** — but Dropbox's price point was $99/year ($9.99/month), making **AdWords economically broken** as the dominant channel.\n- **Activation**: Strong. Once users installed the desktop client and put a file in the Dropbox folder, the experience converted to \"aha\" cleanly (the *file appeared on another device* moment, the core delight from the 2007 video MVP).\n- **Retention**: Strong. Users who activated tended to stay; Dropbox solved a real persistent need.\n- **Referral**: Untapped. Users loved the product and told friends informally, but there was no built-in mechanism.\n- **Revenue**: Decent at the free-to-paid conversion typical for freemium (a few percent).\n\nThe audit identified **Acquisition** as the load-bearing bottleneck — *not because Acquisition was bad in absolute terms, but because the acquisition channel that was working economically did not exist*. Spending more on AdWords would not solve it; only finding a sub-CAC channel would.\n\n**The intervention (September 2008–April 2009):** Dropbox launched a **referral program**. The mechanic: **give 500 MB free storage to both the referrer and the referee** when an invited friend signed up and installed Dropbox. The reward was native to the product (storage), aligned with the product's value (more space), and free for Dropbox to give away (storage was Dropbox's marginal cost, not a cash outlay).\n\nThe result, documented by Drew Houston and reproduced in Sean Ellis & Morgan Brown's *Hacking Growth*: **signups grew 60% as a result of the referral program**, with **35% of daily signups coming through referrals by late 2009**. By April 2010, Dropbox had grown to **4 million users**, then **25 million** by 2011. The referral program effectively **converted a Retention strength into Acquisition output** — Goldratt's \"exploit the constraint\" move applied to a customer funnel.\n\nWalk the audit on Dropbox 2009:\n\n- **Stage definitions (Step 1):** Acquisition = ad click → landing page → signup; Activation = client installed + first file synced; Retention = active in day 7; Referral = invite sent → friend installs + syncs file; Revenue = upgrade from free to paid tier.\n- **Conversions (Step 2):** Acquisition CAC vs LTV economically broken; Activation, Retention, Revenue all benchmarks-fine; Referral nominal.\n- **Bottleneck (Step 4):** Acquisition channel economics — *specifically*, the cost-effective channel did not exist.\n- **Experiment (Step 5):** Build referral program; pre-committed threshold = signup growth ≥ 30% from referral source within 6 months.\n- **Result:** 60% growth; threshold cleared by ~2×. *Persevere*; double down on the channel; iterate on referral mechanics.\n\n**The non-obvious lesson:** Dropbox's \"referral was the bottleneck\" framing would have been wrong. The bottleneck was Acquisition; the *intervention* was building a referral mechanism *because* Retention was strong enough that activated users would actually send referrals. **The referral program was the solution to the Acquisition bottleneck**, not a Referral-stage improvement in isolation.\n\n**Sources:** McClure, Dave. ***Startup Metrics for Pirates*** (slide deck, Startonomics SF, August 2007). Original SlideShare: https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version . **Canonical primary source for AARRR**. Goldratt, Eliyahu M. ***The Goal: A Process of Ongoing Improvement***. North River Press, 1984; rev. 3rd ed. 2014. **Theory of Constraints**, the bottleneck logic underpinning AARRR. Ellis, Sean & Brown, Morgan. ***Hacking Growth: How Today's Fastest-Growing Companies Drive Breakout Success***. Crown Business, 2017, ch. 5 — primary growth-marketing case account of the Dropbox referral program. Houston, Drew. Drew Houston's interviews on growth, including talks at Stanford and the Sequoia Capital growth case study at: https://www.sequoiacap.com/ Dropbox SEC S-1 filing (2018) provided official user-growth milestones: https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm\n\nFile v1.0.2:skill-card.md\n\n## Description:
\nGuides an agent through an AARRR funnel audit to define product-specific acquisition, activation, retention, referral, and revenue stages, measure stage-to-stage conversion, identify the bottleneck, and propose the next experiment.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[deciqai](https://clawhub.ai/user/deciqai)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nExternal teams, operators, product leaders, and growth teams use this skill to turn ambiguous growth stalls into a shared funnel diagnosis and a concrete experiment. It is most useful when product, marketing, or sales need agreement on which AARRR stage is limiting growth.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: Users may share sensitive customer, revenue, or funnel metrics while asking the agent to diagnose growth.
\nMitigation: Use aggregated or anonymized metrics unless the current AI environment is approved for that business data.
\nRisk: A funnel recommendation can be misleading if activation events or benchmark comparisons are poorly defined.
\nMitigation: Require product-specific stage definitions, compare each stage against relevant domain benchmarks, and review the proposed experiment before execution.
\n\n\n## Reference(s):
\n- [Sources - aarrr-pirate-metrics](references/sources.md)
\n- [Startup Metrics for Pirates](https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version)
\n- [Dropbox Inc. Form S-1](https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm)
\n- [Sequoia Capital Dropbox Case Studies](https://www.sequoiacap.com/)
\n\n\n## Skill Output:
\n**Output Type(s):** [Guidance, Markdown, Analysis]
\n**Output Format:** [Markdown funnel audit with stage definitions, conversion table, bottleneck, and experiment recommendation]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [May ask staged follow-up questions before producing the final audit when the user's funnel data is incomplete.]
