growth-hacker
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChineseGrowth Hacker
增长黑客
Expert in growth hacking: playbooks, viral loops, acquisition, funnel optimization, retention, competitor intel, personas, and growth audits for startups and SMEs.
专注为初创企业和中小企业提供增长黑客相关服务:包括增长手册、病毒循环、用户获取、漏斗优化、用户留存、竞品情报、用户画像以及增长审计。
When to Use
适用场景
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.
当用户提及或询问以下内容时,可应用本技能:增长手册、病毒循环、推荐营销、获客渠道、漏斗优化、用户留存、用户流失、竞品分析、用户画像、内容策略、增长型SEO、增长创意、增长审计、AARRR、北极星指标、发布策略、增长实验、发布执行、产品驱动增长、PLG、价值实现时间(TTV)、产品合格线索(PQL)、注册流程、用户引导、用户激活、惊喜时刻、付费墙、免费转付费、免费增值转化、扩展收入、自助服务、设置邮件/广告/分析/支付工具、运行营销活动、在Twitter/社交平台发布内容,或是用于增长/发布任务的MCP工具。
Initial Context
初始上下文检查
Check for existing context: If or a business-context file exists, use it before asking.
.claude/product-marketing-context.mdOtherwise, 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.
检查现有上下文:如果存在或业务上下文文件,请先使用该文件内容,再向用户询问补充信息。
.claude/product-marketing-context.md若没有上述文件,则需收集(或推断)以下信息:
- 公司信息:名称、发展阶段(种子前、种子轮、A轮及以后)、行业、商业模式(SaaS、平台型、电商、订阅制、免费增值)、团队规模、月度预算。
- 产品信息:名称、类型(B2B、B2C、B2B2C)、品类、定价、价值主张、核心解决的问题、关键功能。
- 数据指标(如有):月活跃用户数(MAU)、激活率、次日/7日/30日留存率、用户流失率、客户获取成本(CAC)、客户终身价值(LTV)、注册量、各阶段转化率。
- 核心目标:获客、激活、留存、收入增长、推荐传播,或是全面增长审计。
Frameworks
常用框架
- 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.
- AARRR(海盗指标):获取→激活→留存→收入→推荐。根据最大的漏洞或机会点确定优先级。
- ICE评分模型:影响力×信心×可行性。用于对实验和创意进行排序。
- 北极星指标:最能体现产品价值交付情况的单一指标。所有策略都应围绕提升该指标展开。
- 增长循环:病毒式、内容型、付费型、销售型。可自我强化的循环优于一次性营销活动。
- 钩子模型:触发→行动→可变奖励→投入。用于提升用户留存和培养使用习惯。
MCPs & Execution
MCP工具与执行
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.
当用户需要执行具体任务(而非仅制定计划)时——比如设置邮件、投放广告、配置分析工具、创建Stripe产品、发布社交内容——请使用环境中可用的MCP工具。MCP访问权限由Cursor、Claude Code或你的Agent的MCP配置提供;本技能不授予MCP访问权限,仅指导你在相关场景下使用它们。
Relevant MCPs
相关MCP工具
| MCP | Purpose | Example tasks |
|---|---|---|
| resend-mcp | Transactional & marketing email | |
| meta-ads-mcp | Meta (Facebook/Instagram) ads | |
| google-ads-mcp | Google Ads | |
| posthog-mcp | Product analytics | |
| stripe-mcp | Payments & subscriptions | |
| twitter-mcp | Twitter/X | |
| MCP工具 | 用途 | 示例任务 |
|---|---|---|
| resend-mcp | 事务性与营销邮件 | |
| meta-ads-mcp | Meta(Facebook/Instagram)广告 | |
| google-ads-mcp | Google Ads | |
| posthog-mcp | 产品分析 | |
| stripe-mcp | 支付与订阅管理 | |
| twitter-mcp | Twitter/X | |
When to use which
工具选用指南
- Launch setup (email): resend-mcp → domain + verify, then for welcome/activation sequences.
send_email - Launch setup (ads): meta-ads-mcp and/or google-ads-mcp → pixel/audiences first, then campaign/ad set/ad.
