Strategy
How to Clone Your Best Customers : GTM Playbook
Most teams rebuild their ICP from scratch every quarter. New list, new filters, new debate about which industries to target. Meanwhile, the answer has been sitting in their CRM the entire time. The companies already paying you are the only ICP that matters. They converted. They paid. They stayed. That's a stronger buying signal than any demographic filter you'll ever apply to a database. Here's the system to turn your closed won deals into a repeatable prospecting engine.
Step 1: Pull your closed won domains from the CRM
Export your last 30 to 50 closed won deals from HubSpot, Salesforce, or Pipedrive. Domains only, nothing else. Don't overthink the sample size. Every company on that list already went through your full sales process and made it out the other side as a paying customer. That's the seed.
Step 2: Find lookalike accounts
Drop those seed domains into CompanyEnrich. Their similar companies API runs vector search across more than 30 million companies and finds accounts that match the same profile as your closed won customers. A list of 10 seed domains returns roughly 500 ranked matches. You don't have to guess what good looks like because your closed won data already defined it.
Step 3: Layer in buying signals
A lookalike account is worth outreach. A lookalike account with an active buying signal is worth prioritizing above everything else on your list.
Four signals worth tracking at this stage: a new funding round, new leadership in the buyer seat, a hiring spike in target roles, and a tech stack change that signals intent for a tool adjacent to yours. CompanyEnrich surfaces these on the same accounts you just found. These signals also become the angle for your outreach copy, because nothing performs worse than generic personalization and nothing performs better than a relevant, well timed reason to reach out.
Step 4: Score and tier every account with AI
Feed each account through a classification prompt in Claude, using your closed won data and the signal layer as the scoring rubric. The output is a clean three tier structure.
Tier 1 is strong fit with an active signal. These are the accounts to move on immediately. Tier 2 is strong fit but no active signal yet. These go into a standard sequence now and move up the list when a signal fires. Tier 3 is adjacent fit, accounts that don't match your ICP perfectly but are worth watching as the market evolves. Clay can handle the same classification workflow if you prefer to keep everything inside one tool.
What used to take hours of manual sorting now takes minutes.
Step 5: Enrich the contacts
Once the accounts are scored, source the right contacts at each one. CompanyEnrich runs people search across more than 170 million profiles and finds the titles you care about. Then waterfall Prospeo and FullEnrich to verify emails. Waterfall logic means each tool picks up what the previous one missed, which keeps your deliverability high and your bounce rate low.
Step 6: Personalize outreach by tier
Channel intensity and personalization depth should match the tier. Don't spend an hour writing a custom email for every Tier 3 account, and don't run Tier 1 accounts through a generic mass email sequence.
Tier 1 accounts get one custom prompt per account, written by Claude or Clay using the specific signal as context. These accounts also get the full multichannel treatment: cold call, LinkedIn outreach, and email running in parallel through Nooks, lemlist, and Instantly.
Tier 2 gets templated messaging with one meaningful variable swap, usually the signal or a relevant firmographic detail. Channel is LinkedIn plus email through lemlist and Instantly.
Tier 3 goes into high volume email sequences through Instantly and nothing else until they show more intent.
Step 7: Sync everything back to the CRM
Every reply, every meeting booked, and every closed won deal feeds back into HubSpot, Salesforce, or Pipedrive. The next time you run this system, the input data is cleaner, the lookalike matches are more refined, and your scoring model gets sharper.
This is where the compounding starts. The system gets better with every cycle because your CRM gets richer with every deal.
Why this beats building from scratch
When you build a list from filters and database exports, you're guessing. You're picking industries and company sizes that feel right based on intuition or incomplete data. When you build from closed won, you're using evidence. Real companies, real deals, real buying behavior.
The 30 to 50 deals already in your CRM are the ICP. Everything after that is execution.
Stop redefining who your customer is and start finding more of them.