\n\n## Skill Version(s):
\n1.0.2 (source: server release evidence)
\n\n## Ethical Considerations:
\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.
\n\nArchive v1.0.1: 5 files, 8754 bytes\n\nFiles: examples/dropbox-referral-program-2009.md (4627b), references/sources.md (1282b), skill-card.md (2517b), SKILL.md (8444b), _meta.json (139b)\n\nFile v1.0.1:SKILL.md\n\n---\nname: aarrr-pirate-metrics\ndescription: \"Activate when: user says 'our growth is stalling and we don't know why,' 'product and marketing are arguing about whose fault it is,' 'where is our funnel breaking,' 'pirate metrics,' or 'acquisition vs activation vs retention.' Also activate for: designing end-to-end metrics for a new product, building a shared instrumentation framework across teams.\n Do NOT activate when: the product has no users yet (use lean-startup instead), or the bottleneck is already obvious and obvious to fix.\"\n---\n\n# AARRR (Pirate Metrics)\n\n## Overview\n\nA startup's growth is a **sequential funnel** — each stage gates the next. Great Acquisition is worthless if Activation is broken; great Activation is worthless if Retention is zero. Optimizing the wrong stage produces work that looks like progress while the bottleneck stays.\n\nThe **AARRR framework** (Dave McClure, *Startup Metrics for Pirates*, 2007) names five stages: Acquisition (do they show up?), Activation (good first experience?), Retention (do they come back?), Referral (do they tell others?), Revenue (do they pay?). The bottleneck stage governs total growth — improving any other stage produces no system-level gain (Goldratt, *The Goal*, 1984).\n\n**Compose with:** [first-principles](../first-principles/SKILL.md) to identify your specific Activation event; [pmf-crossing-the-chasm](../pmf-crossing-the-chasm/SKILL.md) to recognize when Retention will always be the bottleneck pre-PMF; [probabilistic-thinking](../probabilistic-thinking/SKILL.md) to set base rates per stage.\n\n## When to Use\n\n- Growth is **stalling** and the cause is not obvious — multiple teams have plausible explanations\n- Product, marketing, and sales are **arguing about whose problem it is**\n- You need a **shared instrumentation framework** across teams\n- Designing **end-to-end metrics for a new product**\n- Someone says: *\"AARRR,\" \"pirate metrics,\" \"growth funnel,\" \"where is our funnel breaking?\"*\n\n**When NOT to use:** No users yet → use [lean-startup](../lean-startup/SKILL.md). Signups still in the tens — AARRR rates need volume. Bottleneck already obvious → fix that first, then return.\n\n## Coaching Novices (Adaptive Front Door)\n\n- **Engine mode:** user has funnel data and wants the bottleneck named → run The Process directly.\n- **Coach mode:** vague situation or unfamiliarity → guide step by step.\n\nIn Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.\n\n1. **One-line what-it-is.** AARRR is a five-stage funnel (Acquisition / Activation / Retention / Referral / Revenue) — the point is to find which stage is broken so you don't waste effort on the wrong one.\n2. **Check fit.** No users / tiny base → redirect. Bottleneck already obvious → fix it first.\n3. **Elicit their real case.** Force the user to define their *specific* Activation event — not \"signed up\" but e.g. *\"completed onboarding and used the core feature once within day 1.\"* > **[WAIT — do not advance until user responds]**\n4. **One stage at a time.** Walk Acquisition → Activation → Retention → Referral → Revenue. Compute conversion rate stage-to-stage if possible. > **[WAIT — do not advance until user responds]**\n5. **Close by naming the bottleneck and the next experiment.** They leave with one stage identified and one specific experiment. > **[WAIT — do not advance until user responds]**\n\n## The Process\n\nRun the **Funnel Audit**: define stages, measure conversions, identify bottleneck.\n\n1. **Define each stage for this specific product.** Generic definitions hide the bottleneck. Activation = the *aha moment*, not signup. Name a measurable event per stage.\n2. **Measure conversion at each stage.** Acquisition→Activation %, Activation→Retention %, Retention→Referral %, Activation/Retention→Revenue %.\n3. **Compare against domain benchmarks.** A stage below benchmark is a bottleneck candidate (B2B SaaS day-30 retention 70%+; consumer freemium paid conversion 1–5% typical, 5–10% strong).\n4. **Identify the load-bearing bottleneck.** Worst conversion *relative to its benchmark* — fixing it produces the largest system-level gain.\n5. **Pre-commit one experiment.** Name the specific change, the metric, and the pre-committed threshold (per [lean-startup](../lean-startup/SKILL.md)).\n6. **Re-measure and re-identify.** After the experiment, the bottleneck moves. Repeat — the audit is iterative.\n\n### Output: Funnel Audit\n\n```\n# AARRR Funnel Audit: \nStage definitions: Acquisition | Activation | Retention | Referral | Revenue \nConversions: [Stage | Users | Conv.% from prior | vs. benchmark ↑↓]\nBottleneck: \nExperiment: Change · Metric · Threshold · Window \n```\n\n*→ Method in Action: [Dropbox's Referral Program (2009)](examples/dropbox-referral-program-2009.md)*\n\n## Funnel Packs\n\nStage definitions and benchmarks are domain-specific. **Consumer freemium:** Activation = first core value; paid conversion 1–5% typical, 5–10% strong. **B2B SaaS:** Activation = first non-trivial team use week 1; gross retention 90%+, net 100%+ mid-market. **Marketplaces:** Activation = first transaction; bottleneck = liquidity-thin side. **E-commerce:** Activation = first order; Retention = second order within 90 days.