AI Guides › Money & Business
The Business Idea Test: Running It Past An Evidence Ledger, Not A Hunch
By Nigel Guy · 2 min read
Most business ideas get evaluated on excitement — does this feel like a good
idea, does the person you told seem enthusiastic. Excitement is a terrible
filter. It correlates with novelty far more than with whether anyone will
actually pay for the thing.
The rule: before you build anything, sort what you actually know about the
idea into fact, assumption, unknown, and constraint — most "obviously good"
ideas turn out to be built almost entirely on assumptions once you do this
honestly.
The mechanism
- List everything you believe about the idea's viability.
- Label each one: a fact you've verified (someone has actually paid
for something like this), an assumption you believe but haven't
checked (people would probably pay for this), an unknown you'd need
to find out, and a constraint that bounds what you can even attempt
(budget, time, skills available).
- Count the facts. Most first-pass ideas have close to zero — that's
normal, not disqualifying, but it changes what the next step should be.
- Turn the biggest assumption into the smallest possible test. Not
"build the product" — the cheapest, fastest way to get a real signal
about whether the core assumption holds.
- Set your threshold before you run the test, not after. Decide what
result means "continue" versus "stop" while you're still neutral about
the outcome.
What to skip
Skip building anything substantial before the riskiest assumption has been
tested at all. And skip treating enthusiasm from friends or family as
evidence — they're answering "do I like you," not "would I pay for this,"
and those are different questions even when they feel similar.
Guardrails
- This is a lightweight, personal-use version of a more rigorous practice —
bigger bets deserve a fuller process, more than one person's perspective,
and a properly designed test, not just a solo labelled list.
- A fact today can become stale — markets move, and "verified six months
ago" isn't the same claim as "verified now" for anything time-sensitive.
- Honest labelling is the entire value of this exercise. Calling an
assumption a fact because you want it to be true defeats the purpose
before you've started.
All 751 AI guides · JulieMango plans from £17/mo