The intro call: what I ask before we build anything

Every build I take on starts with a thirty-minute call. People assume it’s a sales call. It’s closer to a diagnostic — and since the questions aren’t secrets, publishing them helps everyone: you can run the exercise on yourself before we ever speak.

Here’s what I ask, and what each question is actually testing.

“Walk me through the problem — without mentioning the solution.”

The hardest instruction on the call. Most people arrive in love with a solution (“an app that…”) and have to be walked backwards to the problem it supposedly solves. If the problem can’t be described without the solution, that’s my first yellow flag — it usually means the idea started from wouldn’t it be cool rather than this keeps happening and it hurts.

“Who has this problem, and how do they deal with it today?”

Two tests in one. Naming a specific person — “freelance accountants with 30+ clients”, not “businesses” — tells me discovery has started. And the “today” part is load-bearing: if people currently solve the problem with a spreadsheet, WhatsApp, and swearing, that’s good news — it proves the problem is real and shows exactly what the software must beat. If they solve it with nothing at all, I ask why they’d start with your app.

“What does ‘working’ look like in 90 days?”

Not the five-year vision — the smallest observable success. Ten paying users? One workflow that saves your own company four hours a week? This question surfaces whether we’re building a falsifiable bet or a monument. Answers with a number in them make me happy. “We’ll see traction” does not.

“What exists already?”

Domain, brand, an audience, a waitlist, data, a spreadsheet that is the current product — existing assets change the build order completely. Someone with a newsletter of 2,000 readers in their niche needs different first software than someone starting from zero. This is also where I check who owns what exists — untangling account ownership later is misery best avoided.

“Who else says no?”

Co-founders, spouses, investors, the accountant who “just needs to approve it”. Software gets built for the person on the call and killed by the person who wasn’t. I’d rather meet the veto in week zero.

What I’m doing with all this

By minute twenty I’m sketching two things silently: the smallest scope that would teach us the most, and what the fleet could ship in the first two weeks. If those two pictures overlap, I say so and we talk next steps. If they don’t — wrong stage, wrong economics, wrong me — I say that too, and point somewhere better. About a third of calls end that way, on purpose.

Worst case, you leave with sharper questions than you came with. That was the promise on the page, and it’s the cheapest product work you’ll ever do: thirty minutes, before a single line of code exists.