ShipDesk.devblog
the blog

How to validate a startup idea before you build it

George · 4 min read

An idea is not validated when people say they like it. It is validated when people who have the problem give up something for a solution: money, a deposit, or real time. Everything before that is a hypothesis.

This guide is for the stage before the code. If you already have a working product, our posts on cutting your MVP down and finding your first 10 users pick up from there.

Write the idea as a problem, with a cost

Start with one sentence that names a group, a problem and what the problem costs them today. Use a real example as you read: "freelance designers lose a few hours a month chasing clients who pay late, and some invoices are never paid."

Compare that with "an invoicing app with AI reminders". The second sentence describes your solution. The first one describes something you can check: do freelance designers actually chase late payments, and does it cost them enough to care?

If you cannot name the group or the cost, the idea is not ready to test yet. That is useful to learn on day one.

Find people who have the problem, not people who fit a profile

A job title is not evidence. "Freelance designer" includes people who are always paid on time. Look instead for people describing the problem in their own words: posts asking how to handle a client who will not pay, threads comparing invoice tools, complaints about awkward follow-up emails.

Search the communities where the group already talks, and search for the problem, not your product category. Keep a simple sheet: where you found the person, what they said, and what they do about it today.

If two weeks of looking turns up almost nobody describing the problem, you have learned something important without building anything.

Ask about the past, not the future

People are generous about ideas and accurate about their own week. So ask what happened, not what they would do:

  • "When did a client last pay you late? What did you do?"
  • "What have you tried to stop it happening? What did that cost?"
  • "What is the most annoying part of the way you handle it now?"

Avoid "would you use an app that...". Almost everyone says yes to that, and it predicts very little.

Listen for workarounds. A designer who keeps a spreadsheet of unpaid invoices and a folder of reminder email drafts has already shown the problem is worth effort. A designer who shrugs and says "it sorts itself out" has shown the opposite, however polite they are about your idea.

Ask for a commitment

Once you can describe the problem the way they do, ask for something real. Choose the size to fit the product:

  • A pre-order or a paid pilot for a product people would pay for.
  • A small deposit for early access, refundable if you never ship.
  • A signup on a one-page site that says plainly what it will do and what it will cost.

The one-page site does not need to be good. A headline in their words, three lines about how it works, a price and a button are enough. What matters is that a person who reads the price still signs up.

Be honest about where things stand. "This is not built yet; a deposit reserves a spot and is refunded if I do not ship by March" is a fair offer. Pretending a product exists when it does not is not.

Decide with a number you set before you started

Write down what counts as a yes before you begin, then keep to it. For the invoice example: "within two weeks, five freelance designers pay a $10 deposit." Pick a number that would genuinely change what you do next.

If you set the bar after seeing the results, every idea passes. That is how founders spend six months on something three people liked.

Reading the result

A strong result usually looks like this: people describe the problem before you do, they have paid for or built workarounds, and some of them ask when they can have it.

A weak result looks like this: lots of encouragement, no commitments, and a problem that happens once a year or costs very little.

A mixed result is common. Maybe agencies cared and solo freelancers did not. That is not a failure; it tells you who the product is for. Run the same test again with the narrower group.

The short version

  • Write the idea as a problem with a cost, not as a feature.
  • Find people describing the problem, not people with a matching job title.
  • Ask about what they did, not what they would do.
  • Ask for money, a deposit or a signup after they see the price.
  • Decide with a number you wrote down first.

When the test passes, give yourself a date to ship toward. A launch week on ShipDesk is a free one.

pick a launch week on ShipDesk