\n\nContribution: add a pack for your domain — one file with (a) stage definitions, (b) benchmarks, (c) typical bottleneck pattern, (d) canonical experiments.\n\n## Applying It Well\n\n- **Activation is the most often-sloppily-defined stage.** \"Signed up\" is not Activation. Specify the aha moment event.\n- **The bottleneck is relative to benchmark** — not the stage with the lowest absolute number.\n- **Fixing the bottleneck moves the bottleneck.** Re-identify after each experiment.\n- **Each team optimizes the stage they own** — dangerous if the bottleneck is elsewhere. The audit forces cross-team prioritization.\n- **Domain benchmarks matter more than absolute numbers.** 3% paid conversion = great freemium, terrible enterprise SaaS.\n\n*→ Primary sources: [references/sources.md](references/sources.md)*\n\n## Common Rationalizations\n\n**[D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.**\n\n| Fake move | Reality |\n|---|---|\n| [D] **Optimizing Acquisition while Activation is broken** | Most common failure. More users into a broken funnel = more wasted CAC. Fix the bottleneck first. |\n| [D] **\"Activation\" = \"signed up\"** | Signup is end of Acquisition. Activation = first time user experiences the value. Sloppy definition hides the real funnel. |\n| [D] **Treating low absolute conversion as the bottleneck** | Acquisition always has the lowest absolute count. What matters: each stage *relative to its domain benchmark*. |\n| [D] **Adding paid ads when Retention is weak** | A leaky bucket doesn't fill faster when bigger. Fix Retention before scaling Acquisition. |\n| [D] **One-shot audit, never re-run** | The bottleneck moves once fixed. Running the audit once and never again misses the next bottleneck. |\n| *→ Add [O] entries here after each real use — paste the actual failure pattern* | *What went wrong and why* |\n\n## Red Flags\n\n- Activation is defined as \"signed up\"\n- Dashboards show absolute counts but not stage-to-stage conversion rates\n- No domain benchmarks referenced\n- Resource allocation matches which stage is easy to instrument, not which is the bottleneck\n- The funnel audit has been run once and never repeated\n- Teams don't share a bottleneck consensus\n- Acquisition spend scaled while Retention is below benchmark\n\n## Verification\n\n- [ ] Each stage has a product-specific measurable definition (Activation = aha moment, not signup)\n- [ ] Conversion rates between stages measured, not just absolute counts\n- [ ] Each stage compared against a domain benchmark\n- [ ] Load-bearing bottleneck named (worst conversion relative to benchmark)\n- [ ] Experiment pre-committed with metric + threshold + time window\n- [ ] Re-audit scheduled (bottleneck will move)\n- [ ] All teams agree on the bottleneck\n\n---\n\n*Part of **deciqAI Knowledge Skills** — open-source thinking skills that make rigor executable for AI agents. Built by deciqAI · https://deciqai.com · Contributions welcome — see the template at the repo root.*\n\nFile v1.0.1:_meta.json\n\n{\n \"ownerId\": \"kn754b8sk22s8c6gjxt02bftbn88q7ye\",\n \"slug\": \"aarrr-pirate-metrics\",\n \"version\": \"1.0.1\",\n \"publishedAt\": 1783456144812\n}\n\nFile v1.0.1:references/sources.md\n\n# Sources — aarrr-pirate-metrics\n\n> *Primary sources for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\n- **McClure, Dave.** *Startup Metrics for Pirates* (slide deck, Startonomics SF, August 2007). Original SlideShare: https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version **Canonical primary source for AARRR.**\n- **Goldratt, Eliyahu M.** *The Goal: A Process of Ongoing Improvement*. North River Press, 1984; rev. 3rd ed. 2014. **The Theory of Constraints** that underpins the bottleneck logic of AARRR.\n- **Ellis, Sean & Brown, Morgan.** *Hacking Growth: How Today's Fastest-Growing Companies Drive Breakout Success*. Crown Business, 2017. Growth-marketing case studies including the Dropbox referral case.\n- **Dropbox Inc., Form S-1 (SEC, 2018)** — primary-source user growth milestones: https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm\n- **Sequoia Capital case studies on Dropbox** — primary-source venture history: https://www.sequoiacap.com/\n- The popular framing \"growth hacking\" is **not** cited as a source — by this skill's own rule, an aphorism is not evidence. The framework is precisely \"instrument the funnel, find the bottleneck stage relative to benchmark, concentrate effort there, re-measure.\"\n\nFile v1.0.1:examples/dropbox-referral-program-2009.md\n\n# Method in Action: Dropbox's Referral Program (2009)\n\n> *Example for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\nA worked example. Not a marketing legend — primary-source documented in Drew Houston's interviews and growth-marketing case studies.\n\nBy early 2009, **Dropbox** had achieved PMF (see [pmf-crossing-the-chasm](../pmf-crossing-the-chasm/SKILL.md)) and had ~100,000 users from the 2007 video MVP + 2008 launch. The team ran an AARRR audit on the funnel:\n\n- **Acquisition**: Paid Google AdWords campaigns were running, with **customer acquisition cost (CAC) ~$200–$300 per user** — but Dropbox's price point was $99/year ($9.99/month), making **AdWords economically broken** as the dominant channel.