Back to Details

designing-surveys

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Designing Surveys

调研设计

Scope

适用范围

Covers
  • Designing product surveys that answer a specific decision (not “general feedback”)
  • Choosing the right audience, sampling, and timing (including “best customers” cohorts)
  • Writing clear, unbiased questions and using good scales (CSAT vs NPS guidance)
  • Building an instrument that works on mobile (logic, required fields, option visibility)
  • Planning analysis and turning results into decisions and follow-ups
When to use
  • “Design a customer survey for…”
  • “Create an onboarding survey to profile users / separate buyer vs user.”
  • “We need a CSAT/NPS/PMF survey.”
  • “Draft a cancellation / churn survey.”
  • “Help me write survey questions and an analysis plan.”
When NOT to use
  • You need deep “why” stories and context (use 
    conducting-user-interviews
    )
  • You need to measure causal impact of a change (use an experiment/A/B test, not a survey)
  • Your reachable sample is extremely small (n < ~30) and you need directional insight → interviews may be better
  • The topic is high-risk (legal/medical/safety) or requires formal survey science review; involve an expert
涵盖内容
  • 设计用于解答特定决策问题的产品调研（而非“一般性反馈”调研）
  • 选择合适的受众、抽样方式和调研时机（包括“核心客户”群组）
  • 撰写清晰、无偏见的问题，并选择合适的量表（含CSAT与NPS使用指导）
  • 构建适配移动端的调研工具（含逻辑设置、必填项、选项可见性）
  • 规划分析方案，并将调研结果转化为决策与后续行动
适用场景
  • “为……设计一份客户调研”
  • “制作一份新用户入门调研，用于用户画像分析 / 区分购买者与使用者”
  • “我们需要一份CSAT/NPS/PMF调研”
  • “起草一份用户流失调研问卷”
  • “帮我撰写调研问题和分析计划”
不适用场景
  • 需要深度了解“原因”和背景信息（请使用
    conducting-user-interviews
    ）
  • 需要衡量某项变更的因果影响（请使用实验/A/B测试，而非调研）
  • 可触达的样本量极小（n < ~30）且需要方向性洞察→访谈可能更合适
  • 主题涉及高风险（法律/医疗/安全）或需要正式调研科学审核；请咨询专业人士

Inputs

输入信息

Minimum required
  • Product + target user(s)/segment(s)
  • The decision to make (what will change based on the survey) + deadline
  • Survey type (e.g., onboarding profiling, CSAT, NPS, PMF, churn, feature discovery)
  • Distribution channel(s) (in-product, email, customer success, etc.) + sampling constraints
  • Privacy/compliance constraints (what data you can/can’t collect)
Missing-info strategy
  • Ask up to 5 questions from references/INTAKE.md.
  • If still missing, proceed with explicit assumptions and list Open questions that would change the design.
最低要求
  • 产品信息 + 目标用户/用户细分群体
  • 待决策事项（调研结果将如何影响决策）+ 截止日期
  • 调研类型（例如：新用户入门画像调研、CSAT调研、NPS调研、PMF调研、用户流失调研、功能发现调研）
  • 分发渠道（产品内、邮件、客户成功团队等）+ 抽样限制
  • 隐私/合规限制（可收集/不可收集的数据类型）
缺失信息处理策略
  • 从references/INTAKE.md中选取最多5个问题进行询问。
  • 若仍有信息缺失，基于明确假设推进工作，并列出待确认问题，这些问题可能会影响调研设计。

Outputs (deliverables)

输出成果（交付物）

Produce a Survey Pack in Markdown (in-chat; or as files if the user requests):
  1. Context snapshot (decision, audience, channel, constraints)
  2. Survey brief (goal, target population, sampling, timing, success criteria)
  3. Questionnaire (questions with rationale + response types; question IDs)
  4. Survey instrument table (copy/paste-ready for building in a survey tool)
  5. Analysis + reporting plan (segments, cuts, coding plan, decision thresholds)
  6. Launch plan + QA checklist (pilot, mobile QA, bias checks, comms, follow-ups)
  7. Risks / Open questions / Next steps (always included)
Templates: references/TEMPLATES.md
Expanded heuristics: references/WORKFLOW.md
制作Markdown格式的调研包（可在对话中直接提供；若用户要求，也可作为文件交付）：
  1. 背景快照（待决策事项、受众、渠道、限制条件）
  2. 调研简介（目标、目标人群、抽样方式、时机、成功标准）
  3. 问卷（含问题设计依据、响应类型、问题ID）
  4. 调研工具表格（可直接复制粘贴用于调研工具搭建）
  5. 分析与报告计划（用户细分、数据拆分、编码方案、决策阈值）
  6. 发布计划与QA清单（试点测试、移动端适配检查、偏见排查、沟通方案、后续跟进）
  7. 风险 / 待确认问题 / 下一步行动（必须包含）
模板参考：references/TEMPLATES.md
扩展指南：references/WORKFLOW.md

