The go-to-market assumption portfolio
A plan to find your next customers is a stack of assumptions about who buys, why, and what will reach them. This method borrows the discipline a regulated device team uses to rank risk: score every assumption on the chance it’s wrong and the damage if it is, then spend two-week sprints testing the top ones.
How it works
1. List every assumption. Write out everything the plan to find your next customers depends on: the buyer, the problem, the channel, the proof. Nothing is too obvious to write down.
2. Score each one. Rate the chance it’s wrong and the impact if it is, 1 to 5 each, and multiply. The product is the assumption’s score.
3. Test the top ones. Run two-week sprints against the highest scores. Test speed, how many days until a test produces data, breaks ties.
4. Move the scores with what you learned. After each sprint the numbers change, and only with evidence.
It runs on LoopGrow too
I keep my own risk file and run the method on my own business. Two rows from it:
Settled, 7 October 2026. The assumption: “Enterprise buyers’ problems can’t be mined from public sources, so customer development starts from interviews alone.” It was my own working assumption. I’d never heard of a business-to-business executive complaining on Reddit about their most pressing problem. The test: agents searched one market, water and wastewater utilities in the United States and Canada, across five kinds of public source. They found 31 usable statements of the buyer’s problem, each re-checked word for word at its source, in annual reports, public tenders and the council reports behind them, job postings, and talks and podcasts where utility general managers speak. 55 statements were in the voice of a buyer, an executive or the organisation itself. Reddit couldn’t be tested, because it refused automated access. The assumption failed the rule I set before the test, so the free kit now starts by gathering what buyers already say in public, quoted with links.
Open. The assumption: “Founders in tough markets recognise their go-to-market as an engineering problem, and will pay to run it that way.” Impact 5, chance it’s wrong 3, score 15. The test: ten conversations framed around a risk file, counting how many founders ask for the matrix.
Why rank before you buy
There’s no menu. A website, a short founder video, an outreach list, case studies from customers 3 to 10: each one gets built only when it tests an assumption near the top of your list. You pick what you need, and the ranking is how we check that you need it.
For how this compares with hiring, agencies and building it yourself, see your options.
Where to start
Take the qualified prospects scorecard: 12 multiple-choice questions, about 2 minutes. Your results come with the free kit, which runs this method in your own artificial intelligence (AI) assistant, and the option to book a working session built around your answers.