Back to Details

measuring-product-market-fit

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Measuring Product-Market Fit

衡量产品-市场契合度（PMF）

Scope

适用范围

Covers
  • Measuring PMF using a triangulated signal set (survey + behavior + customer evidence)
  • Running and interpreting the Sean Ellis “Very Disappointed” survey (overall + by segment)
  • Reading retention curves / cohort retention as PMF evidence (and knowing when they mislead)
  • Using reference-customer / advocacy signals as an additional PMF proxy
  • Detecting PMF drift (market shifts, rising expectations, competitive resets) and setting a re-measurement cadence
  • Special handling for marketplaces (measure PMF per side; focus on the “hard side” first)
When to use
  • “Do we have PMF? For which segment?”
  • “Run a Sean Ellis PMF survey and tell me what it means.”
  • “Build a PMF scorecard with retention + survey + references.”
  • “Our market shifted—did we lose PMF?”
  • “We want a go/no-go signal for scaling growth spend or launching publicly.”
When NOT to use
  • You haven’t defined the problem/ICP yet (use 
    problem-definition
    ).
  • You only need a survey instrument, not a full PMF measurement system (use 
    designing-surveys
    ).
  • You’re deciding whether/how to pivot (use 
    startup-pivoting
    ) rather than measuring PMF signals.
  • You need a product vision/strategy doc as the primary output (use 
    defining-product-vision
    / 
    ai-product-strategy
    ).
涵盖内容
  • 使用三角验证信号集（调查+行为数据+客户证据）衡量PMF
  • 执行并解读**Sean Ellis“非常失望”**调查（整体结果+细分群体结果）
  • 将留存曲线/cohort留存作为PMF证据进行分析（并了解其可能产生误导的场景）
  • 将参考客户/用户推荐信号作为PMF的补充指标
  • 检测PMF漂移（市场变化、用户期望提升、竞品迭代）并设定重新测量的周期
  • 针对平台型产品的特殊处理（分别衡量双边的PMF；优先聚焦“难获客的一侧”）
适用场景
  • “我们是否达成了PMF？针对哪些细分群体？”
  • “执行Sean Ellis PMF调查并解读结果。”
  • “构建包含留存数据+调查+参考客户的PMF评分卡。”
  • “市场环境变化了——我们是否失去了PMF？”
  • “我们需要一个是否扩大增长投入或公开发布的决策依据。”
不适用场景
  • 尚未明确问题/ICP（理想客户画像）的情况（请使用
    problem-definition
    ）。
  • 仅需要调查工具，而非完整PMF测量体系的情况（请使用
    designing-surveys
    ）。
  • 正在决定是否/如何转型（请使用
    startup-pivoting
    ）而非测量PMF信号的情况。
  • 主要输出为产品愿景/战略文档的情况（请使用
    defining-product-vision
    / 
    ai-product-strategy
    ）。

Inputs

输入要求

Minimum required
  • Product + category + current stage (pre-PMF / early PMF / growth / mature)
  • Business model: B2B / B2C / marketplace (and, for marketplaces, which side you’re focusing on)
  • Your current best guess at the target segment/ICP (and any meaningful segments)
  • Definition of active user and the core value moment (the action that indicates value received)
  • What data you can access: survey channels, product analytics, retention cohorts, revenue, qualitative feedback, reference customers/testimonials
  • Time horizon and constraints (deadline, privacy/PII constraints, internal-only vs shareable)
Missing-info strategy
  • Ask up to 5 questions from references/INTAKE.md, then proceed.
  • If key inputs are missing, proceed with explicit assumptions and label confidence.
  • Do not request secrets. If data includes PII, ask for redacted excerpts or aggregated fields.
最低必填项
  • 产品+品类+当前阶段（Pre-PMF/早期PMF/增长期/成熟期）
  • 商业模式：B2B/B2C/平台型产品（平台型产品需明确聚焦哪一侧）
  • 目前对目标细分群体/ICP的最佳假设（以及所有有意义的细分维度）
  • 活跃用户和核心价值时刻的定义（即用户感知到价值的关键行为）
  • 可获取的数据类型：调查渠道、产品分析数据、留存cohort、收入数据、定性反馈、参考客户/推荐语
  • 时间范围与约束条件（截止日期、隐私/PII限制、内部使用 vs 可共享）
缺失信息处理策略
  • 从references/INTAKE.md中最多提出5个问题，然后继续推进。
  • 若关键输入缺失，基于明确假设推进工作并标注置信度。
  • 不得索要敏感信息。若数据包含PII，请要求提供脱敏片段或聚合字段。

