Find more companies like your ten best customers

What it does

Your best customers are the clearest description of who to sell to next. A lookalike search finds companies that share what makes them good, not just their industry and size. It works when you write down the traits that actually matter, and it fails when you search on industry and headcount alone: that finds your customers’ neighbours, not their twins.

What you need

• Your ten best customers: the ones that pay well, stayed, got results and were easy to work with
• A contact database with company filters, like Apollo or Clay
• Claude or ChatGPT with web access, to read their websites
• A spreadsheet

How it works

1. 1

   Pick your ten best customers

   Best means they pay well, they stayed, they got results and they were easy to work with. Not the biggest logo, and not the one that took a year to close.

2. 2

   Write down what they share

   Prompt for Claude or ChatGPT, with web access on

   Here are my 10 best customers, with their websites and a line on why each one was a great customer:
   [PASTE THE LIST]
   
   Read each website and list what at least 7 of the 10 have in common:
   - what they sell, and to whom
   - how they sell: a sales team, online checkout, or partners
   - their size, by headcount or revenue
   - their business model
   - tools you can see on their site, like a booking widget or a chat tool
   - where they are
   
   For each trait, say how many of the 10 have it, and whether I could filter for it in a contact database, would have to check it by hand, or it looks like a coincidence. Only use what's on their sites. If you can't tell for a company, say so.

3. 3

   Search on the traits a database can filter

   Industry keywords, size band, location, and the tools they use if your database tracks them. Many databases also have a “similar companies” search. Seed it with your best customers, then compare what comes back with your own search before you trust it.

4. 4

   Check the traits you can’t filter

   The traits that only show up on a website, like “runs its own trucks” or “sells through a sales team”, get checked company by company. Do it by hand for a small list, or with AI for a big one: check every company on a list fits.

5. 5

   Keep only close matches, then check 20

   Keep the companies that match most of the traits that matter. Then open 20 at random. If fewer than 8 in 10 look like your best customers, tighten the search and run it again, and write the filters into your sourcing spec.

What good looks like

A list of companies that look like your best customers on what matters, not just on industry and size, with the traits written down so the next search finds the same kind.

Mistakes to skip

• Searching on industry and size alone. That finds your customers’ neighbours, not their twins.
• Picking your ten biggest logos instead of your ten best customers.
• Trusting a “similar companies” button without checking what it matched on.