I built a startup for eight months before I found out nobody wanted it. Menu redesign platform connecting freelance designers with small restaurants. Wrote the code myself, hired a contractor for mobile, even trademarked the name. Launch day came. Eleven signups in the first month. Seven were friends and family. The other four bounced and never came back. That ,000 lesson taught me something I should’ve learned in week one: validation isn’t optional, and asking your friends if your idea is good doesn’t count as validation.
I’m one of those founders, by the way. Or I was. Back in 2017, I poured about eight months into a platform connecting freelance designers with small restaurants for menu redesigns. Wrote code late at night, hired a contractor for the mobile app, even trademarked the name. Launch day came. Eleven signups in the first month. Seven were friends and family. The other four bounced and never returned. That little adventure cost me $40,000 and a year I won’t get back. So yeah, I’ve got some opinions about validation.
But here’s what bugs me. Everybody nods along when you say “validate your idea first.” It’s become startup gospel at this point. And still, most people don’t do it. Why? I think it’s because building feels like progress. Talking to strangers about their problems? That feels awkward and slow. Writing code gives you something to show. A conversation gives you… what, notes? It doesn’t feel productive. But furnishing a house before checking the foundation is a special kind of foolish, and that’s pretty much what coding before validation amounts to.
Do Founders Even Know What Validation Means?
I’ve probably talked to a hundred early-stage founders over the past few years. When I ask “did you validate this idea?” the answer is almost always yes. Then I ask how. And that’s where things fall apart.
“I asked my friends.” Cool, your friends lied to you. They’re being nice. That’s what friends do.
“I posted in a Slack group and people seemed interested.” Seemed interested. Not paid money. Not signed up. Seemed.
“My co-founder and I both have the problem.” Great, so that’s two customers. Can you build a business on two customers?
Polite nods from people who care about your feelings aren’t validation. Your mom saying “that’s wonderful, honey” isn’t validation. Your co-founder agreeing with you is especially not validation — they’re already committed. Real validation means getting uncomfortable. It means hearing “no” from strangers and being willing to scrap your idea if the evidence points that way. Most people would rather build in a bubble than face that possibility. Which, I think, is exactly why that 73% number stays so stubbornly high.
So what does actual validation look like? I’m going to walk through the process I use now — the one that would’ve saved me $40K on the restaurant menu disaster. I’ve run three ideas through it since that failure. Killed two early, which probably saved me $80K or more. The third became a company that got acquired in 2023. Not a bad track record. But the process only works if you’re ruthlessly honest with yourself about what you find. And honestly? That might be the hardest part of all.
Start With the Problem, Not Your Brilliant Solution
Write down the problem you’re solving. One sentence. Not the product, not the features, not the tech stack — just the problem. If you can’t get it into a single clear sentence, you probably don’t understand it well enough yet.
“People need a better way to manage their receipts.” That’s a problem statement. “An AI-powered receipt scanning app with OCR and automatic categorization” — that’s a solution statement. You want the first one. The solution comes way later. And I know this feels backwards because the solution is the exciting part. Nobody daydreams about problem statements. But rushing to solutions is exactly how you end up building something clever that nobody needs.
Now, who actually has this problem? And please don’t say “everyone.” That’s not an answer. “Small business owners spending more than three hours a week on expense tracking” — that’s an answer. Specificity makes everything downstream easier. You’ll know who to interview, where to find them, what questions to ask. Vague problem definitions lead to vague research, which leads to vague conclusions, which leads to a vague product that nobody specifically wants. See the pattern?
One more thing before you move on. Write down your assumptions. All of them. Every idea sits on a stack of assumptions that most founders never bother to spell out. You’re assuming the problem exists at scale. You’re assuming people care enough to pay for a fix. You’re assuming they aren’t already solving it some other way. You’re assuming you can actually reach these people. You’re assuming the market is big enough to matter. List every single one, because the whole point of what comes next is testing these assumptions before you bet your savings on them.
Customer Interviews: The Part Everyone Skips
Of all the validation work you could do, talking to potential customers matters most. It’s also the step founders resist hardest. I get it — I resisted it too. Reaching out to strangers, scheduling calls, asking about their problems. It feels clunky compared to coding. But there’s no substitute for a real conversation where you can dig into how someone actually experiences a problem in their day-to-day life. Surveys give you numbers. Interviews give you understanding. Those are different things.