Outputs (deliverables)

输出成果（交付物）

Produce a PMF Measurement Pack (Markdown in-chat; or as files if requested) containing:
  1. Context snapshot (product, stage, decision, timebox, segments, constraints)
  2. PMF measurement model (core value moment, active user definition, signal set, thresholds as heuristics)
  3. Sean Ellis survey plan + results (sample definition, questions, response counts, “very disappointed” % overall + by segment, top benefits)
  4. Behavioral evidence (retention/cohort summary + engagement frequency; instrumentation gaps + how they affect confidence)
  5. Reference-customer / advocacy evidence (who is willing to vouch; quotes; counts vs heuristic targets)
  6. PMF Scorecard (signals, targets, current state, confidence, evidence links/notes)
  7. Diagnosis + action plan (PMF status by segment; top drivers; prioritized next actions/experiments)
  8. Risks / Open questions / Next steps (always included)
Templates and checklists:
  • references/TEMPLATES.md
  • references/CHECKLISTS.md
  • references/RUBRIC.md
生成PMF测量包（聊天内以Markdown格式呈现；若有需求可生成文件），包含：
  1. 背景快照（产品、阶段、待决策事项、时间限制、细分群体、约束条件）
  2. PMF测量模型（核心价值时刻、活跃用户定义、信号集、作为启发式规则的阈值）
  3. Sean Ellis调查方案+结果（样本定义、问题、响应数量、整体及各细分群体的“非常失望”占比、核心收益主题）
  4. 行为证据分析（留存/cohort总结+参与频率；数据采集缺口及其对置信度的影响）
  5. 参考客户/用户推荐证据（愿意为产品背书的用户；引用内容；数量与启发式目标对比）
  6. PMF评分卡（信号指标、目标值、当前状态、置信度、证据链接/备注）
  7. 诊断结论+行动计划（各细分群体的PMF状态；核心驱动因素；优先级排序的后续行动/实验）
  8. 风险/待解决问题/下一步计划（必含内容）
模板与检查清单：
  • references/TEMPLATES.md
  • references/CHECKLISTS.md
  • references/RUBRIC.md

Workflow (7 steps)

工作流程（7个步骤）

1) Intake + decision framing

1) 需求收集+决策框架梳理

  • Inputs: User context; references/INTAKE.md.
  • Actions: Confirm the decision (scale spend, launch, refocus ICP, pricing), the timebox, and the audience. Define “what will we do differently based on this?”
  • Outputs: Context snapshot + measurement constraints.
  • Checks: A stakeholder can answer: “What decision will this change by <date>?”
  • 输入：用户提供的背景信息；references/INTAKE.md
  • 行动：确认待决策事项（扩大投入、发布、重新聚焦ICP、定价）、时间限制以及受众。明确“基于本次结果我们将做出哪些不同决策？”
  • 输出：背景快照+测量约束条件
  • 校验标准：利益相关者能够回答：“在<日期>前，本次结果将改变哪项决策？”

2) Define the PMF measurement model (and segments)

2) 定义PMF测量模型（及细分群体）

  • Inputs: Product + segment hypotheses; data availability.
  • Actions: Define:
    • The core value moment and active user definition
    • The segment(s) to evaluate (ICP + meaningful slices)
    • The signal set (survey + behavior + customer evidence) and what “good” looks like (as heuristics)
  • Outputs: PMF measurement model + segment plan.
  • Checks: Each signal has (a) a metric definition, (b) a data source, and (c) a limitation note.
  • 输入：产品+细分群体假设；数据可获取性
  • 行动：定义以下内容：
    • 核心价值时刻和活跃用户的定义
    • 需评估的细分群体（ICP+有意义的细分维度）
    • 信号集（调查+行为数据+客户证据）以及“达标”的标准（作为启发式规则）
  • 输出：PMF测量模型+细分群体评估计划
  • 校验标准：每个信号都包含(a)指标定义、(b)数据来源、(c)局限性说明