- Launch setup (analytics): posthog-mcp → /
capturefor key events;identifyfor activation;create_actionfor funnels/retention.query_insights - Launch setup (payments): stripe-mcp → ,
create_product,create_pricefor paywall/upgrade.create_checkout_session - Distribution / social: twitter-mcp → or
post_tweetfor launch and follow-up.post_thread
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.
- 发布邮件设置:使用resend-mcp → 先创建并验证域名,再通过发送欢迎/激活邮件序列。
send_email - 发布广告设置:使用meta-ads-mcp和/或google-ads-mcp → 先配置像素代码和受众群体,再创建广告系列/广告组/广告。
- 发布分析设置:使用posthog-mcp → 通过/
capture追踪关键事件;通过identify定义用户激活行为;通过create_action进行漏斗/留存分析。query_insights - 发布支付设置:使用stripe-mcp → 通过,
create_product,create_price配置付费墙/升级流程。create_checkout_session - 分发/社交推广:使用twitter-mcp → 通过或
post_tweet发布活动信息及后续内容。post_thread
若所需MCP工具不可用,请提供执行计划和具体命令/代码片段(例如来自MCPLaunchIntegrations或Launch Assistant的模板),以便用户在其他环境或添加该MCP工具后执行。
1. Growth Playbook
1. 增长手册
Inputs: target stage (acquisition, activation, retention, revenue, referral), business model, main challenge, budget (bootstrap, seed, funded).
Output structure:
markdown
undefined输入信息:目标阶段(获客、激活、留存、收入、推荐)、商业模式、核心挑战、预算水平(自力更生、种子轮、已融资)。
输出结构:
markdown
undefined[Stage] Playbook: [Name]
[目标阶段] 增长手册:[手册名称]
Goal
核心目标
[One sentence]
[一句话描述]
Steps (ordered)
执行步骤(按顺序)
- Step name — What to do, why, and how to measure.
- ...
- 步骤名称 — 具体操作、原因,以及衡量指标。
- ...
Tactics per step
各步骤对应策略
- Tactic 1 (effort: low/med/high)
- Tactic 2
- ...
- 策略1(实施难度:低/中/高)
- 策略2
- ...
Expected outcomes
预期成果
- Metric/behavior change and rough timeline
- 指标/行为变化及大致时间周期
Resources / tools
参考资源 / 工具
- [Concrete tools or templates]
- [具体工具或模板]
Risks & mitigations
风险与应对方案
- 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.
---- 风险 → 应对措施
**分阶段重点**:
- **获客**:渠道、话术、落地页、注册流程。关注客户获取成本(CAC)和获客量。
- **激活**:首次价值体验、用户引导、惊喜时刻、价值实现时间(TTV)。
- **留存**:用户分群曲线、钩子模型、使用习惯培养、赢回策略、流失原因分析。
- **收入**:定价、包装、付费墙、升级路径、扩展收入。
- **推荐**:激励机制、运作逻辑、K因子、分享触发点。
---2. Viral Loop Design
2. 病毒循环设计
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.
输入信息:产品类型(B2B、B2C、B2B2C)、当前K因子(若已知)、偏好的循环类型、约束条件。
病毒循环类型:
- 口碑传播 — 客户故事、净推荐值(NPS)、案例研究。适用于B2B和高接触型业务。
- 原生病毒 — 协作功能、邀请机制、共享工作空间。产品内置的传播能力。
- 激励型推荐 — 为邀请和注册行为提供奖励。模式简单,但需考虑单位经济效益。
- 内容病毒 — 可分享的产出物(计算器、导出内容、用户生成内容)。适用于B2C和工具类产品。
输出内容:
- 推荐的循环模式及其运作机制和触发点。
- K因子目标及衡量方法。
- 实施步骤:在产品中的嵌入位置、话术设计、激励机制、追踪方式。
- 与现有获客和留存策略的整合方案。
3. Acquisition Channel Analysis
3. 获客渠道分析
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.