\n- **Activation**: Strong. Once users installed the desktop client and put a file in the Dropbox folder, the experience converted to \"aha\" cleanly (the *file appeared on another device* moment, the core delight from the 2007 video MVP).\n- **Retention**: Strong. Users who activated tended to stay; Dropbox solved a real persistent need.\n- **Referral**: Untapped. Users loved the product and told friends informally, but there was no built-in mechanism.\n- **Revenue**: Decent at the free-to-paid conversion typical for freemium (a few percent).\n\nThe audit identified **Acquisition** as the load-bearing bottleneck — *not because Acquisition was bad in absolute terms, but because the acquisition channel that was working economically did not exist*. Spending more on AdWords would not solve it; only finding a sub-CAC channel would.\n\n**The intervention (September 2008–April 2009):** Dropbox launched a **referral program**. The mechanic: **give 500 MB free storage to both the referrer and the referee** when an invited friend signed up and installed Dropbox. The reward was native to the product (storage), aligned with the product's value (more space), and free for Dropbox to give away (storage was Dropbox's marginal cost, not a cash outlay).\n\nThe result, documented by Drew Houston and reproduced in Sean Ellis & Morgan Brown's *Hacking Growth*: **signups grew 60% as a result of the referral program**, with **35% of daily signups coming through referrals by late 2009**. By April 2010, Dropbox had grown to **4 million users**, then **25 million** by 2011. The referral program effectively **converted a Retention strength into Acquisition output** — Goldratt's \"exploit the constraint\" move applied to a customer funnel.\n\nWalk the audit on Dropbox 2009:\n\n- **Stage definitions (Step 1):** Acquisition = ad click → landing page → signup; Activation = client installed + first file synced; Retention = active in day 7; Referral = invite sent → friend installs + syncs file; Revenue = upgrade from free to paid tier.\n- **Conversions (Step 2):** Acquisition CAC vs LTV economically broken; Activation, Retention, Revenue all benchmarks-fine; Referral nominal.\n- **Bottleneck (Step 4):** Acquisition channel economics — *specifically*, the cost-effective channel did not exist.\n- **Experiment (Step 5):** Build referral program; pre-committed threshold = signup growth ≥ 30% from referral source within 6 months.\n- **Result:** 60% growth; threshold cleared by ~2×. *Persevere*; double down on the channel; iterate on referral mechanics.\n\n**The non-obvious lesson:** Dropbox's \"referral was the bottleneck\" framing would have been wrong. The bottleneck was Acquisition; the *intervention* was building a referral mechanism *because* Retention was strong enough that activated users would actually send referrals. **The referral program was the solution to the Acquisition bottleneck**, not a Referral-stage improvement in isolation.\n\n**Sources:** McClure, Dave. ***Startup Metrics for Pirates*** (slide deck, Startonomics SF, August 2007). Original SlideShare: https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version . **Canonical primary source for AARRR**. Goldratt, Eliyahu M. ***The Goal: A Process of Ongoing Improvement***. North River Press, 1984; rev. 3rd ed. 2014. **Theory of Constraints**, the bottleneck logic underpinning AARRR. Ellis, Sean & Brown, Morgan. ***Hacking Growth: How Today's Fastest-Growing Companies Drive Breakout Success***. Crown Business, 2017, ch. 5 — primary growth-marketing case account of the Dropbox referral program. Houston, Drew. Drew Houston's interviews on growth, including talks at Stanford and the Sequoia Capital growth case study at: https://www.sequoiacap.com/ Dropbox SEC S-1 filing (2018) provided official user-growth milestones: https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm\n\nFile v1.0.1:skill-card.md\n\n## Description:
\nGuides teams through an AARRR funnel audit to define stage metrics, compare conversion rates against benchmarks, identify the growth bottleneck, and choose the next experiment.
\n\nThis skill is ready for commercial/non-commercial use.
\n\n## Publisher:
\n[deciqai](https://clawhub.ai/user/deciqai)
\n\n### License/Terms of Use:
\nMIT-0
\n\n\n## Use Case:
\nProduct, growth, marketing, and sales teams use this skill when growth is stalling, funnel ownership is disputed, or a product needs shared end-to-end growth instrumentation. The skill helps define AARRR stage events, measure stage-to-stage conversion, identify the bottleneck relative to benchmarks, and select a focused experiment.
\n\n### Deployment Geography for Use:
\nGlobal
\n\n## Known Risks and Mitigations:
\nRisk: Company funnel numbers, revenue metrics, and growth plans shared with the agent may be sensitive business information.
\nMitigation: Use anonymized or aggregated metrics when possible and avoid sharing confidential raw data unless the deployment environment is approved for that information.
\nRisk: Funnel recommendations may be misleading when stage definitions, conversion rates, or benchmark comparisons are incomplete or inaccurate.
\nMitigation: Review the resulting funnel audit with product and growth stakeholders before acting on the recommended experiment.
