The 10-Client Test

Most B2B firms that grew on referrals can’t say, with evidence, why their best clients bought. The referrer did the positioning and the vouch did the selling. That works until you try to scale. Then every new channel amplifies a go-to-market nobody wrote down.

The 10-Client Test is the check to run before you spend on growth. Take your last ten won clients and answer six questions for each:

1. Who they were, and what made them start looking

2. Why they chose you, in their words

3. How they found you

4. What they paid

5. What they bought

6. Who closed it, and whether anyone besides the founder could have

The skill does the digging. It finds your recent closed-won deals in your CRM, pulls the first sales call for each from your call recorder, and fills in the table with a source on every cell. Where it can’t find evidence, the cell says “no evidence.” It never guesses.

Then it reads the table for you. Each column comes back as validated, mixed, or a guess. You get the growth stage that pattern usually points to, and one concrete test for your weakest column: a message tried on a narrow buyer, a price tried on real proposals, or one channel run long enough to read.

What you get

• A filled ten-client table, every cell sourced to a call or a CRM field

• A column-by-column verdict: validated, mixed, or guess

• The one thing to prove next, and how to test it

Your data stays in your Claude. The skill runs inside your own Claude conversation. It doesn’t send, post, or upload anything, and we never see your transcripts or CRM.

No recordings? It still works. Upload transcript files or a CRM export, or type in your ten clients and answer the questions yourself. The table and the readout work the same way.