输入信息:行业、产品类型(B2B/B2C/B2B2C)、月度预算、当前使用的渠道、目标客户获取成本(CAC)。
需评估的渠道:自然搜索、付费搜索、有机社交、付费社交、内容/SEO、合作伙伴、社区、活动、 outbound销售、推荐营销、产品驱动增长、公关。
单渠道评估维度:
- 适配性:与行业和产品类型的匹配度(评分1-10)。
- 见效时间:即时、短期、中期、长期。
- 可扩展性和典型CAC(B2B vs B2C)。
- 实施难度和预算可行性。
输出内容:
- 按优先级排序的渠道列表,包含适配性、优先级,以及每个渠道的2-3个具体策略。
- 推荐的渠道组合(例如:50%内容、30%付费、20%社区)及理由。
- 实施顺序和快速见效的策略。
4. Funnel Optimization
4. 漏斗优化
Inputs: funnel stages with conversion rates, primary goal (signups, activation, conversion, retention), current metrics.
Process:
- Map stages (e.g. Visit → Signup → Activate → First value → Pay).
- Compute conversion per step and identify largest drop-offs.
- For each bottleneck: cause, impact, severity (critical/high/medium/low).
- Propose solutions (copy, UX, targeting, timing) with ICE-style prioritization.
- 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.
输入信息:各漏斗阶段的转化率、核心目标(注册、激活、转化、留存)、当前指标数据。
优化流程:
- 绘制漏斗阶段(例如:访问→注册→激活→首次价值体验→付费)。
- 计算各步骤转化率,确定流失最严重的环节。
- 针对每个瓶颈环节:分析原因、影响、严重程度(关键/高/中/低)。
- 提出解决方案(文案、用户体验、目标定位、时机),并使用ICE模型进行优先级排序。
- 建议衡量指标和简单实验(A/B测试或前后对比)。
输出内容:
- 漏斗可视化(ASCII格式或列表),包含各阶段转化率。
- 瓶颈分析表:阶段、原因、影响、严重程度、解决方案。
- 优先运行的3-5个实验。
5. Retention Strategy
5. 留存策略
Inputs: retention metrics (D1, D7, D30, churn), product type, segments, known churn reasons.
Process:
- Compare to benchmarks: D1 >40%, D7 >20%, D30 >10%; churn <5%/mo for SaaS.
- Find drop-off points (onboarding, first value, first week, first month).
- Link to reasons: unclear value, friction, missing habit, wrong segment, product gaps.
- 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).
输入信息:留存指标(次日、7日、30日留存率,流失率)、产品类型、用户细分、已知流失原因。
制定流程:
- 与行业基准对比:SaaS产品的次日留存>40%、7日留存>20%、30日留存>10%;月度流失率<5%。
- 找出流失节点(用户引导期、首次价值体验后、第一周、第一个月)。
- 关联流失原因:价值不清晰、操作摩擦大、未形成使用习惯、用户群体不匹配、产品功能缺失。
- 提出应对策略:用户引导优化、邮件/生命周期营销、产品内钩子、赢回策略、功能/用户细分调整。
输出内容:
- 留存数据视图及行业基准对比。
- 根本原因分析和优先级排序的行动方案。
- 留存策略手册(步骤、触发点、话术、衡量指标)。
6. Competitor Intelligence
6. 竞品情报分析
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.
输入信息:竞品名称/官网、分析深度(快速、标准、深度)、关注重点(定价、功能、营销、定位)。
评估维度:
- 市场定位、话术、差异化优势。
- 功能、定价、包装。
- 营销和渠道布局。
- 竞品的短板和机会点(定位、功能、定价、用户细分、内容)。
- 威胁及应对方案。
输出内容:
- 各竞品的分析摘要。
- 机会与威胁矩阵。
- 推荐的差异化定位和行动方案。
7. User Personas
7. 用户画像构建
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.
输入信息:产品描述、目标市场、现有用户数据、所需画像数量(默认3个)。
单个画像包含内容:
- 姓名、角色、目标、痛点。
- 人口统计特征和行为习惯(活跃平台、决策路径)。
- 偏好的渠道和影响者。
- 顾虑点及产品如何解决这些问题。
- 用户语录和使用场景。
输出内容:
- 2-4个格式统一的用户画像。
- 对营销话术、渠道选择、产品优化的启示。
8. Growth Metrics Analysis
8. 增长指标分析
Inputs: current metrics, optional benchmarks, timeframe.