You need at least twenty conversations. Not five, not ten. Twenty. Why so many? Because the first handful will be noisy. People have different contexts, different vocabularies, different priorities. You’ll hear contradictory stuff. That’s normal. Somewhere around conversation fifteen, though, patterns start showing up. You’ll notice the same frustrations described in similar language, the same janky workarounds, the same moments of “ugh, I hate when that happens.” That’s signal. If you hit twenty conversations and there’s still no pattern? That probably tells you something important about how widespread this problem really is.
And here’s the part that trips up almost every first-timer: don’t pitch during the interview. I know. You’re excited. You want to describe your vision and watch their eyes light up. Resist. Hard. The moment you start talking about your solution, the conversation stops being about their reality and becomes about your fantasy. Instead, ask about behavior. “Tell me about the last time you dealt with [problem].” “Walk me through what you did.” “How much time did that eat up?” “What was the worst part?” “Have you tried fixing it some other way?” “What happened with that?” These questions give you ground truth. Your pitch gives you nothing but polite bias.
Rob Fitzpatrick wrote a whole book about this called The Mom Test, and if you haven’t read it, maybe stop reading this and go grab a copy. Seriously. His core idea is dead simple: ask about their life, not your idea. Bad question: “Would you use an app that does X?” Good question: “How do you currently handle X?” People will absolutely fib to be nice when you ask about your concept. They can’t fib about stuff they’ve already done. Past behavior beats stated intentions every single time, and it’s the only data from an interview you should actually trust.
Where do you find twenty people to interview? Depends on your market. LinkedIn works well for B2B. Reddit, niche forums, and Facebook groups work for consumer stuff. Industry meetups. Twitter. Wherever your target customer already hangs out. Don’t be spammy about it. A simple message works: “I’m researching how [type of person] deals with [problem], and I’d love to hear about your experience. Would you have 20 minutes for a quick call?” You’d be surprised how many people say yes when you approach with genuine curiosity instead of a sales pitch.
Landing Page Tests: Where Talk Meets Action
Interviews confirmed a real problem exists? Good. Now test whether people would pay for your specific solution. This is where the landing page test comes in. And I should be clear — this isn’t “put up a website and see what happens.” There’s a method, and skipping it turns the whole exercise into a vanity metric machine.
Build a simple page describing your product as though it already exists. Clear headline stating the benefit. A few points explaining what it does. Maybe a mockup or diagram. And then a call to action — this is the part that actually matters. Don’t use “join our newsletter” or “sign up for the waitlist.” Those are low-friction actions that tell you almost nothing. Someone giving you their email is not the same as someone reaching for their wallet. Use “Buy now,” “Start your free trial,” or “Pre-order for $X.” You want to see if people will take an action that costs them something real, even if it’s just entering credit card details.
Buffer did this years ago and it’s still one of my favorite examples. Joel Gascoigne threw up a landing page describing what Buffer would do, complete with pricing plans. When someone clicked a plan, they saw a message saying the product wasn’t ready yet, and could leave their email. The conversion rate — people who actually clicked a pricing plan — gave him real demand signal before he’d written any code. That’s the kind of data you’re after.
Drive traffic with small paid ad campaigns. Google Ads, Facebook Ads, even Reddit Ads these days — you can start at $10-20 per day. Run it for a week or two. You aren’t building a business here; you’re getting enough visitors to see a pattern. If 500 people land on your page and zero click the buy button, that’s a strong signal. If 5% click “Buy now,” you might be onto something. Exact thresholds vary by industry and price point, but anything above 2-3% conversion on cold traffic is probably encouraging.
Don’t overthink the design. Use Carrd, Unbounce, or a basic WordPress page. I’ve seen founders spend three weeks perfecting their landing page layout and then run the actual test for two days. That’s completely backwards. Spend two days on the page. Spend three weeks collecting data. The data is what you’re here for, not a pretty gradient.
Smoke Tests, Fake Doors, and Other Creative Cheats
Smoke tests push the landing page idea further. You simulate the product’s value without building the full thing. This is where creativity matters more than engineering skill. Zappos started by photographing shoes at local stores and posting them online. When someone ordered a pair, the founder went to the store, bought them at retail, and shipped them out. No inventory system. No warehouse. Just a scrappy test of whether people would buy shoes on the internet. That test validated what became a billion-dollar company.
A “fake door” test is a specific flavor of smoke test. You put a button or link for something that doesn’t exist yet, track how many people click it, and show them a “coming soon — leave your email” message. Works great if you’ve already got a product and want to gauge demand for a new feature. But it works for brand new products too. Put a “Download the App” button on your landing page. When someone taps it, log that click and show a launch date. The click-through rate tells you whether the promise was compelling enough to act on.
