growth-hacker（增长黑客）

Growth Hacker

Expert in growth hacking: playbooks, viral loops, acquisition, funnel optimization, retention, competitor intel, personas, and growth audits for startups and SMEs.

When to Use

Apply this skill when the user mentions or asks for: growth playbook, viral loop, referral, acquisition channels, funnel optimization, retention, churn, competitor analysis, personas, content strategy, SEO for growth, growth ideas, growth audit, AARRR, North Star metric, launch strategy, growth experiments, launch execution, product-led growth, PLG, time-to-value, TTV, PQL, product-qualified lead, signup flow, onboarding, activation, aha moment, paywall, free-to-paid, freemium conversion, expansion revenue, self-serve, set up email/ads/analytics/payments, run campaigns, post to Twitter/social, or MCP for growth/launch tasks.

Initial Context

Check for existing context: If .claude/product-marketing-context.md or a business-context file exists, use it before asking.

Otherwise, gather (or infer):

• Company: name, stage (pre-seed, seed, Series A+), industry, business model (saas, marketplace, ecommerce, subscription, freemium), team size, monthly budget.
• Product: name, type (b2b, b2c, b2b2c), category, pricing, value prop, core problem, key features.
• Metrics (if any): MAU, activation rate, D1/D7/D30 retention, churn, CAC, LTV, signups, conversion by stage.
• Goals: acquisition, activation, retention, revenue, referral, or full audit.

Frameworks

• AARRR (Pirate Metrics): Acquisition → Activation → Retention → Revenue → Referral. Prioritize by where the biggest leak or opportunity is.
• ICE: Impact × Confidence × Ease. Use to rank experiments and ideas.
• North Star Metric: One metric that best reflects value delivered. Align tactics to move it.
• Growth Loops: Viral, content, paid, sales. Reinforcing loops beat one-off campaigns.
• Hook Model: Trigger → Action → Variable Reward → Investment. For retention and habit.

MCPs & Execution

When the user wants to execute (not just plan)—set up email, run ads, configure analytics, create Stripe products, post to social—use any available MCP tools from the environment. MCP access is provided by Cursor, Claude Code, or your agent’s MCP config; this skill does not grant MCP access, it directs you to use it when relevant.

Relevant MCPs

When to use which

• Launch setup (email): resend-mcp → domain + verify, then send_email for welcome/activation sequences.
• Launch setup (ads): meta-ads-mcp and/or google-ads-mcp → pixel/audiences first, then campaign/ad set/ad.
• Launch setup (analytics): posthog-mcp → capture/identify for key events; create_action for activation; query_insights for funnels/retention.
• Launch setup (payments): stripe-mcp → create_product, create_price, create_checkout_session for paywall/upgrade.
• Distribution / social: twitter-mcp → post_tweet or post_thread for launch and follow-up.

If an MCP is not available, deliver the plan and concrete commands/snippets (e.g. from MCPLaunchIntegrations or Launch Assistant patterns) so the user can run them elsewhere or after adding the MCP.

1. Growth Playbook

Inputs: target stage (acquisition, activation, retention, revenue, referral), business model, main challenge, budget (bootstrap, seed, funded).

Output structure:

# [Stage] Playbook: [Name]

## Goal
[One sentence]

## Steps (ordered)
1. **Step name** — What to do, why, and how to measure.
2. ...

## Tactics per step
- Tactic 1 (effort: low/med/high)
- Tactic 2
- ...

## Expected outcomes
- Metric/behavior change and rough timeline

## Resources / tools
- [Concrete tools or templates]

## Risks & mitigations
- Risk → mitigation

Stage-specific focus:

• Acquisition: channels, messaging, landing, signup. CAC and volume.
• Activation: first value, onboarding, aha moment, time-to-value.
• Retention: cohort curves, hooks, habit, win-backs, churn reasons.
• Revenue: pricing, packaging, paywall, upgrade paths, expansion.
• Referral: incentives, mechanics, K-factor, sharing triggers.

2. Viral Loop Design

Inputs: product type (b2b, b2c, b2b2c), current K-factor if known, preferred type, constraints.

Viral loop types:

• Word-of-mouth — Stories, NPS, case studies. Good for b2b and high-touch.
• Inherent virality — Collaboration, invites, shared workspaces. Product-native.
• Incentivized referral — Reward for invite + signup. Simple, needs unit economics.
• Content/viral — Shareable output (calcs, exports, UGC). Good for b2c and tools.

Output:

• Recommended loop(s) with mechanism and triggers.
• K-factor target and how to measure.
• Implementation steps: where in product, messaging, incentives, tracking.
• Integration with existing acquisition and retention.

3. Acquisition Channel Analysis

Inputs: industry, product type (b2b/b2c/b2b2c), monthly budget, current channels, target CAC.

Channels to evaluate: organic-search, paid-search, social-organic, paid-social, content/SEO, partnerships, community, events, outbound/sales, referral, product-led, PR.

Per channel, assess:

• Fit to industry and product type (score 1–10).
• Time to results: immediate, short, medium, long.
• Scalability and typical CAC (b2b vs b2c).
• Difficulty and budget feasibility.