\n\n\n## Reference(s):
\n- [Primary Sources](references/sources.md)
\n- [Dropbox Referral Program Example](examples/dropbox-referral-program-2009.md)
\n- [Startup Metrics for Pirates](https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version)
\n- [Dropbox Form S-1](https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm)
\n- [Sequoia Capital Dropbox Case Studies](https://www.sequoiacap.com/)
\n\n\n## Skill Output:
\n**Output Type(s):** [Analysis, Guidance, Markdown]
\n**Output Format:** [Markdown]
\n**Output Parameters:** [1D]
\n**Other Properties Related to Output:** [Produces a funnel audit with stage definitions, conversions, bottleneck, and experiment recommendation.]
\n\n## Skill Version(s):
\n1.0.1 (source: server release evidence)
\n\n## Ethical Considerations:
\nUsers should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.
\n\nArchive v1.0.0: 5 files, 8870 bytes\n\nFiles: examples/dropbox-referral-program-2009.md (4627b), references/sources.md (1282b), skill-card.md (2746b), SKILL.md (8444b), _meta.json (139b)\n\nFile v1.0.0:SKILL.md\n\n---\nname: aarrr-pirate-metrics\ndescription: \"Activate when: user says 'our growth is stalling and we don't know why,' 'product and marketing are arguing about whose fault it is,' 'where is our funnel breaking,' 'pirate metrics,' or 'acquisition vs activation vs retention.' Also activate for: designing end-to-end metrics for a new product, building a shared instrumentation framework across teams.\n Do NOT activate when: the product has no users yet (use lean-startup instead), or the bottleneck is already obvious and obvious to fix.\"\n---\n\n# AARRR (Pirate Metrics)\n\n## Overview\n\nA startup's growth is a **sequential funnel** — each stage gates the next. Great Acquisition is worthless if Activation is broken; great Activation is worthless if Retention is zero. Optimizing the wrong stage produces work that looks like progress while the bottleneck stays.\n\nThe **AARRR framework** (Dave McClure, *Startup Metrics for Pirates*, 2007) names five stages: Acquisition (do they show up?), Activation (good first experience?), Retention (do they come back?), Referral (do they tell others?), Revenue (do they pay?). The bottleneck stage governs total growth — improving any other stage produces no system-level gain (Goldratt, *The Goal*, 1984).\n\n**Compose with:** [first-principles](../first-principles/SKILL.md) to identify your specific Activation event; [pmf-crossing-the-chasm](../pmf-crossing-the-chasm/SKILL.md) to recognize when Retention will always be the bottleneck pre-PMF; [probabilistic-thinking](../probabilistic-thinking/SKILL.md) to set base rates per stage.\n\n## When to Use\n\n- Growth is **stalling** and the cause is not obvious — multiple teams have plausible explanations\n- Product, marketing, and sales are **arguing about whose problem it is**\n- You need a **shared instrumentation framework** across teams\n- Designing **end-to-end metrics for a new product**\n- Someone says: *\"AARRR,\" \"pirate metrics,\" \"growth funnel,\" \"where is our funnel breaking?\"*\n\n**When NOT to use:** No users yet → use [lean-startup](../lean-startup/SKILL.md). Signups still in the tens — AARRR rates need volume. Bottleneck already obvious → fix that first, then return.\n\n## Coaching Novices (Adaptive Front Door)\n\n- **Engine mode:** user has funnel data and wants the bottleneck named → run The Process directly.\n- **Coach mode:** vague situation or unfamiliarity → guide step by step.\n\nIn Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.\n\n1. **One-line what-it-is.** AARRR is a five-stage","readmeExcerpt":"Skill: AARRR (Pirate Metrics) Owner: deciqai Summary: Activate when: user says 'our growth is stalling and we don't know why,' 'product and marketing are arguing about whose fault it is,' 'where is our funnel br... Tags: latest:1.0.6 Version history: v1.0.6 | 2026-07-16T17:51:05.759Z | user Description tail link + agents machine-readable metadata line (deciqai.com/s/aarrr-pirate-metrics.json) v1.0.5 | 2026-07-09T11:1","codeSnippets":[],"executableExamples":[{"language":"text","snippet":"# AARRR Funnel Audit: \nStage definitions: Acquisition | Activation | Retention | Referral | Revenue \nConversions: [Stage | Users | Conv.% from prior | vs. benchmark ↑↓]\nBottleneck: \nExperiment: Change · Metric · Threshold · Window "},{"language":"text","snippet":"# AARRR Funnel Audit: \nStage definitions: Acquisition | Activation | Retention | Referral | Revenue \nConversions: [Stage | Users | Conv.% from prior | vs. benchmark ↑↓]\nBottleneck: \nExperiment: Change · Metric · Threshold · Window "},{"language":"text","snippet":"# AARRR Funnel Audit: \nStage definitions: Acquisition | Activation | Retention | Referral | Revenue \nConversions: [Stage | Users | Conv.% from prior | vs. benchmark ↑↓]\nBottleneck: \nExperiment: Change · Metric · Threshold · Window "},{"language":"text","snippet":"# AARRR Funnel Audit: \nStage definitions: Acquisition | Activation | Retention | Referral | Revenue \nConversions: [Stage | Users | Conv.% from prior | vs. benchmark ↑↓]\nBottleneck: \nExperiment: Change · Metric · Threshold · Window "},{"language":"text","snippet":"# AARRR Funnel Audit: \nStage definitions: Acquisition | Activation | Retention | Referral | Revenue \nConversions: [Stage | Users | Conv.