How to Validate Product Ideas Before You Waste 6 Months Building the Wrong Thing
The fastest way to validate product ideas is to find 10-15 people in your target audience, talk to them about the problem (not your solution), and see if any...
I've watched too many founders spend a year building something nobody wanted. The pattern is always the same: they fell in love with the idea and skipped the boring part where you find out if anyone cares.
What is validating product ideas?
Validating a product idea means gathering evidence that a specific group of people has a problem painful enough that they'll pay to solve it, before you commit serious money or months of your life to building.
The keyword there is evidence. Not opinions from your friends. Not upvotes on Reddit. Not your cousin saying "I'd totally use that." Evidence looks like: signed LOIs, pre-orders, waitlists with real emails, people paying you to solve the problem manually before the product exists.
Why it matters: CB Insights analyzed 483 startup post-mortems and found "no market need" was the #1 reason startups fail, cited in 35% of cases. Not bad tech. Not bad marketing. Just building something nobody wanted.
How to validate product ideas in 7 steps
- Write down the problem in one sentence. Not the solution. The problem. If you can't do this without mentioning your product, you don't understand the problem yet.
- Identify 15-20 people who have this problem right now. Real names. LinkedIn profiles. Not "small business owners" as a category.
- Interview at least 10 of them. Ask about the last time they hit this problem. What did they do? What did they try? How much did it cost them (time or money)?
- Look for the pain signal. Are people already spending money on bad workarounds? Duct-taping spreadsheets together? That's gold. If nobody's doing anything about the problem, the problem isn't real enough.
- Make a fake door test. Landing page, one line describing the product, an email signup or a "Buy for $X" button that leads to a "coming soon" message. See who clicks.
- Ask for money before you build. Pre-orders, deposits, annual subscriptions at a discount. Verbal "I'd buy this" is worthless. Credit cards on file are not.
- Do it manually first. If you're building a service, deliver it by hand to 3-5 customers before automating anything. This is what Paul Graham calls "doing things that don't scale" and it's the single most useful piece of startup advice I've read.
Skip any of these steps and you're gambling.
The stats that should scare you
- 35% of failed startups blamed "no market need" according to CB Insights' analysis of 483 post-mortems (updated 2021).
- 9 out of 10 new products fail in some form, based on decades of consumer packaged goods research summarized by Harvard Business Review (2011).
- 42% of surveyed founders told CB Insights their startup died because there was "no market need" for the product, tying with running out of cash as a top cause.
The takeaway isn't that startups are hard. Everyone knows that. The takeaway is that most of these failures were preventable with two weeks of customer conversations.
Validation methods compared
Not all validation methods are equal. Some feel productive but tell you almost nothing.
| Method | Signal strength | Time to run | Common mistake |
|---|---|---|---|
| Friends/family feedback | Very weak | 1 day | Believing them |
| Surveys | Weak | 1-2 weeks | Asking "would you buy this?" (answer is always yes) |
| Customer interviews | Medium-strong | 2-3 weeks | Pitching instead of listening |
| Landing page + ads | Strong | 1-2 weeks | Not driving real traffic |
| Pre-sales / LOIs | Very strong | 2-4 weeks | Skipping because it feels scary |
| Concierge MVP | Very strong | 2-6 weeks | Automating too early |
If you're going to pick one, pick pre-sales. Nothing else comes close. Money moving into your account is the only signal that isn't polluted by politeness.
What most people get wrong
They ask leading questions. "Would you use a tool that does X?" Everyone says yes because it's free and easy to be nice.
Instead, ask about the past. Rob Fitzpatrick's The Mom Test is the whole playbook here and I recommend it to anyone who'll listen. Ask what people did last Tuesday, not what they'd hypothetically do next month.
The other thing people get wrong: they validate the wrong thing. You don't need to validate that the problem exists. You already know it does, that's why you thought of the idea. You need to validate that (a) enough people have it, (b) they'll pay real money to solve it, and (c) they'll pay you to solve it, not the 8 competitors already in the space.
One more thing. If your validation plan doesn't include the possibility of finding out you're wrong, it's not validation. It's confirmation bias with extra steps.
Where do good ideas to validate come from?
Real complaints from real people. That's it.
Review sites, support forums, Reddit threads, app store reviews. People spelling out what's broken in their current tools. You can spend a weekend reading 200 reviews of a competitor and come out with a better product roadmap than most companies have.
Key Takeaways
- 35% of startups fail from "no market need." Most of this is preventable.
- Validation means evidence, not opinions. Pre-orders beat surveys every time.
- Interview at least 10 real people who have the problem right now.
- Ask about past behavior, not hypothetical future behavior.
- If you're not willing to be proven wrong, you're not validating anything.
Pick one idea you've been sitting on. Book 5 customer calls this week. If you want a shortcut to finding what people are already complaining about, dig through real product reviews and let the pain points surface themselves. Try the Review2Idea free analysis tool to pull opportunities straight from app reviews people have already written.
Frequently Asked Questions
Q: How long should it take to validate a product idea?
A: Two to six weeks for most software ideas. If you're spending 3+ months validating, you're either overthinking or avoiding the scary part (asking for money). Give yourself a deadline.
Q: How many customer interviews do I need?
A: At least 10, ideally 15-20. You'll start hearing the same problems repeated around interview 8, which is when patterns become useful. Fewer than 10 and you're generalizing from noise.
Q: Can I validate an idea without talking to customers?
A: Sort of. Landing page tests and ad campaigns can measure interest at scale. But they tell you what people click, not why. You'll need interviews to understand the why, which is what shapes the actual product.
Q: What if my idea is so new that nobody knows they need it?
A: This is what founders tell themselves right before wasting 18 months. Even truly new products solve an existing problem. Uber solved "cabs never come." Figure out the underlying pain that already exists today.
Q: Should I build an MVP to validate?
A: Usually no, not first. A concierge version (you doing the work manually for 3-5 customers) or a pre-sale beats an MVP for early validation because it's faster and forces real commitment. Build code only after people have paid you to.
Find this kind of gap in your own category
Run one free analysis to see the complaint patterns in a competitor, then upgrade to Pro when you need a full research sprint.