Output:

• Ranked channel list with fit, priority, and 2–3 concrete tactics each.
• Recommended channel mix (e.g. 50% content, 30% paid, 20% community) and rationale.
• Implementation order and quick wins.

4. Funnel Optimization

Inputs: funnel stages with conversion rates, primary goal (signups, activation, conversion, retention), current metrics.

Process:

1. Map stages (e.g. Visit → Signup → Activate → First value → Pay).
2. Compute conversion per step and identify largest drop-offs.
3. For each bottleneck: cause, impact, severity (critical/high/medium/low).
4. Propose solutions (copy, UX, targeting, timing) with ICE-style prioritization.
5. Suggest metrics and simple experiments (A/B or before/after).

Output:

• Funnel viz (ascii or list) with conversion %.
• Bottlenecks table: stage, cause, impact, severity, solutions.
• Top 3–5 experiments to run first.

5. Retention Strategy

Inputs: retention metrics (D1, D7, D30, churn), product type, segments, known churn reasons.

Process:

1. Compare to benchmarks: D1 >40%, D7 >20%, D30 >10%; churn <5%/mo for SaaS.
2. Find drop-off points (onboarding, first value, first week, first month).
3. Link to reasons: unclear value, friction, missing habit, wrong segment, product gaps.
4. Propose plays: onboarding, email/lifecycle, in-app hooks, win-back, feature/segment tweaks.

Output:

• Retention view and benchmarks.
• Root causes and prioritized actions.
• Retention playbook (steps, triggers, messaging, metrics).

6. Competitor Intelligence

Inputs: competitor names/sites, depth (quick, standard, deep), focus (pricing, features, marketing, positioning).

Assess:

• Positioning, messaging, and differentiators.
• Features, pricing, packaging.
• Marketing and channel presence.
• Gaps and opportunities (positioning, feature, pricing, segment, content).
• Threats and possible responses.

Output:

• Summary per competitor.
• Opportunity vs threat matrix.
• Recommended differentiators and moves.

7. User Personas

Inputs: product description, target market, any user data, number of personas (default 3).

Per persona:

• Name, role, goals, pain points.
• Demographics and behavior (where they are, how they decide).
• Preferred channels and influencers.
• Objections and how the product addresses them.
• Quotes and use cases.

Output:

• 2–4 personas in a consistent template.
• Implications for messaging, channels, and product.

8. Growth Metrics Analysis

Inputs: current metrics, optional benchmarks, timeframe.

Process:

1. Define North Star and supporting metrics.
2. Check health: trends, segment performance, funnel, retention.
3. Compare to benchmarks where possible.
4. Call out anomalies and likely causes.
5. Suggest next metrics to add or refine.

Output:

• Metric review and trend comments.
• Issues and hypotheses.
• 3–5 recommended actions or experiments.

9. Content & SEO Strategy

Inputs: industry, target audience, goals, existing content, competitors.

Deliver:

• Topics (pillar + clusters) aligned to intent and keywords.
• Content types (blog, guides, tools, comparison, G2/Capterra, etc.).
• SEO: keywords, on-page, internal linking, technical basics.
• Distribution: organic, social, email, partnerships.
• Cadence and quick wins.

Output:

• Content pillars and 10–20 topic ideas.
• SEO and distribution checklist.
• 90-day plan outline.

10. Growth Ideas / Experiments

Inputs: business context, constraints, previous experiments, risk tolerance (conservative, moderate, aggressive).

Process:

1. Generate quick wins (low effort, fast learning).
2. Medium-term plays (new channels, loops, segments).
3. Moonshots (high impact, higher risk).
4. Score with ICE and filter by risk and resources.

Output:

• 5–10 ideas per bucket with hypothesis, method, and success metric.
• Top 3–5 to run next.

11. Product-Led Growth (PLG)

Inputs: product type (b2b, b2c, b2b2c), business model (freemium, free trial, product-led sales), current signup/activation/free-to-paid rates, North Star (if any).

Product-Led Growth means the product itself is the primary driver of acquisition, activation, conversion, and expansion. Self-serve and low-touch beat high-touch sales for many SaaS and tools.

PLG fundamentals

• PQL (Product-Qualified Lead): A user who has experienced enough value in-product to be a strong sales or upgrade candidate. Define the in-app behavior that signals intent.
• Time-to-Value (TTV): Minutes or actions from signup to “aha moment.” Shorter TTV = higher activation and conversion.
• Self-serve funnel: Try → Activate → Convert → Expand. Each step should be measurable and improvable in-product.

Signup flow (first touch)

• Minimize required fields: Email + password (or social) first. Defer name, company, role to onboarding or progressive profiling.
• Value before ask: Let users see or try value before signup when possible (demos, calculators, limited use).
• Reduce perceived effort: Progress indicators, smart defaults, inline validation, clear “what happens next.”
• Social / SSO: Prominent Google, Apple, Microsoft, or GitHub; often converts better than email-only.

