When to use it
Use this when targeting has gone soft: reps are working anyone with a pulse, campaigns are spraying, and win rates are sliding because you're selling to accounts you can't actually serve. It earns its keep at the start of a planning cycle, when entering a new segment, when win rates drop, or when onboarding reps who need a crisp picture of who you win with and who quietly churns. The output is only as good as the honesty of your inputs, feed it real lost-deal and churn reasons, not your pitch deck.
Do NOT use it to invent an ICP from scratch with no customer data; it will produce a confident, plausible, and useless fiction. If you have fewer than ~10 closed deals, you don't have an ICP yet, you have a hypothesis, and you should label it that way. Also do not treat the output as permanent; ICPs drift as the product, pricing, and market move, so this is a quarterly exercise, not a one-time artifact.
The principle: your ICP is defined more by who you reject than who you chase. Any team can list dream-customer traits; the teams that win are ruthless about disqualifiers, the signals that an account will churn, discount you to death, or demand a build you can't staff. This prompt forces that discipline and converts it into a scoring checklist a rep can run in 30 seconds.
The prompt
You are a GTM strategist who builds ICPs from evidence, not aspiration. You are skeptical of dream-customer fantasies and you treat disqualifiers as the most valuable output. You never invent customer names, stats, or market data; you reason only from what I provide and flag every inference.
CONTEXT
Product + core value: {{PRODUCT_AND_VALUE}}
Our best customers (describe industry, size, motion, why they're great, names optional): {{BEST_CUSTOMERS}}
Deals we lost or churned, and the real reason: {{BAD_FIT_NOTES}}
Market / segment we serve: {{SEGMENT}}
Pricing / deal-size context (ACV range, pricing model): {{PRICING_CONTEXT}}
TASK
Produce a usable ICP, a buying-committee map, and a scoring model, sharp enough that a new rep could disqualify a bad account from a single discovery call.
METHOD
1. Find the PATTERN across the best customers (what they share beyond surface firmographics, the operating conditions that make us win).
2. Find the ANTI-PATTERN across lost/churned deals (the conditions that predict failure). This is the source of your disqualifiers.
3. Translate patterns into observable, checkable signals, not vibes. 'Has a dedicated CS team of 3+' beats 'cares about customers'.
4. Map the buying committee and name who blocks deals, not just who signs them.
5. Build a scoring checklist a rep can actually run.
OUTPUT FORMAT
PART A, IDEAL CUSTOMER PROFILE (account-level)
- Firmographics: industry, employee count, revenue band, geography.
- Environment / technographics: tools, motions, or conditions that signal fit.
- Trigger events (4-6): what makes an account need us NOW.
- Hard disqualifiers (4-6): specific signals an account is NOT a fit. Be concrete and a little brutal.
- One-sentence ICP statement.
PART B, BUYING COMMITTEE (2-3 personas)
For each: title/role, what they own, top 3 pains, what success looks like to them, what they fear, where they research, and the single message that lands. Label each as ECONOMIC BUYER, CHAMPION, or LIKELY BLOCKER. Note who is most often the silent deal-killer.
PART C, FIT SCORING CHECKLIST
5-7 yes/no questions a rep can ask early, with a simple rule: how many 'yes' = good / medium / poor fit, and the one answer that should auto-disqualify regardless of the rest.
CONSTRAINTS
- Tag every inference (inferred). Never fabricate logos, stats, or funding.
- No hype language. Keep it skimmable.
- If my BAD_FIT_NOTES are thin or missing, say the disqualifiers will be weak and ask me for two real lost-deal reasons before finalizing.