% from prior | vs. benchmark ↑↓]\nBottleneck: \nExperiment: Change · Metric · Threshold · Window "},{"language":"text","snippet":"# AARRR Funnel Audit: \nStage definitions: Acquisition | Activation | Retention | Referral | Revenue \nConversions: [Stage | Users | Conv.% from prior | vs. benchmark ↑↓]\nBottleneck: \nExperiment: Change · Metric · Threshold · Window "}],"parameters":null,"dependencies":[],"permissions":[],"extractedFiles":[{"path":"SKILL.md","content":"---\nname: aarrr-pirate-metrics\ndescription: \"Activate when: user says 'our growth is stalling and we don't know why,' 'product and marketing are arguing about whose fault it is,' 'where is our funnel breaking,' 'pirate metrics,' or 'acquisition vs activation vs retention.' Also activate for: designing end-to-end metrics for a new product, building a shared instrumentation framework across teams.\n Do NOT activate when: the product has no users yet (use lean-startup instead), or the bottleneck is already obvious and obvious to fix. More: deciqai.com/c/aarrr-pirate-metrics\"\n---\n\n# AARRR (Pirate Metrics)\n\n## Overview\n\nA startup's growth is a **sequential funnel** — each stage gates the next. Great Acquisition is worthless if Activation is broken; great Activation is worthless if Retention is zero. Optimizing the wrong stage produces work that looks like progress while the bottleneck stays.\n\nThe **AARRR framework** (Dave McClure, *Startup Metrics for Pirates*, 2007) names five stages: Acquisition (do they show up?), Activation (good first experience?), Retention (do they come back?), Referral (do they tell others?), Revenue (do they pay?). The bottleneck stage governs total growth — improving any other stage produces no system-level gain (Goldratt, *The Goal*, 1984).\n\n**Compose with:** first-principles to identify your specific Activation event; pmf-crossing-the-chasm to recognize when Retention will always be the bottleneck pre-PMF; probabilistic-thinking to set base rates per stage.\n\n## When to Use\n\n- Growth is **stalling** and the cause is not obvious — multiple teams have plausible explanations\n- Product, marketing, and sales are **arguing about whose problem it is**\n- You need a **shared instrumentation framework** across teams\n- Designing **end-to-end metrics for a new product**\n- Someone says: *\"AARRR,\" \"pirate metrics,\" \"growth funnel,\" \"where is our funnel breaking?\"*\n- An **AI-native product has strong signups but weak activation/retention** — cheap AI-hype acquisition masks empty-state and post-novelty leaks; or you're facing AI-native competition on a commoditized model and need to find where your funnel actually loses users\n\n**When NOT to use:** No users yet → use lean-startup. Signups still in the tens — AARRR rates need volume. Bottleneck already obvious → fix that first, then return.\n\n## Coaching Novices (Adaptive Front Door)\n\n- **Engine mode:** user has funnel data and wants the bottleneck named → run The Process directly.\n- **Coach mode:** vague situation or unfamiliarity → guide step by step.\n\nIn Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output only that step's question, then stop.\n\n1. **One-line what-it-is.** AARRR is a five-stage funnel (Acquisition / Activation / Retention / Referral / Revenue) — the point is to find which stage is broken so you don't waste effort on the wrong one.\n2. **Check fit.** No users / tiny base → redirect. Bottleneck already obvious → fix it first.\n3. **Elicit their real case."},{"path":"_meta.json","content":"{\n \"ownerId\": \"kn754b8sk22s8c6gjxt02bftbn88q7ye\",\n \"slug\": \"aarrr-pirate-metrics\",\n \"version\": \"1.0.6\",\n \"publishedAt\": 1784224265759\n}"},{"path":"references/sources.md","content":"# Sources — aarrr-pirate-metrics\n\n> *Primary sources for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\n- **McClure, Dave.** *Startup Metrics for Pirates* (AARRR). The framework was first presented in 2007 (Ignite Seattle); the widely-circulated \"long version\" slide deck linked here is the Startonomics SF edition (October 2008). SlideShare: https://www.slideshare.net/dmc500hats/startup-metrics-for-pirates-long-version **Canonical primary source for AARRR.**\n- **Goldratt, Eliyahu M.** *The Goal: A Process of Ongoing Improvement*. North River Press, 1984; rev. 3rd ed. 2014. **The Theory of Constraints** that underpins the bottleneck logic of AARRR.\n- **Ellis, Sean & Brown, Morgan.** *Hacking Growth: How Today's Fastest-Growing Companies Drive Breakout Success*. Crown Business, 2017. Growth-marketing case studies including the Dropbox referral case and Facebook's \"7 friends in 10 days\" activation metric.\n- **Palihapitiya, Chamath.** \"How We Put Facebook on the Path to 1 Billion Users\" (public talk, c. 2012; widely circulated 2013). First-person account of Facebook's growth team and the \"7 friends in 10 days\" activation metric. Note: the \"7 friends in 10 days\" figure is a widely-repeated growth-team anecdote rather than a formally published metric; treat as illustrative.\n- **Dropbox Inc., Form S-1 (SEC, 2018)** — primary-source user growth milestones: https://www.sec.gov/Archives/edgar/data/1467623/000119312518055809/d451946ds1.htm\n- **Sequoia Capital case studies on Dropbox** — primary-source venture history: https://www.sequoiacap.com/\n- **OpenAI.