Build a 90-Day Growth Plan with AI

1. Go to claude.ai and start a new chat (Sonnet is fine). 2. Paste this prompt (fill in the brackets): "I'm building a 90-day growth plan for my business. Help me lock ONE North Star metric. Context: - What I sell: [product or service in 1 sentence] - Stage: [pre-revenue / early / scaling] - Current monthly numbers: revenue [$], paying users [#], leads/week [#], traffic/mo [#] - Why I want growth in the next 90 days: [reason / event / runway] Do this in order: 1. Suggest 3 candidate North Star metrics that are leading enough to move in 90 days but tied to revenue. 2. For each, say what it predicts and what it ignores. 3. Pick ONE and write it as: 'By Day 90, [metric] goes from [X] to [Y]'. 4. Sanity-check the target: is it 1.5–3x baseline, or am I sandbagging / fantasizing?" 3. Pick the metric. Reject any vanity picks (followers, impressions, list size) unless they're a confirmed input to revenue. 4. Open a fresh doc and paste the final one-line goal at the top — everything else this guide produces will live under it.

1. In the same Claude chat, paste: "Now diagnose where my funnel is weakest. Rough numbers per stage are fine. - Acquisition (visitors / leads per month): [#] - Activation (signed up or completed first key action): [#] - Retention (still active week 4 or month 2): [#] - Revenue (paying or qualified): [#] - Referral (came from word of mouth or share): [#] Tasks: 1. Compute conversion % between each stage. 2. Tell me the ONE stage with the lowest conversion and the ONE stage where a small fix would move my North Star most. They might be different — say which is the bigger lever and why. 3. Give me a one-paragraph hypothesis for *why* that stage leaks based on what you can infer about my offer. 4. List 3 questions I should answer (talk to users, check analytics) before I commit to fixing it." 2. Read the answer. If Claude picked the wrong lever (you know your business better), tell it why and ask it to redo. 3. Add to your doc: 'Lever: [stage]. Hypothesis: [one line].'

1. Same chat, paste: "Generate 8–12 candidate growth bets that would move [chosen lever from Step 2] over 90 days. Mix channel work, product/onboarding fixes, and lifecycle/email plays — not just paid ads. For each bet: - Hypothesis: 'If we [do X], then [North Star metric] moves because [why]' - Impact (1–10): how much it moves the North Star if it works - Confidence (1–10): how sure I am it works given my context - Ease (1–10): how doable solo in 30–90 days - ICE score: average of the three Then: 1. Rank by ICE. 2. Recommend the top 3 — but force diversity (don't recommend three flavors of the same thing). 3. For the top 3, write a one-paragraph 'what success looks like' description." 2. Read the top 3. Veto any that secretly need money or a team you don't have — ask Claude to swap. 3. Add to your doc: 'Bet 1 / Bet 2 / Bet 3' with the hypothesis line for each.

1. Same chat, paste: "For my 3 bets, build a 30/60/90 milestone plan. Rules: - Each month gets ONE theme (e.g. 'M1: foundations', 'M2: distribution', 'M3: optimize') — pick what fits. - Per bet, list 2–3 milestones per month. Milestones are *deliverables shipped*, not activities. - Earlier months should reduce risk for later months (research/build before scale). - Flag any milestone that depends on someone else (designer, customer interview, integration) — those need lead time. Output a markdown table: Month | Theme | Bet 1 milestones | Bet 2 milestones | Bet 3 milestones | Risk/dependency notes" 2. Read the table. Cut any month that has more than 6 total milestones across bets — that's a wishlist, not a plan. 3. Add the table to your doc under '30/60/90 plan'.

1. Same chat, paste: "For Month 1 only, give me a weekly plan for Weeks 1–4. Per week: - Theme (one line) - Actions (5 or fewer total across all 3 bets — be ruthless) - Leading KPI to watch this week (an activity or input metric, e.g. 'demo calls booked', 'onboarding emails sent') - Reflection question for end of week Also: - Define 1 lagging KPI = my North Star (tracked monthly). - Define 2–3 leading KPIs that should predict it (tracked weekly). Output as markdown: a 4-row weekly table + a small KPI block underneath." 2. If any week has more than 5 actions, ask: 'Cut to 5. Tell me what got dropped and why.' 3. Add to your doc under 'Month 1 weekly plan' + 'KPI dashboard'. 4. Re-run this prompt at the end of Month 1 (for Month 2) and end of Month 2 (for Month 3). Don't pre-plan all 12 weeks now — you'll waste it once reality hits.

1. Open Google Sheets (or Notion if you live there) and create a doc titled '[Business] — 90-Day Growth Plan — [start date]'. 2. Create these tabs/sections: - **North Star** — the one-line goal + baseline + target - **Bets** — your 3 bets with hypothesis (you are the owner) - **30/60/90** — the milestones table from Step 4 - **Month 1 weekly plan** — the table from Step 5 - **KPI dashboard** — lagging + leading metrics, one column per week, fill weekly - **Review log** — date, what shipped, what slipped, lesson 3. Paste the content from your working doc into each section. 4. Block 2 recurring calendar events: - **Friday Growth Review — 15 min** (every Friday): update the KPI dashboard, log what shipped vs slipped, decide if next week's plan needs a cut. - **Month-End Retro — 30 min** (last weekday of each month): re-run Step 5's prompt for the next month using your actual KPI data. 5. Pin the doc somewhere you'll see daily — browser bookmark, Notion sidebar, Slack saved messages.