AI 分析
痛点：独立开发者花钱推广前难以判断是产品需求错还是投放渠道错，容易把大量预算浪费在没有紧迫需求的产品上；而面向有明确考试日期人群的细分备考产品已被验证可付费。
目标用户：有明确考试日期的职业/技工考试考生；以及想低成本验证需求、避免烧钱投广告的独立开发者。
证据：Spent £4,000 across every channel and got zero customers, then £25 and got paying customers in nine days. ... a stranger paid full price nine days later, then two more, it's at 78 users and 8 subscribers now on about £25 of ads ... The difference was the second product had people with an exam date who already knew they had a problem
变现：按细分职业考试做订阅制备考产品：题库、模拟考、错题复习、学习计划；先以早鸟价或预售验证付费意愿，再扩到多个证书/考试品类。
MVP：选一个具体职业考试，用AI/Claude Code在2周内做出500-1000道题+模考+考试倒计时+错题本+Stripe订阅；去对应subreddit/论坛发免费样题和备考攻略，落地页直接标价，收早期订阅。
竞品：Quizlet、Anki、UWorld、Kaplan、Trade-specific exam prep apps
摘要：作者花£4,000在全渠道推广家庭回忆App，六个月后只有177注册、0付费；后来用Claude Code几周做出自己行业的考试备考产品，只在相关subreddit发帖，£25广告费，9天内陌生人全价付费，现有78用户、8订阅者。关键差异是第二个产品面对有明确考试日期和已知问题的人群，而不是“以后会回来用”的泛需求产品。