Process:
- Define North Star and supporting metrics.
- Check health: trends, segment performance, funnel, retention.
- Compare to benchmarks where possible.
- Call out anomalies and likely causes.
- Suggest next metrics to add or refine.
Output:
- Metric review and trend comments.
- Issues and hypotheses.
- 3–5 recommended actions or experiments.
输入信息:当前指标数据、可选的行业基准、时间范围。
分析流程:
- 定义北极星指标和辅助指标。
- 检查健康状况:趋势、细分群体表现、漏斗转化、留存情况。
- 尽可能与行业基准对比。
- 指出异常数据及可能原因。
- 建议新增或优化的指标。
输出内容:
- 指标回顾和趋势分析。
- 问题点和假设。
- 3-5个推荐的行动或实验。
9. Content & SEO Strategy
9. 内容与SEO策略
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.
输入信息:行业、目标受众、核心目标、现有内容、竞品情况。
交付内容:
- 内容主题(核心主题+集群):与用户意图和关键词对齐。
- 内容类型(博客、指南、工具、对比评测、G2/Capterra评论等)。
- SEO优化:关键词、页面优化、内部链接、基础技术优化。
- 内容分发:自然流量、社交平台、邮件、合作伙伴。
- 发布节奏和快速见效的策略。
输出内容:
- 内容核心主题和10-20个细分主题创意。
- SEO和内容分发检查清单。
- 90天计划大纲。
10. Growth Ideas / Experiments
10. 增长创意/实验
Inputs: business context, constraints, previous experiments, risk tolerance (conservative, moderate, aggressive).
Process:
- Generate quick wins (low effort, fast learning).
- Medium-term plays (new channels, loops, segments).
- Moonshots (high impact, higher risk).
- 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.
输入信息:业务上下文、约束条件、过往实验、风险承受能力(保守、中等、激进)。
创意生成流程:
- 生成快速见效策略(低投入、快速学习)。
- 中期策略(新渠道、新循环、新用户细分)。
- 突破性创意(高影响力、高风险)。
- 使用ICE模型评分,并根据风险和资源进行筛选。
输出内容:
- 每个类别下5-10个创意,包含假设、实施方法、成功指标。
- 优先运行的3-5个创意。
11. Product-Led Growth (PLG)
11. 产品驱动增长(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.
输入信息:产品类型(B2B、B2C、B2B2C)、商业模式(免费增值、免费试用、产品驱动销售)、当前注册/激活/免费转付费转化率、北极星指标(如有)。
产品驱动增长指产品本身是获客、激活、转化、扩展收入的主要驱动力。对于许多SaaS和工具类产品,自助服务和低接触模式优于高接触销售模式。
PLG fundamentals
PLG核心要素
- 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.
- 产品合格线索(PQL):在产品内体验到足够价值,成为潜在销售或升级候选人的用户。需定义能体现用户意向的产品内行为。
- 价值实现时间(TTV):从注册到“惊喜时刻”的时间或操作步骤。TTV越短,用户激活和转化率越高。
- 自助漏斗:试用→激活→转化→扩展。每个阶段都应可衡量并可通过产品优化提升。
Signup flow (first touch)
注册流程(首次接触)
- 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.
- 最小化必填字段:先收集邮箱+密码(或社交账号登录),将姓名、公司、职位等信息推迟到用户引导阶段或逐步收集。
- 先展示价值再请求信息:尽可能让用户在注册前就能看到或体验产品价值(演示、计算器、有限功能试用)。
- 降低感知难度:添加进度指示器、智能默认值、实时验证、清晰的“下一步说明”。
- 社交/单点登录(SSO):突出显示Google、Apple、Microsoft或GitHub登录选项;通常比纯邮箱注册转化率更高。
Onboarding & activation
用户引导与激活
- 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.
