How to tell if a business idea is any good, before you spend money
Last updated: November 7, 2026
Written by the Zylar team. How we research and check our guides
The most honest test of a business idea is whether a stranger will pay for it before it exists. Before you get there, check five things: who has the problem, who already solves it, whether you can deliver it at a price that works, what could break it, and what you would need to prove first.
Why check first
Of the new US business locations that opened in the year to March 2021, 79% made it through their first year, but only 47% were still open five years later (US Bureau of Labor Statistics). When CB Insights studied 385 failed startups, 70% had run out of money and 43% had poor product-market fit: too few people wanted what they built. Running out of money is usually the last thing that happens. Checking demand first is the cheap part.
Built it first. Sam buys a domain, a logo and 500 branded leashes before checking his dog walking idea, and launch day brings zero bookings. Take two: Zylar's AI agents tell him to prove it first with three doorman referrals, and two weeks later he has three bookings. Concept demo with a hypothetical idea.
Five checks you can do this week
| Check | How to do it yourself | Warning sign |
|---|---|---|
| Who has the problem? | Talk to five people who have it. Ask about the last time it happened, not whether they'd buy. | They call it mildly annoying |
| Who already solves it? | Search Google Maps, the app stores and forums. List five options and their prices. | You find nothing at all, or many identical cheap ones |
| Can you deliver it at a price that works? | Take the price per customer, minus what it costs to deliver, including your own hours. | It only works if your time is free |
| What could break it? | List the licences, insurance, seasons and single platforms or suppliers you'd depend on. | One risk you can't control decides everything |
| Will anyone pay now? | Ask for a deposit, a pre-order or a signed letter of intent. | Plenty of "I'd totally buy that" and no money |
Planning a January start?
US business applications jump every January. Over the last four years they rose between 27% and 83% from December to January (US Census Bureau). If you're one of them, the next seven weeks are the time to run these checks.
Questions
Is a friend saying "I'd buy that" a good sign? Not much. Saying yes costs them nothing. A deposit from someone who doesn't know you tells you far more.
How many people should I talk to? Start with five who have the problem. Keep going until the answers start repeating.
Can AI tell me if my idea is good? AI can gather the evidence quickly: the market, the competitors and the risks. It can't tell you whether people will pay. That's what the last check is for.
Zylar is being built to run the first four checks with AI agents that cite their sources and score the idea out of 100, then tell you what to prove next. It's in early alpha. Join the waitlist.