Then there’s the concierge MVP, which I think might be the most underrated validation tactic out there. Instead of building software, you do the job manually for your first customers. Building a meal planning app? Manually create meal plans for ten people based on their preferences. Email the plans to them. Charge money for it. This does two things at once: it proves people will pay for the outcome, and it gives you a ridiculously deep understanding of what the software actually needs to do. You’ll learn more in a week of manual delivery than you’d discover in months of user-testing a built product. From what I’ve seen, founders who start with concierge delivery build way better v1 products.
The Wizard of Oz approach is similar but sneakier. Customers think they’re using a fully automated product, but behind the curtain, a human is doing the work. It’s more deceptive than the concierge method, so tread carefully. But it can be surprisingly effective for testing complex ideas. A company I advised in 2022 was building an AI contract review tool. Before they built the AI, they had a lawyer reviewing contracts manually and delivering results through the product interface. Customers got answers in 24 hours instead of 2 minutes, but the quality was high. They validated that businesses would pay $200/month for the service before spending anything on AI development. Smart move.
Pre-Sales: When Money Does the Talking
Nothing validates an idea like someone giving you actual money. Not “I’d totally pay for that.” Actual dollars moving from their account to yours. Right now. Today. If you can get ten people to pre-pay for something that doesn’t exist yet, you’ve got stronger proof than a thousand survey responses could ever give you. People say all sorts of things in interviews and surveys. Wallets don’t lie.
Kickstarter and Indiegogo turned pre-sales into a mainstream strategy, but you don’t need a crowdfunding platform. Send an email to the people you interviewed. Tell them you’re building the solution, describe what it’ll do, and offer a founding member discount — say, 50% off the first year — if they commit now. Stripe or Gumroad can handle the payment. If nobody bites even at a steep discount, that means something. If they do pay, you’ve got first customers and some capital to start with.
One thing I want to be upfront about: pre-sales come with an obligation. You actually have to deliver. Don’t take money and disappear. Set clear expectations on timelines, and be ready to refund anyone who changes their mind. But the act of asking for money — and the response you get — teaches you more in a week than months of desk research ever could. When I validated my second company, I emailed 50 potential customers with a pre-sale offer. Twelve bought. That gave me $3,600 in revenue and enough confidence to leave my consulting work and go full-time. Those twelve people also became my most engaged early users. Their feedback shaped the product for the next two years.
Does the Lean Startup Framework Still Hold Up?
I can’t write about idea validation without mentioning Eric Ries and the Lean Startup, because his framework popularized most of these concepts. And his core insight — build-measure-learn, minimum viable products, validated learning — is right. You should run experiments instead of making guesses. You should measure real behavior instead of collecting opinions. You should learn from data, not from your gut feeling.
But I’ve got some issues with how the lean startup movement played out in practice. Mostly around “MVP.” That term became a buzzword people use to justify shipping half-baked garbage. An MVP isn’t a broken, barely functional product. It’s the smallest thing that tests your riskiest assumption. Sometimes that’s a landing page. Sometimes it’s a spreadsheet. Sometimes it’s literally just a conversation. The “product” in MVP doesn’t have to be software at all — it has to be something you can learn from. Too many founders hear “MVP” and think “build version 0.1 and push it live,” which is often way more work than validation actually requires.
The other thing I think Ries underemphasized — and maybe I’m wrong about this — is how much work should happen before you build anything at all. His build-measure-learn cycle assumes you’re already in building mode. But the highest-value validation happens before a single line of code. Customer interviews, problem definition, assumption mapping. That’s the pre-work that determines whether your build-measure-learn loop is pointed in a useful direction or just spinning you in circles.
How Do You Actually Know If Your Idea Passed?
Everyone asks this, and there’s no clean formula. Validation isn’t a binary pass/fail — it’s more like a spectrum of confidence. But after running this process a few times, certain signals start to feel reliable. Here’s what I look for, roughly in order of how convincing they are:
- Interview consistency: At least 70% of your conversations point to the same core problem, described in similar terms, with similar levels of frustration. If fifteen out of twenty people are telling you basically the same story, that’s hard to ignore.
- Active spending already happening: Your target customers are paying for partial solutions right now — other products, consultants, manual workarounds that eat their time or money. If nobody’s spending a dime to fix this problem today, they probably won’t spend on your fix either.
- Landing page conversion above 3%: Cold traffic taking a high-commitment action like entering payment info, booking a demo, or pre-ordering. Anything north of 2-3% from strangers who’ve never heard of you? That’s encouraging.
