How to Validate a SaaS Idea Before You Build It
Validation is evidence someone will pay, not evidence someone is polite. Four tests, cheapest first, and what each one actually proves.
Validation means evidence that someone will pay, not evidence that someone said your idea was good. Run the cheapest tests first: problem interviews, then a landing page with a real signup, then a manual version delivered by hand, then a paid pilot. Build the software last, if at all.
The expensive mistake isn’t building the wrong thing. It’s building the right thing for people who were being nice to you.
What counts as validation?
A useful test: would this evidence convince someone who wanted you to be wrong?
| Signal | What it proves | Strength |
|---|---|---|
| “That’s a great idea” | They are polite | None |
| “We have this problem” | Problem exists | Weak |
| Describes a current workaround | Problem is felt, not theoretical | Moderate |
| Already paying for something | Budget exists | Strong |
| Pre-pays or signs a pilot | They will pay you | Decisive |
The jump from moderate to strong is the one that matters. Someone with a spreadsheet workaround has a problem. Someone paying for a bad tool has a budget line, which is a different and much better thing.
The four tests, cheapest first
1. Problem interviews
Ask about the last time the problem happened, what they did, and what it cost. Past behaviour only. The moment you describe your solution, the conversation becomes about being encouraging and the data is gone.
2. Landing page with a real action
Describe the outcome, not the features, and ask for something that costs the visitor a little friction — an email, a short form. Traffic without a signup means the pitch or the audience is wrong, and that’s useful before you write code.
3. Do it manually
Deliver the outcome by hand for a handful of people. Spreadsheets, email, your own labour. This is unglamorous and it’s the best test on the list: it tells you whether the outcome is wanted independently of whether your software is any good.
4. Paid pilot
Charge, even a small amount. Payment converts opinion into evidence and changes the feedback you get — people who have paid tell you what’s wrong.
This ordering assumes you’re building for a market you can reach. If your buyers are inside large organisations with procurement cycles, a pilot may take months regardless of interest, and the manual test becomes far more important than the landing page.
How long should validation take?
Weeks, not months. The failure mode is validating forever because it’s safer than launching. If three tests come back strong, that’s your answer — further interviews are procrastination wearing a lab coat.
- Evidence of payment beats evidence of enthusiasm.
- Ask about the last time it happened, never about your idea.
- An existing paid workaround is the strongest pre-launch signal.
- Deliver it manually before you build it.
- Charging early changes the quality of feedback.
What to do this week
Book three conversations with people who have the problem. Ask only about the last occurrence and what they did. If two of the three describe a workaround they built themselves, you have something worth the next test.