** \"Introducing ChatGPT\" (November 30, 2022): https://openai.com/blog/chatgpt — dates the acquisition-side inflection for the 2023–2026 AI SaaS funnel example (AI-hype traffic makes acquisition cheap).\n- **Sequoia Capital / Cahn, David.** \"AI's $600B Question\" (2024): https://www.sequoiacap.com/article/ais-600b-question/ — widely-cited analysis of the gap between AI infrastructure/capex build-out and durable end-application revenue and retention; context for why activation/retention (not acquisition) is the AI-SaaS bottleneck.\n- The popular framing \"growth hacking\" is **not** cited as a source — by this skill's own rule, an aphorism is not evidence. The framework is precisely \"instrument the funnel, find the bottleneck stage relative to benchmark, concentrate effort there, re-measure.\""},{"path":"examples/ai-saas-funnel-leaks-2023-2026.md","content":"# Method in Action: Where AI SaaS Funnels Leak (2023–2026)\n\n> *Example for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\nA pattern-level worked example — not a single company, but the recurring funnel shape observed across the wave of AI-native SaaS products that launched after ChatGPT's release in November 2022. The generative-AI boom pushed a distinctive failure mode to the surface: acquisition became historically *easy* while activation and retention became the real leaks. This example applies the Funnel Audit to that pattern. Where a specific number would be fragile, it is qualified or omitted.\n\nThe context in one line (drawn from the 2023–2025 record; the pattern appears to persist but recent periods should be re-checked against current data): after the November 2022 ChatGPT launch, curiosity and press drove enormous top-of-funnel traffic to anything labeled \"AI,\" and by early 2025 the AI-native SaaS category was crowded with near-identical wrapper products competing on the same underlying foundation models. In that environment, *showing up* is cheap and *staying* is hard — the inverse of the classic pre-AI SaaS funnel, where distribution was the scarce resource.\n\nWalk the Funnel Audit (matching the SKILL.md Process steps):\n\n**Step 1 — Define each stage for this specific product.** For a generic AI SaaS assistant:\n- **Acquisition** = visitor lands on the site and creates an account (often driven by AI-hype PR, launch-day virality, or a viral demo).\n- **Activation** = the user reaches first real value — i.e., they get past the **empty state** and complete one genuinely useful task with the AI (a good draft, a correct answer on *their own* data, a workflow actually finished), not merely \"typed one prompt.\"\n- **Retention** = the user returns and uses the product in a later week for real work, *after the novelty of trying an AI toy has worn off*.\n- **Referral** = the user shares an output or invites a colleague.\n- **Revenue** = the user converts from free trial to a paid subscription.\n\nThe single most common definitional error here is calling \"signed up and sent one message\" *Activation*. That is the end of Acquisition. AI products are unusually vulnerable to this because a first prompt is trivially easy, so the vanity signal looks great while no durable value has been delivered.\n\n**Step 2 — Measure conversion at each stage.** The characteristic AI-SaaS reading:\n- Acquisition→Activation: **leaky.** Many curiosity-driven signups hit a blank chat box or empty canvas, don't know what to type, get a generic or hallucinated first result, and never reach first value. This is the **empty-state / first-value** leak.\n- Activation→Retention: **the deepest leak.** Even users who got one impressive result often don't come back, because the initial \"wow\" was novelty, not a solved recurring job. This is the **novelty-decay** leak.\n- Retention→Referral and →Revenue: mostly downstream symptoms — thin because too few users retained.\n\n**Step 3 — Compare agains"},{"path":"examples/dropbox-referral-program-2009.md","content":"# Method in Action: Dropbox's Referral Program (2009)\n\n> *Example for the [aarrr-pirate-metrics](../SKILL.md) skill.*\n\nA worked example. Not a marketing legend — primary-source documented in Drew Houston's interviews and growth-marketing case studies.\n\nBy early 2009, **Dropbox** had achieved PMF (see [pmf-crossing-the-chasm](../pmf-crossing-the-chasm/SKILL.md)) and had ~100,000 users from the 2007 video MVP + 2008 launch. The team ran an AARRR audit on the funnel:\n\n- **Acquisition**: Paid Google AdWords campaigns were running, with **customer acquisition cost (CAC) ~$200–$300 per user** — but Dropbox's price point was $99/year ($9.99/month), making **AdWords economically broken** as the dominant channel.\n- **Activation**: Strong. Once users installed the desktop client and put a file in the Dropbox folder, the experience converted to \"aha\" cleanly (the *file appeared on another device* moment, the core delight from the 2007 video MVP).\n- **Retention**: Strong. Users who activated tended to stay; Dropbox solved a real persistent need.\n- **Referral**: Untapped. Users loved the product and told friends informally, but there was no built-in mechanism.\n- **Revenue**: Decent at the free-to-paid conversion typical for freemium (a few percent).\n\nThe audit identified **Acquisition** as the load-bearing bottleneck — *not because Acquisition was bad in absolute terms, but because the acquisition channel that was working economically did not exist*. Spending more on AdWords would not solve it; only finding a sub-CAC channel would.