3) Run the Sean Ellis PMF survey (must-have test)

3) 执行Sean Ellis PMF调查（必填测试）

  • Inputs: Target population list (active users); distribution channel; references/TEMPLATES.md (PMF block).
  • Actions: Draft and run:
    • “How would you feel if you could no longer use <product>?” (Very / Somewhat / Not disappointed)
    • Follow-up: “What is the primary benefit you receive?” (text)
    • Segment respondents (persona/ICP, use case, tenure) to find the “must-have” cohort
  • Outputs: Survey plan + results table (overall + by segment) + top benefit themes.
  • Checks: Sample definition is explicit; results include counts (n), not only percentages; major bias risks are listed.
  • 输入：目标用户列表（活跃用户）；分发渠道；references/TEMPLATES.md（PMF模块）
  • 行动：设计并执行调查：
    • “如果无法再使用<产品>，你会有何感受？”（非常失望/有些失望/不失望）
    • 跟进问题：“你从产品中获得的核心收益是什么？”（开放式文本）
    • 对受访者进行细分（用户画像/ICP、使用场景、使用时长）以找出“不可或缺”的用户群体
  • 输出：调查方案+结果表格（整体+各细分群体数据）+核心收益主题汇总
  • 校验标准：样本定义明确；结果包含样本量(n)，而非仅百分比；列出主要偏差风险

4) Analyze behavioral evidence (retention + engagement)

4) 分析行为证据（留存+参与度）

  • Inputs: Product usage data or best-available proxy; activation definition.
  • Actions: Build a minimal behavioral picture:
    • Cohort retention (or repeat usage/purchase) by segment and tenure
    • Retention curve shape (improving/flat/decaying) and interpretation
    • Engagement frequency vs the product’s natural cadence (daily/weekly/monthly)
  • Outputs: Retention/engagement summary + confidence notes + instrumentation gaps.
  • Checks: Retention is measured from a clear cohort start; analysis separates activation from retention.
  • 输入：产品使用数据或最佳可用替代数据；激活定义
  • 行动：构建极简行为分析视图：
    • 按细分群体和使用时长划分的Cohort留存（或重复使用/购买数据）
    • 留存曲线形态（上升/平稳/下降）及解读
    • 参与频率与产品自然使用周期（每日/每周/每月）的对比
  • 输出：留存/参与度分析总结+置信度说明+数据采集缺口
  • 校验标准：留存数据基于明确的cohort起始点；分析将激活与留存区分开

5) Collect reference-customer / advocacy evidence

5) 收集参考客户/用户推荐证据

  • Inputs: Customer list; CS/sales notes; reviews; testimonials.
  • Actions: Identify users willing to vouch publicly/privately:
    • B2B heuristic target: 6–8 reference customers
    • B2C heuristic target: 15–25 strong references/advocates
    • Capture the “why” (benefit) and the segment they represent
  • Outputs: Reference evidence log + gaps by segment.
  • Checks: References map to the intended ICP/segment; evidence is current (not from a different market era).
  • 输入：客户列表；客户成功/销售记录；评论；推荐语
  • 行动：识别愿意公开/私下为产品背书的用户：
    • B2B启发式目标：6–8个参考客户
    • B2C启发式目标：15–25个核心推荐者/支持者
    • 记录其认可的“核心收益”及所属细分群体
  • 输出：参考客户证据日志+各细分群体的缺口分析
  • 校验标准：参考客户与目标ICP/细分群体匹配；证据为近期数据（非不同市场时期的旧数据）

6) Synthesize into a PMF scorecard + diagnosis (by segment)

