定位是被讨论得最少、做错代价最高的一项产品决策。下面 15 个 Prompt 把 April Dunford 框架——竞争替代、独特属性、价值、市场框架、目标 segment——套到你的产品。覆盖横向转纵向、“新品类”框架，以及最难的情况：表面看像 12 个同类产品时怎么讲一个强故事。
适合哪些场景
写 v1 叙事的创始人、改 homepage hero 的 PMM、在能赢的单子上输给会讲故事竞品的销售 lead、把 roadmap 对齐到更锐故事的产品 lead。
什么时候不建议这样写 Prompt
战术性 messaging A/B 测试不要用——定位在 messaging 之上。早期采用 segment 还没清晰时也不要用，先访谈。
Prompt 结构公式
定位 Prompt 一定带这六个要素：
- 角色：让 AI 扮演谁（资深 PM / 独立创始人 / 产品设计师 / 独立开发者 / 增长负责人）。
- 上下文：阶段（想法 / MVP / 增长 / 规模化）、团队规模、流量或 ARR、平台（web / iOS / Android）、受众、限制。
- 目标：一个具体交付物——一段 PRD、一组用户故事、一个实验设计、一篇上线公告。
- 限制：时间线（本 sprint / 本季度）、要砍的范围、不能动的东西（现有流程、计费、合规）。
- 输出格式：表格、清单、可贴 ticket 的 JSON、或带标签的段落，能直接粘到 Linear / Notion / Jira。
- 示例 / 信号：1-2 份你欣赏的参考或竞品、加 1 个想避开的反例。
这套 Prompt 适合用在哪
- 新产品 v1 叙事
- Homepage hero 重定位
- 横向产品的纵向定位
- 竞品 pivot 后重新锚定
- 团队内部对齐文档
15 个可直接复制的 Prompt 模板
1. April Dunford 5 要素画布
默认型。每个产品跑一次，每年刷新。
You are a product positioning consultant in the April Dunford tradition. For {product}, output a 5-component canvas: (1) Competitive alternatives — what users do today (not just competitors, but spreadsheets, status quo, hired help). (2) Unique attributes — capabilities only we have. (3) Value — what those attributes enable that matters. (4) Target market — best-fit segment with characteristics. (5) Market category — the frame we sit in. Each component 2-4 bullets.
Context: {paste product brief}可替换变量： 产品 brief、目标客户、竞品
优化建议： 价值部分泛时追加：“For every value bullet, name the moment in the user week when that value becomes acute.”
2. “显然属于哪个品类”自检
For {product}, propose 3 alternative market categories we could position into. For each: who would we be compared against, what would buyers expect, what would we be measured on. Pick one and explain why it gives us the strongest position.
Context: {paste}3. 替代选项阶梯
List the 5 things users currently do instead of buying {product}. Rank them by frequency, not by sophistication. For each: what it does well, what it fails on, where {product} clearly wins. Be ruthless about including "do nothing" as an alternative.4. 独特属性审计
Below is our claimed list of unique attributes for {product}. For each, judge: actually unique (no competitor has it), table-stakes (everyone has it), or marketing claim (cannot be verified). Cut the table-stakes and marketing claims. End with 3 attributes a competitor truly cannot copy in 12 months.
{paste attributes}5. 目标 segment 收窄
For {product}, narrow the target segment 3 times. Each pass cuts who we serve by adding a constraint (stage, industry, role, behavior, frequency). End with a segment where we are obviously the best fit, even if it is much smaller than current targeting.6. 横向产品的纵向定位
Our product is horizontal ({describe}). Recommend 3 verticals to position into. For each: why we win here, what we would emphasize, what we would mute, what landing page changes. Pick the strongest and write the new hero line.7. “从 X 到 Y” 重定位
Frame our repositioning as "from {old category / target / value} to {new}". Output: what changes for the buyer perception, what changes for the team, what stays the same. End with the 3 sentences we use at all-hands to announce it.8. 定位陈述（一段话）
Compress our positioning canvas into a 4-line positioning statement: "For {target}, who {pain}, {product} is a {category} that {key benefit}. Unlike {alternative}, we {differentiator}." Then write a 30-word version and a 10-word version of the same.