Run it from the terminal
Variables to swap
| Variable | What to put there |
|---|---|
{{PRODUCT_AND_VALUE}} | What you sell and the core outcome it delivers |
{{BEST_CUSTOMERS}} | 3-5 of your best customers, industry, size, motion, and why they're great fits |
{{BAD_FIT_NOTES}} | Real patterns from lost or churned deals and the honest reason they failed |
{{SEGMENT}} | The market or segment you serve (e.g. mid-market B2B SaaS in North America) |
{{PRICING_CONTEXT}} | ACV range and pricing model, shapes which accounts can actually afford you |
Worked example
Input
PRODUCT_AND_VALUE: A customer onboarding platform that cuts time-to-first-value for B2B SaaS
BEST_CUSTOMERS: Fast-growing Series B/C SaaS, 100-500 employees, PLG + sales-assist motion, dedicated CS team of 4-8 feeling onboarding pain
BAD_FIT_NOTES: Tiny startups with no CS owner churned within 3 months; large enterprises wanted bespoke builds we couldn't justify; one services-led shop never adopted because onboarding was human-led by design
SEGMENT: B2B SaaS, North America + UK
PRICING_CONTEXT: $18k-$45k ACV, per-seat + platform fee, annual contracts
Output
PART A, IDEAL CUSTOMER PROFILE
- Firmographics: B2B SaaS, 100-500 employees, ~$15M-$60M ARR (inferred from headcount + ACV fit), US/UK.
- Environment: PLG or hybrid motion, dedicated CS team of 3+, already running a CRM + product analytics, onboarding currently manual or spreadsheet-driven.
- Trigger events: new VP/Head of CS hired; recent funding round; rising first-90-day churn; launching a new product tier or self-serve motion; CS headcount growing faster than process.
- Hard disqualifiers: fewer than ~50 employees or no dedicated CS owner; services-led business where onboarding is intentionally human; enterprise demanding bespoke builds; flat or declining growth (no urgency); no single owner for onboarding (no one to champion or adopt).
- ICP statement: Growth-stage B2B SaaS companies with a dedicated CS team feeling the strain of onboarding that won't scale.
PART B, BUYING COMMITTEE
1. VP / Head of Customer Success, ECONOMIC BUYER. Owns retention + NRR. Pains: inconsistent onboarding, no visibility into account progress, can't scale without hiring. Success: faster time-to-value, flat or improving NRR while growing. Fear: buying a tool the team won't adopt. Researches: peer Slack groups, CS communities, analyst content. Message: 'Standardize onboarding without adding headcount.'
2. CS Ops / Onboarding Manager, CHAMPION. Owns the playbook daily. Pains: manual busywork, tool sprawl, no single source of truth. Success: fewer manual steps, clear visibility per account. Fear: more admin, not less. Message: 'Cut the busywork and see every account's progress in one place.'
3. CFO / Finance, LIKELY BLOCKER and most common silent deal-killer. Pains: tool sprawl and unproven ROI. Fear: another seat-based tool nobody uses. Message: 'Tie onboarding speed directly to retained revenue.'
PART C, FIT SCORING CHECKLIST
1. 100-500 employees? 2. Dedicated CS team of 3+? 3. Growing (funded/hiring in last 18 mo)? 4. A single named owner for onboarding? 5. Standard (not bespoke) needs? 6. Onboarding currently manual/spreadsheet-based? 7. SaaS product (not services-led)?
Rule: 6-7 yes = good fit; 4-5 = medium (qualify hard); <4 = poor fit. AUTO-DISQUALIFY: no dedicated CS owner, historically our fastest churn.
Tips to get more out of it
- The disqualifiers section is only as good as your honesty about churn, paste real lost-deal reasons, not sanitized CRM dropdowns. If you only have CRM stage-loss data, dig up two actual stories first.
- Pressure-test the output with a follow-up: 'Which disqualifier do reps most often ignore in the heat of a quarter, and why?' That surfaces the trap your team keeps walking into.
- Operationalize the scoring checklist or it dies in a doc, turn it into a CRM picklist, a lead-scoring rule, or a required field on the opportunity, so it actually gates pipeline.
- Re-run quarterly with new closed-won and closed-lost data; an ICP defined a year ago is selling to a company you no longer are.
- Pair each persona's 'single message' with the cold-email and ad prompts so messaging stays consistent from first touch to demo.
- If the model produces a generic persona ('busy decision-maker who values efficiency'), it's working from too little, reject it and feed more specific best-customer detail.
- Ask it to separate 'must-have' firmographics from 'nice-to-have', reps over-index on the wrong traits (industry) and ignore the predictive ones (a named onboarding owner).