6) 整合为PMF评分卡+细分诊断

  • Inputs: Survey + behavior + reference evidence.
  • Actions: Triangulate signals to answer:
    • Do we have PMF for any segment? Which one is strongest?
    • What are the top drivers of “must-have” value?
    • What’s blocking PMF for adjacent segments?
    • Are we at risk of PMF drift (market shift, expectations rising)?
  • Outputs: PMF Scorecard + diagnosis narrative + confidence rating.
  • Checks: Diagnosis is segment-specific and evidence-backed; “unknowns” are explicit.
  • 输入：调查+行为数据+参考客户证据
  • 行动：交叉验证信号以回答以下问题：
    • 我们是否在某些细分群体中达成了PMF？哪个群体的契合度最高？
    • “不可或缺”价值的核心驱动因素是什么？
    • 哪些因素阻碍了相邻细分群体达成PMF？
    • 我们是否面临PMF漂移风险（市场变化、用户期望提升）？
  • 输出：PMF评分卡+诊断说明+置信度评级
  • 校验标准：诊断针对细分群体且有证据支持；明确列出“未知事项”

7) Quality gate + action plan + cadence

7) 质量校验+行动计划+测量周期

  • Inputs: Draft pack; references/CHECKLISTS.md and references/RUBRIC.md.
  • Actions: Run the checklist + score with rubric. Produce:
    • Prioritized next actions/experiments (what to change, how to measure impact)
    • A PMF re-measurement cadence + drift triggers
    • Risks / Open questions / Next steps
  • Outputs: Final PMF Measurement Pack.
  • Checks: Actions are concrete enough to execute next sprint/quarter; measurement plan includes owners and dates (if known).
  • 输入：草稿版测量包；references/CHECKLISTS.md和references/RUBRIC.md
  • 行动：执行检查清单+按评分标准打分。生成以下内容：
    • 优先级排序的后续行动/实验（需调整的内容、衡量影响的方式）
    • PMF重新测量周期+漂移触发条件
    • 风险/待解决问题/下一步计划
  • 输出：最终版PMF测量包
  • 校验标准：行动足够具体可在下一个迭代/季度执行；测量计划包含负责人和时间（若已知）

Quality gate (required)

质量校验（必填）

  • Use references/CHECKLISTS.md and references/RUBRIC.md.
  • Always include: Risks, Open questions, Next steps.
  • 使用references/CHECKLISTS.md和references/RUBRIC.md
  • 必含内容：风险、待解决问题、下一步计划

Examples

示例

Example 1 (B2B SaaS, early growth):
“Use 
measuring-product-market-fit
. Product: AI meeting notes for account executives. Segments: mid-market sales teams vs SMB founders. Data: 90-day cohorts + in-app survey. Decision: whether to scale paid acquisition next quarter. Output: a PMF Measurement Pack.”
Example 2 (Marketplace, supply-first):
“We’re building a caregiver marketplace. We have early demand, but supply is thin. Measure PMF for the supply side first using a PMF survey + retention proxies. Output a scorecard and a plan to strengthen the core value exchange.”
Boundary example (insufficient inputs):
“Do we have PMF?”
Response: ask up to 5 intake questions (segment, active user definition, data sources, survey channel, decision), then produce a minimal PMF Measurement Pack with explicit assumptions and confidence limits.
示例1（B2B SaaS，早期增长阶段）：
“使用
measuring-product-market-fit
。产品：面向销售专员的AI会议纪要工具。细分群体：中大型销售团队 vs SMB创始人。数据：90天cohort数据+应用内调查。待决策事项：是否在下一季度扩大付费获客投入。输出：PMF测量包。”
示例2（平台型产品，供给侧优先）：
“我们正在构建一个护工匹配平台。需求端已有早期用户，但供给端资源不足。优先使用PMF调查+留存替代指标衡量供给侧的PMF。输出评分卡及强化核心价值交换的计划。”
边界示例（输入信息不足）：
“我们是否达成了PMF？”
回应：提出最多5个需求收集问题（细分群体、活跃用户定义、数据来源、调查渠道、待决策事项），然后基于明确假设生成极简版PMF测量包并标注置信度限制。