ValidationJuly 30, 2026By Maarten Follaets

How to choose between two startup ideas

Dark slide with the headline Two ideas, let the evidence pick, and a callout box noting 1,000 paying customers is the floor for a business, sourced to Jason Cohen at A Smart Bear.

Quick answer: When you are stuck between two startup ideas, do not pick the one that excites you more. Excitement is not evidence, and you are biased toward whichever you thought of most recently. Instead, run the same cheap test on both and score each on four things: is the problem real, is the market reachable, will people actually pay, and can you differentiate. The idea with the stronger evidence wins, even if it is the less exciting one. And watch for convergence: as you talk to more people, the better idea gets clearer and the pain and price line up, while the weaker one gets fuzzier the more you dig.

"Which one excites me more" is the wrong question

You have two ideas. You have gone back and forth for weeks. Both feel good on different days.

So you try to decide by feel, and feel keeps flipping.

Here is the trap. Excitement tracks novelty, not viability. The idea you thought of last week feels more alive simply because it is newer. Ask yourself which one you are more excited about and you will get a different answer depending on the morning. That is not a decision method. That is a mood.

The founders who choose well do not ask which idea they love. They ask which idea has the better evidence behind it. I have flip-flopped on this myself, and the times it went wrong were always the times I let the fun of an idea outvote the facts.

Solving a problem is not the bar

The most common mistake is to pick the idea that solves the more interesting problem. But solving a problem is not the same as having a business.

Jason Cohen makes this point hard in his essay Excuse me, is there a problem?. Plenty of startups identify a real problem, build a product that genuinely solves it, and still die. As he puts it, solving a problem is "not nearly enough to build a successful company".

So when you compare two ideas, do not stop at "does it solve a real problem". Both might. Score each on the things that turn a solved problem into a company:

Real problem: do specific people already try to solve this today, with time or money? A problem people work around is real. A problem people shrug at is not.
Reachable market: can you get in front of these people affordably? Cohen points out you need on the order of 1,000 paying customers for even a small, sustainable business, which means you need a channel to reach thousands, not a vague "everyone".
Willingness to pay: will they actually hand over money, not just praise? This is usually where one of your two ideas quietly falls apart. I wrote more on that in is your SaaS or app idea worth pursuing.
Differentiation: is there a reason to pick you over what exists? Being 10 percent better at something 10 percent of the market cares about is not a wedge.

Give each idea an honest read on all four. Often one idea wins three of them and the "exciting" one wins zero. That is your answer.

The convergence test

Scoring on paper only gets you so far, because you will grade your favourite generously. So put both ideas in front of real people and watch what happens as you dig.

Cohen describes the tell from his own validation of two ventures. With the idea that became a real business, the experience was convergent: the more people he spoke with, the more they agreed on the pain, on whether the solution was acceptable, and on what they would pay. With the weaker idea, the "truth" diverged the more he looked, every conversation pulling in a different direction.

That is the test you can run this week. Talk to ten potential customers for each idea. Ask about their life and their past behaviour, not your pitch. Then look at the shape of the answers.

The idea where people keep describing the same pain, in the same words, willing to pay in the same range, is converging. Back it.

The idea where every conversation adds a new caveat, a new "well, it depends", a new segment, is diverging. That fuzziness is not you needing more research. It is the market telling you the problem is not sharp enough yet.

I dig into how to read those conversations honestly in how do I know if my startup idea is good.

The catch: you grade your favourite on a curve

The reason this is hard is not the framework. It is you.

You will score your favourite idea's problem as more real, its market as more reachable, its buyers as more willing, because you want it to win. Every founder does. The bias does not go away because you read a blog post about it.

This is where an objective second read earns its place. Ventropolis, powered by Foxy, is built to be the check you cannot be for yourself: put both ideas in, and it presses on each with the same cold questions, so the one that wins wins on evidence rather than on which one you walked in loving. Not a magic button, just the referee you cannot be for your own two ideas. You can pressure-test both at Ventropolis for startups.

Pick the same cheap test. Run it on both. Score problem, reach, payment, and difference. Then let the evidence, not the mood, cast the deciding vote.

So which of your two ideas would you still choose if you had to bet on the numbers instead of the feeling?

Sources

Why solving a real problem is not enough to make a company, the rough 1,000-customer floor, and the convergent-versus-divergent signal when validating: Jason Cohen, Excuse me, is there a problem?.

Frequently asked questions

How do I choose between two startup ideas?

Do not choose by which excites you more. Run the same cheap test on both and score each on four things: is the problem real, is the market reachable, will people pay, and can you differentiate. Pick the one with the stronger evidence, not the stronger feeling.

Should I pick the idea with the bigger market?

Not on its own. A big market on a slide says the prize could be large. It says nothing about whether anyone wants your specific product or whether you can reach them affordably. A smaller market you can actually win beats a huge one you cannot.

What if I am equally excited about both?

Good, then excitement is not the tie-breaker and you can let evidence decide. Talk to ten potential customers for each. The idea where the pain and the price converge as you dig is the stronger bet. The one that gets fuzzier is the weaker one.

How long should choosing take?

Days to a couple of weeks, not months. The whole point is to spend a little evidence to avoid spending months building the wrong one. If you cannot get a signal cheaply, that itself is a signal the market is hard to reach.

What if both ideas look weak after testing?

Then testing did its job. You just saved yourself months. Kill both, keep the customer relationships you built, and use what you learned about the problem to find a better idea in the same space.

Is it bad to work on two ideas at once?

For a short evaluation, no. Running a cheap parallel test is fine and even useful for comparison. For building, no. Once you have picked, commit fully. Splitting build effort across two products is how you get two half-products nobody wants.