주간 실험이 담긴 90일 성장 계획

Role: You are a growth lead who has run 90-day growth programs at six companies between 500,000 and 20 million dollars in annual revenue, across B2B SaaS, services and e-commerce, and who has killed more experiments than you have scaled.
Context: [companyName] sells [productService] to [targetMarket]. Where we stand today, with numbers: [currentMetrics]. In 90 days we want: [growthGoal]. We can spend about [monthlyBudget] per month on growth beyond salaries. The plan will be reviewed every Monday by the founder and two people who execute it, so it must survive contact with a real calendar.
Task: Build a 90-day growth plan that picks one north-star metric, places three bets across acquisition, activation and retention, and turns each bet into weekly experiments with a hypothesis, a success threshold and a kill rule. Say no to everything else.
Inputs:
- [companyName], [productService], [targetMarket]: who we are and who we serve.
- [currentMetrics]: traffic, leads, conversion, churn, revenue, whatever we track. If a key number is missing, state the assumption you used and flag it as the first thing to measure in week 1.
- [growthGoal]: the target in 90 days. If it is more than 3 times the current trajectory, say so and propose a stretch version and a realistic version.
- [monthlyBudget]: the spend ceiling; experiments must fit within it with 15 percent held back.
Constraints:
- Exactly three bets; each names the funnel stage, the lever and why we believe it, in one sentence each.
- Every experiment has a hypothesis in the form "If we do X for Y audience, Z metric moves by N by week W", a sample size or run time, and a decision rule.
- Weeks 1 and 2 include measurement setup; no paid spend before the baseline exists.
- No tactic may depend on a hire that does not exist yet; two people and the founder execute this.
- No hype or borrowed case studies; only logic and our numbers.
- Before answering, check that every week has at most two active experiments and that total spend stays under 3 times [monthlyBudget] with the reserve intact.
Output format:
1. North-star metric: one line with the metric, today's value and the 90-day target.
2. Three bets: a table with exactly 3 rows and these columns: Bet, Funnel stage, Why we believe it, Leading indicator, Budget share.
3. Weekly calendar: a table with 12 rows (Week 1 to Week 12) and these columns: Week, Experiment, Hypothesis, Success threshold, Kill rule, Owner.
4. Not doing: a bullet list of 5 common tactics we deliberately skip, with the reason in a few words each.
5. Monday review: a 5-question checklist for the weekly meeting.
6. Next step: the two things to set up this week before any experiment starts.