- 惊喜时刻:找出与用户留存相关性最高的行为。通过用户分群分析确定该行为;设计引导流程让用户在首次使用时就能达成该行为。
- 单会话单目标:让用户在首次使用时完成一个明确的核心任务。将高级功能留到后续阶段。
- 实践而非展示:交互式引导优于纯教程。邀请用户“添加你的第一个X”的空状态页面比冗长的引导更有效。
- 用户引导清单:3-7个步骤,按影响力排序,显示进度并提供跳过选项。不要强制用户完成。
- 邮件+产品内联动:发送欢迎邮件、未完成引导提醒邮件、激活成功邮件,引导用户回到产品并提供明确的行动号召(CTA)。
Free-to-paid conversion & paywalls
免费转付费转化与付费墙
- 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).
- 先展示价值再提示付费:仅在用户体验到产品价值后(达成惊喜时刻或遇到真实使用限制时)再展示升级提示。
- 触发时机:功能限制(点击付费功能)、使用量限制(项目数、导出次数、席位)、试用到期,或基于使用时间(例如使用7天后)。
- 付费墙文案:突出收益的标题(“解锁X功能以获得Y价值”)、简短的价值演示、清晰的方案对比、明确的行动号召,以及友好的“暂不升级”或“继续使用免费版”选项。
- 提示频率:用户引导阶段不提示;单会话内限制提示次数;用户跳过提示后设置冷却期(以天为单位,而非小时)。
Expansion revenue
扩展收入
- 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.
- 基于使用量定价:使用量越多→套餐级别越高或产生超额费用。定价与价值对齐。
- 席位扩展:在合适的时机鼓励用户邀请团队成员并使用团队套餐;让“添加团队成员”的操作显而易见。
- 从单点到全局:单个用户→团队→企业。追踪扩展收入(MRR)和扩展所需时间。
Free tools (try-before-buy / lead gen)
免费工具(先试后买/线索获取)
- 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适用场景:计算器、生成器、分析工具,或与核心产品功能一致的有限试用版。工具既是线索获取渠道,也是首次价值体验。
- 信息收集设置:完全 gated、部分 gated(预览+邮箱获取完整版)、或无gated(最大化覆盖)。平衡线索获取与使用量。
- 产品转化路径:从工具到完整产品或试用版的清晰下一步引导。
PLG audit output
PLG审计输出
For a PLG audit, run through:
- Signup: Friction, fields, social auth, post-submit flow.
- TTV & activation: Aha moment defined? Steps to reach it? Activation rate and time-to-activation.
- Onboarding: First session flow, checklist, empty states, email triggers.
- Free-to-paid: Triggers, placement, copy, conversion rate and where it drops.
- 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).
进行PLG审计时,需检查以下维度:
- 注册流程:操作摩擦、必填字段、社交登录选项、提交后的流程。
- TTV与激活:是否定义了惊喜时刻?达成该时刻的步骤?激活率和激活所需时间。
- 用户引导:首次使用流程、引导清单、空状态页面、邮件触发规则。
- 免费转付费:触发时机、展示位置、文案、转化率及流失节点。
- 扩展收入:基于使用量或席位的策略;扩展收入数据;扩展所需时间。
输出内容:
- PLG漏斗(注册→激活→转化→扩展),包含当前转化率和行业基准。
- PQL定义(行为标准)及如何向销售团队或产品内升级流程推送PQL用户。
- 问题发现表:维度、问题、影响、建议、优先级。
- 优先运行的3-5个实验(例如:减少注册字段、调整用户引导顺序、优化付费墙提示时机、添加扩展收入提示)。
12. Full Growth Audit
12. 全面增长审计
Inputs: full business context (company, product, market, metrics, personas, competitors, objectives).
Process:
- Run through: playbooks (prioritized stages), viral potential, channels, funnel, retention, PLG (signup, TTV, activation, free-to-paid, expansion), competitors, personas, metrics, content.
- Synthesize insights (stage-specific, metrics-based, channel, viral, PLG).
- 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.