Workflow (7 steps)

工作流程（7个步骤）

1) Intake + decision framing

1) 需求收集与决策框架搭建

  • Inputs: User context; references/INTAKE.md.
  • Actions: Clarify the decision, timeline, primary audience, and distribution channel(s). Name the “unknowns” the survey must resolve.
  • Outputs: Context snapshot + survey goal.
  • Checks: You can state the decision in one sentence (“We are deciding whether to… by <date>”).
  • 输入信息：用户提供的背景信息；references/INTAKE.md
  • 行动：明确待决策事项、时间线、核心受众和分发渠道。列出调研需要解决的“未知问题”。
  • 输出成果：背景快照 + 调研目标
  • 检查标准：可通过一句话表述待决策事项（“我们将在<日期>前决定是否……”）

2) Define the audience + sampling plan (who, when, how many)

2) 定义受众与抽样计划（对象、时机、样本量）

  • Inputs: Context snapshot.
  • Actions: Choose primary segment(s) and a sampling frame. Prefer behavior/recency-based cohorts (e.g., “signed up 3–6 months ago and active”) when you need accurate recall.
  • Outputs: Sampling plan (in brief) + segment cuts.
  • Checks: You can explain why each segment is included and what decision it informs.
  • 输入信息：背景快照
  • 行动：选择核心用户细分群体和抽样框架。若需要准确的用户回忆，优先选择基于行为/活跃度的群组（例如：“注册3-6个月且仍活跃的用户”）
  • 输出成果：抽样计划（简介）+ 用户细分维度
  • 检查标准：可解释每个细分群体被纳入的原因，以及其对应的决策事项

3) Choose the measurement design (metrics, scales, prioritization)

3) 选择测量设计（指标、量表、优先级设置）

  • Inputs: Survey goal + audience.
  • Actions: Pick the core metric(s) (often CSAT); add 1–2 diagnostic questions that force prioritization (e.g., “pick top 3 barriers”) and frequency/impact weighting.
  • Outputs: Measurement plan (metric + diagnostics) + draft question list.
  • Checks: Every question maps to a decision, hypothesis, or segment cut; no “nice-to-have” questions.
  • 输入信息：调研目标 + 受众
  • 行动：确定核心指标（通常为CSAT）；添加1-2个诊断性问题以明确优先级（例如：“选出排名前三的障碍”），并设置频率/影响权重。
  • 输出成果：测量计划（指标 + 诊断问题）+ 问卷草稿
  • 检查标准：每个问题都对应一项决策、假设或用户细分维度；无“可问可不问”的问题

4) Draft the questionnaire (sections, wording, and logic)

4) 起草问卷（板块、措辞、逻辑）

  • Inputs: Measurement plan; templates.
  • Actions: Write questions using neutral wording, single concepts per question, and consistent scales. Add segmentation/profile questions only if you will use them in analysis.
  • Outputs: Questionnaire with question IDs, response types, options, and skip logic notes.
  • Checks: No double-barreled or leading questions; completion time target ≤ 3–6 minutes for most surveys.
  • 输入信息：测量计划；模板
  • 行动：使用中立措辞撰写问题，每个问题仅包含单一概念，使用统一量表。仅添加将用于分析的用户细分/画像问题。
  • 输出成果：含问题ID、响应类型、选项、跳转逻辑说明的问卷
  • 检查标准：无双重含义或诱导性问题；大多数调研的完成时间目标≤3-6分钟

5) Build the instrument table + QA it (mobile + bias)

