How to Test a Software Startup Idea

Don't

• Don't skip steps. Finish each one before starting the next.
• Don't start building the product. For god's sake.
• And definitely don't form a company yet.

Talk to people first

1. Find 10-15 potential users. Not friends. People who actually have the problem you think you're solving.^[1] Customer interview discipline is a core theme in The Mom Test too.^[2]
2. Ask them about the problem, not your solution. How do they deal with it today? What have they tried? How much does it cost them?
3. Listen more than you talk. You're looking for patterns.

Prep

1. Round up a few advisors—the best you can get quickly.
2. Brand v0.1—register a domain, make a logo. Please don't spend too much time here.
3. Build a landing page with a couple lines about the product and a way to join a waitlist.

Demo video v1

1. Start with the exact words you'd say. Share with your advisors. Iterate until you can't think of how to improve it.
2. Create the screens you'll show during the voiceover. These should be completely faked. Use AI to generate mockups—Figma, screenshots of things that don't exist, whatever gets the idea across fast. Don't build anything real yet.
3. Have someone with a great voice record the voiceover. Not you.
4. Add the demo video to your landing page.

Get out of the building (again)

1. Share the demo with people you don't know—potential users, investors, strangers on the internet.
2. Watch how they react. What confuses them? What excites them? What do they ask about that you didn't include?
3. Iterate on the faked demo until you're confident you know what to build.

Build the smallest thing

1. Build the smallest possible version that delivers the core value. One feature. Maybe two.
2. Swap the fake screens in your demo for real product.
3. Keep sharing with people you don't know. Now you're looking for people who'll actually use it.

Build?

If people are using it—and asking for more—now you can really build.

— Ry