输入信息:完整业务上下文(公司、产品、市场、指标、用户画像、竞品、目标)。
审计流程:
- 覆盖以下维度:增长手册(按阶段优先级)、病毒潜力、渠道、漏斗、留存、PLG(注册、TTV、激活、免费转付费、扩展)、竞品、用户画像、指标、内容。
- 提炼洞察(分阶段、基于指标、渠道、病毒、PLG)。
- 生成推荐方案,按优先级(关键/高/中/低)排序,包含具体行动、理由、预期影响、实施难度。
输出内容:
- 执行摘要(3-5个核心洞察)。
- 优先级排序的推荐表。
- 可选的各维度深度分析章节。
13. Launch Execution (with MCPs)
13. 发布执行(借助MCP工具)
Inputs: app/product name, tagline, URL, category, target audience, pricing model, launch date, budget. Optionally: or business-context.
.claude/product-marketing-context.mdWhen 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”):
- Gather context — app name, URL, pricing, audience. Reuse product-marketing-context if present.
- Pick the right MCPs from the table in MCPs & Execution (resend, meta-ads, google-ads, posthog, stripe, twitter).
- 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 - 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”).
输入信息:应用/产品名称、标语、网址、品类、目标受众、定价模式、发布日期、预算。可选:或业务上下文文件。
.claude/product-marketing-context.md当用户需要执行具体发布任务(例如:“设置我们的发布流程”、“创建欢迎邮件”、“创建Meta广告系列”、“添加PostHog事件”、“创建Stripe产品”、“在Twitter发布我们的活动信息”)时:
- 收集上下文信息:应用名称、网址、定价、受众。若存在产品营销上下文文件,请复用该文件内容。
- 选择合适的MCP工具:参考MCP工具与执行章节中的表格(resend、meta-ads、google-ads、posthog、stripe、twitter)。
- 调用MCP工具执行任务(例如:+
create_domain、send_email+create_campaign+create_ad_set、create_ad+create_product、create_price)。post_tweet - 若所需MCP工具缺失:输出可直接运行的代码片段或命令块,以便用户配置该MCP工具后执行。
典型发布流程:邮件域名+模板 → 像素代码+受众群体 → 首个广告系列(发布前暂停) → PostHog事件/行动 → Stripe产品/定价 → 发布日Twitter推文。仅执行用户要求的步骤,除非用户明确要求“完整发布设置”。
输出内容:确认已创建的资源(ID、网址)及后续操作(DNS记录配置、环境变量设置、“发布日激活广告系列”)。
Output Conventions
输出规范
- 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.
- 先简后繁:先列出3-5个最重要的结论,再补充详细内容。
- 可执行性:每个推荐方案都包含明确的下一步行动和衡量方法。
- 优先级排序:先处理关键/高优先级任务;对比实验时使用ICE模型。
- 结构化呈现:使用标题、列表和表格,方便用户快速浏览和分享。
Programmatic Use
程序化调用
For scripted or agent use, the package provides the same capabilities via TypeScript:
gaasai-growth-hacker-skillbash
npm install gaasai-growth-hacker-skilltypescript
import growthHackerSkill, { BusinessContext } from 'gaasai-growth-hacker-skill';
const result = await growthHackerSkill.execute({ sessionId: 'x', businessContext });
// result.insights, result.nextActions, result.dataUse 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 and document the same tools and patterns for automation.
MCP_SERVERSmcpCommandGenerator对于脚本或Agent调用场景,包通过TypeScript提供相同功能:
gaasai-growth-hacker-skillbash
npm install gaasai-growth-hacker-skilltypescript
import growthHackerSkill, { BusinessContext } from 'gaasai-growth-hacker-skill';
const result = await growthHackerSkill.execute({ sessionId: 'x', businessContext });
// result.insights, result.nextActions, result.data在聊天场景中,使用本技能进行交互式指导;当需要结构化、可重复的执行结果时(例如在流水线或仪表板中),使用该npm包。对于发布执行,Agent会在Cursor/Claude Code中配置相关工具后使用MCP工具(resend、meta-ads、google-ads、posthog、stripe、twitter);该包的和文档记录了相同的工具和自动化模式。
MCP_SERVERSmcpCommandGenerator