Canvas: {paste}9. 销售对话试金石
Stress-test our positioning by simulating a 5-minute sales conversation with a skeptical buyer in {segment}. Buyer asks 6 hard questions ("Why not X?", "Why now?", "Why this category?"). Write the answer for each. Flag any answer that sounds rehearsed or evasive.10. 由定位改 Homepage hero
Given this positioning canvas, write 3 homepage hero variants: (a) category-led, (b) outcome-led, (c) anti-status-quo. Each: 8-12 word headline, 20-30 word subhead, single CTA. Mark which variant fits which stage of company (early, growth, scaling).
Canvas: {paste}11. 定位批评轮
Critique this draft positioning on 7 dimensions from "Obviously Awesome": (1) clear competitive alternatives, (2) defensible unique attributes, (3) value tied to attributes, (4) target market specific enough, (5) category obvious, (6) story coherent, (7) free of buzzwords. Score each 1-5.
{paste positioning draft}12. “隐藏定位”提取
Below are 10 recent sales call transcripts. Infer the positioning the team is ACTUALLY pitching (not the one in the deck). Compare with our stated positioning. Where do they diverge? Recommend either updating the doc or retraining the team.
{paste transcripts}13. 新品类框架练习
We want to create a new category called {name}. For this to work: (1) what existing category does it replace, (2) what does it make obsolete, (3) what 3 attributes only fit the new category, (4) what would a leading analyst (Gartner / Forrester) require to recognize it. Recommend yes / no with reasoning.14. 重定位推进计划
We are repositioning {product}. Produce a rollout plan: internal alignment (week 1), website + materials (weeks 2-4), customer comms (week 5), sales enablement (weeks 5-6), public launch (week 7). For each phase: deliverables, owner, success signal, kill criteria.15. 长期定位 vs 近期切入
For {product}, separate the long-term positioning (where we want to be in 3 years) from the near-term hook (what gets us our first 100 customers). These can diverge. Output a 2-column doc and explain the bridge between them.容易踩的坑
- 把定位当 messaging——定位在上游，很少变；messaging 跟着它。
- 只列竞品、漏掉”什么都不做”和”现在用的 Excel”——这两个赢了最多单。
- 把 SSO、移动端、集成当独特属性——这些是 table-stakes。
- segment 宽到说不出 5 个真客户。
- 只因为某品类搜索量大就进——必须能赢的品类才进。
- 改 homepage 却没改销售话术——内外不一致摧毁可信度。
- 在现有品类还没赢之前就想造新品类。
优化技巧
- 替代选项一定要包含”什么都不做”和”他们现在用的 Excel”。
- 每年刷新一次定位；市场会绕着你动。
- 销售通话转录（模板 12）是真相源——团队说的就是用户听的。
- 模板 9 找一个真的怀疑型买家做，不要只内部演。
- 定位变更必须配 30 天推进——突然变会打乱销售。
- 把 10 字版（模板 8）背下来——leader 背不下来就还不够锐。
- 区分品类和 segment——Slack 在”团队消息”品类，但目标 segment 窄得多。
- 定位锐化之后，价格形状要与信息对齐，可参考上线前用 AI 生成定价假设的工作流——message 和 price 要么互证，要么一起翻车。
FAQ
- 定位和 messaging 差在哪？: 定位是地基：品类、替代、segment、价值。Messaging 是表层措辞。定位很少改，messaging 经常改。
- AI 能替我发明定位吗？: 不能。AI 能结构化和批评，但定位必须来自真实访谈、丢单数据、竞品事实。
- 什么时候该重定位？: 一直输本该赢的单子时、销售描述和 homepage 出现偏差时、竞品 pivot 进你领地时。
- 该不该造新品类？: 年轻公司几乎不要。先在现有品类赢。造品类要花数百万和数年。
- 定位文档应该多长？: 一页够。超过 3 页就不是定位了——那是市场同人小说。