5) 制作调研工具表格并进行QA（移动端适配 + 偏见排查）

  • Inputs: Questionnaire draft.
  • Actions: Convert to an instrument table for implementation (IDs, types, options, required, logic). Check mobile rendering (all scale points visible) and option order/randomization.
  • Outputs: Survey instrument table + QA checklist items.
  • Checks: Scale labels are unambiguous; required questions are minimal; “Other (free text)” exists when appropriate.
  • 输入信息：问卷草稿
  • 行动：将问卷转换为便于落地的工具表格（含ID、类型、选项、必填项、逻辑）。检查移动端显示效果（所有量表选项可见）以及选项顺序/随机化设置。
  • 输出成果：调研工具表格 + QA清单项
  • 检查标准：量表标签清晰明确；必填问题数量最少；在合适的位置设置“其他（自由文本）”选项

6) Plan the launch (pilot, comms, monitoring, follow-ups)

6) 规划发布方案（试点、沟通、监控、后续跟进）

  • Inputs: Instrument + sampling plan.
  • Actions: Define pilot (small n), launch dates, reminders, incentives, and a monitoring plan. If the goal is message validation, consider a behavioral “survey” via ad/landing tests instead of asking opinions.
  • Outputs: Launch plan + monitoring metrics (response rate, drop-off, segment mix).
  • Checks: You have a plan for low response rate and for closing the loop with respondents.
  • 输入信息：调研工具 + 抽样计划
  • 行动：定义试点测试（小样本量）、发布日期、提醒机制、激励措施和监控计划。若目标是验证信息传递效果，可考虑通过广告/落地页测试开展行为式“调研”，而非直接询问用户意见。
  • 输出成果：发布计划 + 监控指标（响应率、流失率、用户细分占比）
  • 检查标准：制定了低响应率应对方案，以及与受访者的反馈闭环方案

7) Analysis + report plan + quality gate

7) 分析与报告计划 + 质量审核

  • Inputs: Final instrument + goals.
  • Actions: Define how you’ll analyze (segments, cuts, coding of open-ended), the decision thresholds, and how results will be communicated. Run references/CHECKLISTS.md and score references/RUBRIC.md. Add Risks/Open questions/Next steps.
  • Outputs: Final Survey Pack.
  • Checks: A stakeholder can review async and decide “ship / adjust / investigate” without another meeting.
  • 输入信息：最终调研工具 + 目标
  • 行动：定义分析方式（用户细分、数据拆分、开放式问题编码）、决策阈值，以及结果沟通方式。使用references/CHECKLISTS.md进行检查，并通过references/RUBRIC.md进行评分。添加风险/待确认问题/下一步行动。
  • 输出成果：最终调研包
  • 检查标准：利益相关者可异步审阅并直接决定“发布 / 调整 / 深入调研”，无需额外会议

Quality gate (required)

质量审核（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进行审核。
  • 必须包含：风险、待确认问题、下一步行动

Examples

示例

Example 1 (Onboarding): “Design an onboarding survey to identify the buyer vs user and route leads appropriately.”
Expected: short profiling questions (3–4 screens), clear segmentation fields, and a follow-up plan to avoid irrelevant outreach.
Example 2 (Product friction): “Design a CSAT survey to find the top 3 productivity blockers for active users and how often they occur.”
Expected: CSAT + forced-ranking diagnostics + frequency weighting, plus an analysis plan that yields a ranked backlog of issues.
Boundary example: “We want to know if feature X caused retention to improve—send a survey.”
Response: push back; recommend experiment/instrumentation for causality, and use a survey only for qualitative context (or run interviews).
示例1（新用户入门调研）：“设计一份新用户入门调研，用于区分购买者与使用者，并对线索进行合理分配。”
预期成果：简短的画像问题（3-4个页面）、清晰的用户细分字段，以及避免无关触达的后续跟进计划。
示例2（产品摩擦调研）：“设计一份CSAT调研，找出活跃用户排名前三的生产力障碍及其发生频率。”
预期成果：CSAT量表 + 强制排名诊断问题 + 频率权重，以及可生成问题优先级清单的分析计划。
边界案例：“我们想了解功能X是否提升了用户留存率——发布一份调研。”
回应：建议采用实验/埋点分析来验证因果关系，仅在需要定性背景信息时使用调研（或开展访谈）。