←Back to Home
Reading time: 5 min
Day 5

Day 5: 90-Day Experiment

OPC 7-Day One-Person Company - Day 5: MVP drawn but don't know how to verify? Minimum Viable Experiment + Conversion Funnel + Failure Signals + Control Group

Day 5: 90-Day Experiment

Xu Zhixing finished the lean-canvas but freezes: "How do I verify it?"

Day 5 Overview

Day 5 shifts from wall to practical. The 5-beat formula:

  • Problem scene: lean-canvas done, but no validation plan
  • Zhou Wangpu's method: 4 elements — Minimum Viable Experiment (MVE) + Conversion Funnel + Failure Signals + Control Group
  • Mo Xiaozhai demo: Generates A/B test template but makes 1 error in the control group (sets B at 49 vs A at 99) — Xu Zhixing catches and corrects
  • Deliverable: experiment-design.md (90-day rhythm: 30 days MVE / 30 days iterate / 30 days scale)
  • Next day: Run the experiment + prepare launch

Key Quote

"4 elements. Missing 1 and the experiment goes off-track." (4 个元素，缺 1 个实验就跑偏。) — Zhou Wangpu

"I set the control group wrong in 2008 too — B was half the price of A. B 'won' — but the reason was price, not quality. 3 months wasted." (我 2008 年对照组也设错了——B 组比 A 组便宜一半——90 天后 B 组赢——但原因是价格不是质量——白白浪费 3 个月。)

Today's 3 Takeaways (placeholder for full translation)

  1. MVE (Minimum Viable Experiment): the smallest test that can validate or kill the hypothesis
  2. Conversion Funnel: from awareness → interest → trial → payment, where's the leak?
  3. Failure Signals (Red Light Rules): pre-define what makes you stop (e.g., conversion rate < 0.5% after 1000 visitors)

Recommended Tool

  • VWO + Optimizely + 吆喝科技 (AppAdhoc) — A/B testing platforms

For full Chinese content, see: content/days/day5.md