Here’s something most founders don’t want to hear about validating product ideas.
Running ads to test your idea doesn’t actually validate your idea.
I know. It’s the playbook everyone recommends. “Build a landing page, run some Facebook ads, see if people sign up.” Sounds logical. Spend a little money upfront to test demand before building the whole product.
But here’s what actually happens. You get traffic to the landing page (because that part is easy), but the conversion rates are incredibly low. You end up with a list of random people who might be vaguely interested, but you still don’t know if you have a good idea. All you’ve proven is that you can execute that channel reasonably well.
And look, if you have the budget, resources, skills, and team to execute a paid channel flawlessly, this could work for you. But most founders don’t. Most founders are working with limited runway and haven’t built expertise in paid acquisition yet. So they waste money on ads that don’t actually tell them what they need to know.
There are too many variables in the “run ads to test your idea” approach. You need the right copy, right creative, right targeting, right funnel, right messaging. Even if you get clicks, you’re still not sure if the idea is good or if you just need to optimize the funnel. Google and Meta can’t optimize your campaigns until you have volume and paying customers anyway.
This playbook used to work. But now it’s too expensive and too competitive. The bar is much higher.
So what actually works for validating product ideas? Let me walk through the approaches I’ve seen work again and again.
The 60-70% Rule When Building a Product
Before we get into specific tactics, you need to understand this foundational principle.
It’s better to be 60-70% confident based on real data than 100% confident based on no data.
Technical founders especially struggle with this. They want 90-100% certainty before building anything. But great CEOs operate with 70% information and make decisions anyway. Spending too much time fact-finding doesn’t pay off. You need to execute with good-enough data and learn as you go.
The validation approaches below will get you to that 60-70% confidence threshold. That’s enough to start building. That’s enough to commit.
Talk to Real Humans to Validate your Ideas
This is the foundation. You have to talk to people.
But not just any conversations. Most founders ask terrible questions that don’t actually validate anything.
Don’t ask leading questions like “I’m building this, do you want it?” or “Would you pay for this?” People will tell you yes because they’re polite or they like you or they think it sounds interesting. That’s not validation.
Instead, ask about their current process. Ask about their challenges. Observe the problems they’re dealing with right now.
I recommend reading The Mom Test by Rob Fitzpatrick. It’s a one-hour read (hot pink book, you can’t miss it), and it completely changes how you approach customer conversations. The core idea is you should be able to ask your mom about your business idea and get useful information, because you’re not asking about your idea at all. You’re asking about her life, her problems, her current solutions.
You’re learning to listen for real opportunity versus waste of time.
And if you can’t reach your exact target buyer yet, talk to proxies. Talk to an expert in the industry. Talk to someone in an adjacent role. You can still learn a ton about the problem space and whether there’s real opportunity there.
Build in Public
This one is slower than ads, but it’s more sustainable and attracts higher quality interest.
Start talking publicly about the problem you want to solve before you build anything. Write about it. Share your thinking. Explain why you think this problem matters.
Building in public attracts three groups: buyers (people who have this problem), connectors (people who know others with this problem), and feedback (people who can help you refine your thinking).
The pace is different from paid ads. You’re not going to see a spike of traffic overnight. But the people who show up are genuinely interested, and they’ll stick around. They’ll tell you if you’re on the right track or if you’re missing something critical.
Deliver the Service Manually First
This is one of my favorite validation approaches because it proves demand without requiring you to build software.
I worked with a client named Rachel who was building Motivo, a marketplace connecting therapists with clinical supervisors. Before she built any platform, she manually matched people. She found therapists who needed supervision. She found supervisors who had capacity. She made the introductions and facilitated the relationships.
She had hundreds of buyers before she ever got funding or built the platform.
This approach validates that people actually have the problem, that they’re willing to pay to solve it, and that your proposed solution makes sense. Once you have proof of demand, you can raise money or commit to building the software. But you’re not betting your runway on an untested idea.
Quiet Launch
Not every founder wants to build in public or run a big flashy launch. Some people prefer a calm, measured approach.
Rand Fishkin’s Alert Mouse is a great example of this. He built it for himself first (low risk, because he needed it anyway). Then he soft launched it to his network and email list. He started with beta pricing, figuring out monetization later.
This isn’t for everyone. But if you’re the kind of person who wants to move deliberately and build something sustainable without the pressure of a big launch, this can work.
The key is you’re still validating. You’re watching how people use it. You’re learning what resonates. You’re gathering feedback from real users in low-pressure environments.
Prototyping and Design Sprints
This approach comes from Shape Up by Basecamp, and it’s incredibly useful when you’re trying to decide how to build something.
Create 2-3 different versions or approaches as prototypes. These don’t have to be functional. You can do this in Figma. You’re just creating enough fidelity to put in front of prospects and get live feedback.
The questions you’re asking are not “Do you think this is valuable?” You’re asking “What do you think this does?” You’re watching where they get confused. You’re seeing which version resonates most.
This validates your direction before you commit to a full build. You’re learning which approach makes the most sense to your target users.
Why the “Run Ads to Test Your Idea” Myth Persists
I want to come back to this because I know some of you are still thinking, “But I’ve heard so many people say to run ads.”
The reason this myth persists is it sounds logical. Test demand with a small budget before building. Seems smart.
But in practice, there are too many variables. Did the campaign fail because the idea is bad? Or because your targeting was off? Or because your landing page copy didn’t connect? Or because your creative wasn’t compelling enough? Or because the funnel was wrong?
You can’t isolate whether the idea is good when there are 47 other execution variables in the mix.
The approaches above let you isolate the idea itself. Talk to people, and you learn if they have the problem. Deliver the service manually, and you learn if they’ll pay to solve it. Build prototypes, and you learn if your proposed solution makes sense to them.
These methods get you to 60-70% confidence based on real data. That’s what you need to move forward.
What This Looks Like in Practice
Let’s say you’re thinking about building a tool for SaaS founders to track their customer research insights.
The “run ads” approach: Build a landing page, run some Facebook ads to SaaS founders, see how many people sign up for early access. Spend $2,000-$5,000, get a list of emails, still not sure if the idea is good.
The validation approach: Start talking to SaaS founders about how they currently track customer research. Ask what tools they use. Ask where they get stuck. Ask what’s frustrating. Do 6-12 of these conversations (The Mom Test style). Maybe tweet about the problem and see who responds. Maybe deliver a manual version of the service (help a few founders organize their research notes in a spreadsheet and charge them for it).
After 6-12 conversations and maybe 2-3 manual clients, you’ll have real data. You’ll know if this problem is painful enough that people will pay to solve it. You’ll know what features matter most. You’ll know how they think about the problem and what language resonates.
That’s 60-70% confidence. That’s enough to start building.
Operating with Imperfect Information
The hardest part of this for technical founders is accepting that you’ll never have perfect information.
You’re going to make bets with incomplete data. That’s part of being a founder. You can’t fact-find your way to certainty.
But you can gather real data from real conversations with real prospects. You can test demand by delivering the service manually. You can validate direction by putting prototypes in front of people.
Those validation methods get you to 60-70% confidence, which is enough. Then you execute, you learn as you go, and you adjust based on what you discover.
The founders who spend months trying to get to 100% confidence before building anything are usually the ones who never ship. They’re stuck in analysis mode.
The founders who operate at 60-70% confidence and make decisions anyway? They ship. They learn. They iterate. And they’re the ones who build sustainable businesses.