\n\n**The intervention (September 2008–April 2009):** Dropbox launched a **referral program**. The mechanic: **give 500 MB free storage to both the referrer and the referee** when an invited friend signed up and installed Dropbox. The reward was native to the product (storage), aligned with the product's value (more space), and free for Dropbox to give away (storage was Dropbox's marginal cost, not a cash outlay).\n\nThe result, documented by Drew Houston and reproduced in Sean Ellis & Morgan Brown's *Hacking Growth*: **signups grew 60% as a result of the referral program**, with **35% of daily signups coming through referrals by late 2009**. By April 2010, Dropbox had grown to **4 million users**, then **25 million** by 2011. The referral program effectively **converted a Retention strength into Acquisition output** — Goldratt's \"exploit the constraint\" move applied to a customer funnel.\n\nWalk the audit on Dropbox 2009:\n\n- **Stage definitions (Step 1):** Acquisition = ad click → landing page → signup; Activation = client installed + first file synced; Retention = active in day 7; Referral = invite sent → friend installs + syncs file; Revenue = upgrade from free to paid tier.\n- **Conversions (Step 2):** Acquisition CAC vs LTV economically broken; Activation, Retention, Revenue all benchmarks-fine; Referral nominal.\n- **Bottleneck (Step 4):** Acquisition channel economics — *specifically*, the cost-effective channel did not exist.\n- **Ex"}],"languages":[],"docsSourceLabel":"CLAWHUB","editorialOverview":"Activate when: user says 'our growth is stalling and we don't know why,' 'product and marketing are arguing about whose fault it is,' 'where is our funnel br... Skill: AARRR (Pirate Metrics) Owner: deciqai Summary: Activate when: user says 'our growth is stalling and we don't know why,' 'product and marketing are arguing about whose fault it is,' 'where is our funnel br... Tags: latest:1.0.6 Version history: v1.0.6 | 2026-07-16T17:51:05.759Z | user Description tail link + agents machine-readable metadata line (deciqai.com/s/aarrr-pirate-metrics.json) v1.0.5 | 2026-07-09T11:1","editorialQuality":{"score":100,"threshold":65,"status":"ready","wordCount":2133,"uniquenessScore":48,"reasons":[]}},"media":{"evidence":{"source":"no-media","verified":false,"confidence":"low","updatedAt":"2026-10-10T19:44:42.349Z","emptyReason":"No screenshots, media assets, or demo links are available."},"primaryImageUrl":null,"mediaAssetCount":0,"assets":[],"demoUrl":null},"ownerResources":{"evidence":{"source":"unclaimed","verified":false,"confidence":"low","updatedAt":"2026-10-10T19:44:42.349Z","emptyReason":"This page has not been claimed by the agent owner."},"hasCustomPage":false,"customPageUpdatedAt":null,"customLinks":[],"structuredLinks":{"docsUrl":null,"demoUrl":null,"supportUrl":null,"pricingUrl":null,"statusUrl":null},"customPage":null},"relatedAgents":{"evidence":{"source":"protocol-neighbors","verified":false,"confidence":"medium","updatedAt":"2026-10-10T23:48:30.544Z","emptyReason":null},"items":[{"id":"8ebccd8e-3863-4187-8355-c3f14e1f9edf","entityType":"agent","canonicalPath":"/agent/iofficeai-aionui","slug":"iofficeai-aionui","name":"AionUi","description":"Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!","url":"https://github.com/iOfficeAI/AionUi","homepage":"https://www.aionui.com","source":"GITHUB_REPOS","protocols":["MCP","OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-10-09T19:11:12.944Z","createdAt":"2026-02-25T03:38:16.584Z","downloads":null},{"id":"b917f68a-ebff-438e-84f8-3f4b2494c0bc","entityType":"agent","canonicalPath":"/agent/activepieces-activepieces","slug":"activepieces-activepieces","name":"activepieces","description":"AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents","url":"https://github.com/activepieces/activepieces","homepage":"https://www.activepieces.com","source":"GITHUB_REPOS","protocols":["OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-04-15T02:22:12.426Z","createdAt":"2026-02-25T03:38:12.412Z","downloads":null},{"id":"5cb26759-3a39-483f-94cf-276a98c13bb8","entityType":"agent","canonicalPath":"/agent/cherryhq-cherry-studio","slug":"cherryhq-cherry-studio","name":"cherry-studio","description":"AI productivity studio with smart chat, autonomous agents, and 300+ assistants. Unified access to frontier LLMs","url":"https://github.com/CherryHQ/cherry-studio","homepage":"https://cherry-ai.com","source":"GITHUB_REPOS","protocols":["MCP","OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-04-11T14:38:40.986Z","createdAt":"2026-02-25T03:38:19.379Z","downloads":null},{"id":"6f6582d0-5d76-4f0f-b81d-86520247950b","entityType":"agent","canonicalPath":"/agent/copilotkit-copilotkit","slug":"copilotkit-copilotkit","name":"CopilotKit","description":"The Frontend for Agents & Generative UI. React + Angular","url":"https://github.com/CopilotKit/CopilotKit","homepage":"https://docs.copilotkit.ai","source":"GITHUB_REPOS","protocols":["OPENCLAW"],"capabilities":[],"safetyScore":100,"overallRank":70,"updatedAt":"2026-03-25T09:50:57.846Z","createdAt":"2026-02-25T03:39:14.617Z","downloads":null}],"links":{"hub":"/agent","source":"/agent/source/clawhub","protocols":[{"label":"OpenClaw","href":"/agent/protocol/openclew"}]}}}