Onboarding & activation

• Aha moment: The action that correlates most with retention. Find it via cohort analysis; design the flow to reach it in the first session.
• One goal per session: Get one clear win in the first use. Save advanced features for later.
• Do, don’t show: Interactive > tutorial. Empty states that invite “add your first X” beat long tours.
• Onboarding checklist: 3–7 items, ordered by impact, with progress and a dismiss option. Don’t trap.
• Email + in-app: Welcome, incomplete-onboarding, and activation-achieved emails that drive back into the product with a specific CTA.

Free-to-paid conversion & paywalls

• Value before ask: Show the upgrade only after the user has felt value (post–aha moment or when hitting a real limit).
• Trigger points: Feature gates (clicking a paid feature), usage limits (projects, exports, seats), trial expiration, or time-based (e.g. after 7 days of use).
• Paywall copy: Headline on benefit (“Unlock X to get Y”), short value demo, clear plan comparison, specific CTA, and a respectful “Not now” or “Continue with Free.”
• Timing: Not during onboarding; limit frequency per session; cool-down after dismiss (days, not hours).

Expansion revenue

• Usage-based: More usage → higher plan or overage. Align pricing with value.
• Seat expansion: Encourage team invites and team plans; make “add teammate” obvious at the right moment.
• Land-and-expand: Single user → team → org. Track expansion MRR and time-to-expand.

Free tools (try-before-buy / lead gen)

• When it fits PLG: Calculators, generators, analyzers, or limited-use versions that mirror the core product. Tool = lead and first value.
• Gating: Fully gated, partial (preview + email for full), or ungated for reach. Balance capture vs. usage.
• Path to product: Clear next step from tool to full product or trial.

PLG audit output

For a PLG audit, run through:

1. Signup: Friction, fields, social auth, post-submit flow.
2. TTV & activation: Aha moment defined? Steps to reach it? Activation rate and time-to-activation.
3. Onboarding: First session flow, checklist, empty states, email triggers.
4. Free-to-paid: Triggers, placement, copy, conversion rate and where it drops.
5. Expansion: Usage-based or seat-based plays; expansion MRR; time-to-expand.

Output:

• PLG funnel (Signup → Activate → Convert → Expand) with current rates and benchmarks.
• PQL definition (behavioral criteria) and how to surface PQLs to sales or in-app upgrade.
• Findings table: area, issue, impact, recommendation, priority.
• Top 3–5 experiments (e.g. reduce signup fields, reorder onboarding, new paywall timing, expansion prompt).

12. Full Growth Audit

Inputs: full business context (company, product, market, metrics, personas, competitors, objectives).

Process:

1. Run through: playbooks (prioritized stages), viral potential, channels, funnel, retention, PLG (signup, TTV, activation, free-to-paid, expansion), competitors, personas, metrics, content.
2. Synthesize insights (stage-specific, metrics-based, channel, viral, PLG).
3. Produce recommendations with priority (critical/high/medium/low), action, rationale, expected impact, effort.

Output:

• Executive summary (3–5 insights).
• Prioritized recommendations table.
• Optional deeper sections per area.

13. Launch Execution (with MCPs)

Inputs: app/product name, tagline, URL, category, target audience, pricing model, launch date, budget. Optionally: .claude/product-marketing-context.md or business-context.

When the user wants to execute (e.g. “set up our launch,” “create the welcome email,” “create a Meta campaign,” “add PostHog events,” “create Stripe products,” “post our launch on Twitter”):

1. Gather context — app name, URL, pricing, audience. Reuse product-marketing-context if present.
2. Pick the right MCPs from the table in MCPs & Execution (resend, meta-ads, google-ads, posthog, stripe, twitter).
3. Call the MCP tools to perform the task (e.g. create_domain + send_email, create_campaign + create_ad_set + create_ad, create_product + create_price, post_tweet).
4. If an MCP is missing — output a ready-to-run snippet or command block so the user can run it after configuring that MCP.

Typical launch flow: email domain + templates → pixel + audiences → first campaign (paused until launch) → PostHog events/actions → Stripe products/prices → launch-day Twitter post. Run only the steps the user asked for, unless they request a “full launch setup.”

Output: confirm what was created (IDs, URLs) and any follow-up (DNS records, env vars, “activate campaign on launch day”).

Output Conventions

• Concise first: lead with the 3–5 most important points, then detail.
• Actionable: every recommendation = clear next step and how to measure.
• Prioritized: critical/high first; use ICE when comparing experiments.
• Structured: use headers, lists, and tables so the user can skim and share.

Programmatic Use

For scripted or agent use, the gaasai-growth-hacker-skill package provides the same capabilities via TypeScript:

npm install gaasai-growth-hacker-skill

import growthHackerSkill, { BusinessContext } from 'gaasai-growth-hacker-skill';

const result = await growthHackerSkill.execute({ sessionId: 'x', businessContext });
// result.insights, result.nextActions, result.data

Use this skill for interactive guidance in chat; use the package when you need structured, repeatable runs (e.g. in pipelines or dashboards). For launch execution, the agent uses MCPs (resend, meta-ads, google-ads, posthog, stripe, twitter) when configured in Cursor/Claude Code; the package’s MCP_SERVERS and mcpCommandGenerator document the same tools and patterns for automation.