The million-or-billion-dollar question
The answer costs you the most expensive thing you own — time. There are three ways to find it, and only one is both fast and safe.
Your job: turn your one sentence into a testable one-page model — without the blank-page paralysis.
Start an idea (or import a Lean Canvas PDF) Read `business_model://current`. If it's empty, draft a first-pass Lean Canvas from this one sentence — "[your idea here]" — and name the customer segment specifically, never "users." claude mcp add --transport http leanspark https://leanspark.ai/mcp Connect first →
What you get back
Reads your one sentence — or a Lean Canvas you upload as a PDF — and fills the nine boxes into a structured model you can actually test. You bring the idea; LeanSpark does the transcription.
A populated Lean Canvas — problem, customer segments, UVP, revenue — the blueprint the rest of this runbook stress-tests.
Every founder starts in the same place: an idea you’ve been circling — for weeks, maybe for years — and a quiet voice asking, is this the one, or am I about to spend years of my life finding out the hard way?
That question — “is my idea worth pursuing?” — is the million-or-billion-dollar question, because the answer costs you the most expensive thing you own. Not money. Time. The years you put into this are years you don’t get back. So the real question underneath it is: is this worth two or three years of my life?
There are exactly three approaches founders use to answer it. I’ve used all three. Two of them cost me dearly.
Approach one: build first. You’re convinced, so you go build it. Nine to twelve months heads-down, then another six trying to find customers for the thing you already made. By the time you learn whether anyone wanted it, you’ve burned eighteen months and most of your runway. It’s the riskiest, most wasteful approach there is. Life is too short to spend a year building something nobody wants.
Approach two: customer research first. Talk to customers before you build. Weeks of interviews, then an MVP based on what you heard. Genuinely better than approach one — but it has two problems. It’s still slow. And the dangerous one: interviews can hand you a false positive. People are nice. They’ll say they’d “totally use that,” and you’ll walk away certain when all you collected was politeness.
I’ve been there — CloudFire. Early on I built three products nobody wanted. CloudFire taught me the most, because I did everything approach two says to do: I interviewed, I built an MVP, and it worked — real, paying customers. Then I did the math I’d been avoiding. It cost me about twelve dollars to acquire a customer worth about four. Twelve in, four out. Desirable… and completely broke. That’s the trap: you can have a desirable product, real customers, and a model that still doesn’t close.
Approach three: business model first. Before you build, and before you even interview, you lay the whole business model on one page — a Lean Canvas — and stress-test it. You find the cracks while they’re still cheap to fix, on paper, before you’ve spent a single real week on the wrong thing. Designing the model used to take one to two weeks. With LeanSpark it takes one to two hours — which makes approach three faster than customer-first and dramatically safer than build-first. That’s the approach this runbook walks you through.