- Pre-sale traction: At least 10% of the people you ask actually put money down before the product exists. Not “sounds interesting, keep me posted.” Money. On the table.
- Pull instead of push: People asking you when the product will be ready, rather than you chasing them down to explain why they should care. When potential customers start following up with you, something real is happening.
- Kill signals to watch for: Polite interest but zero action. People acknowledging the problem but ranking it low priority. Nobody currently spending on any solution. Every interview revealing a different problem. Struggling to even find twenty people who fit your target profile. Any of these should make you seriously question whether to keep going.
None of these metrics alone should make or break your decision. But stack a few positive signals together and you’ve got something. Stack a few negative ones and, well — better to know now than after you’ve spent six months building.
Validation Hurts. Do It Anyway.
Can I be honest for a second? This process is emotionally rough. You’re taking something you’re pumped about — something you’ve been daydreaming about, maybe for months — and shoving it under harsh light. And that light often says “nah.” When I killed my second idea (a subscription box for home office supplies) after three weeks of interviews, a landing page, and some ad spend, the data was painfully clear: not enough people cared. I moped about it for a while. It felt personal.
But here’s how I eventually reframed it. Killing a bad idea early is a win. Not a consolation prize. A genuine win. You didn’t waste half a year. You didn’t drain your savings. You didn’t hire people and then have to let them go. You spent maybe a few hundred dollars and a few weeks of evenings, and you learned that this particular path was a dead end. That’s not failure — that’s cheap tuition.
There’s also this sneaky tendency to cherry-pick your data. You’ll do twenty interviews, seventeen of which are lukewarm, and then you’ll fixate on the three who were excited. “But those three really loved it!” Sure. And 85% of your sample didn’t. Be honest about what the full picture shows. Confirmation bias will absolutely wreck your validation if you let it. Seek out the “no” as hard as you seek the “yes.” Maybe more so.
You Don’t Need Much Money For This
Some founders act like validation is this big expensive production. It’s not. Recently I helped someone run through this entire process for under $500. Customer interviews cost nothing — Zoom and Google Meet are free. A landing page on Carrd runs about $19 per year, or you can use Unbounce’s free trial. Ad spend to drive traffic: maybe $200-500 total at $15-25 a day for a couple weeks. Mockups in Figma’s free tier or Canva. Pre-sale collection through Stripe at 2.9% plus thirty cents per transaction. No upfront cost there.
All in, you’re looking at under $600 and three to four weeks of part-time work. Compare that to the $50,000-100,000 and six to twelve months a typical MVP build costs. Validation might be the highest-ROI activity in the entire startup process. It’s kind of wild to me that people skip it because they’re “in a hurry” to start building. In a hurry to do what, exactly? Build something nobody wants? The market isn’t going anywhere. Take the few weeks. Do the work.
B2B vs. B2C: Same Idea, Different Tactics
Quick note here, because the process shifts depending on who you’re selling to. B2B validation is, in some ways, easier. Business buyers tend to be more rational about their problems. They’re spending company money to solve company problems, and they can usually describe their pain points clearly. You can find them on LinkedIn, at trade events, through warm introductions. And asking for money feels more natural. “We’re building a tool that does X. Would your team pay $500/month for it?” — that’s a perfectly normal question to ask a VP of Operations.
B2C is trickier. Consumer behavior is less predictable. People say they want things and then don’t buy them. Emotional stuff, brand perception, social pressure — it all factors in. Landing page tests and smoke tests become even more important for consumer products because you need to observe what people actually do, not what they say they’d do. B2C also tends to need bigger sample sizes. If you’re building a consumer app, you might want 50-plus interviews and thousands of landing page visitors before you feel confident in your read of the market.
After Validation: What Changes
Say your idea passes. People have a real problem, they’re willing to pay, and your specific angle resonates. What now? You build. But you build differently than you would’ve without validation. You’ve got real customer language to use in marketing — their words, not yours. You’ve got a prioritized feature list based on what people actually asked for, not what you imagined they’d want. You’ve got potential beta users who already raised their hands. And you’ve got something that’s hard to put a price on: the confidence that comes from knowing, not just hoping, that there’s demand on the other side.
Start with the smallest version that delivers the core value you validated. Not every feature someone mentioned. Not the nice-to-haves. The one thing that, if it works, makes people go “yes, this fixes my problem.” Ship that to your pre-sale customers or your beta list. Collect feedback. Iterate. You’re still learning — that never stops — but now you’re learning about execution rather than whether the market exists. Much better place to be.
The best time to validate was before you started building. The second best time is right now. Stop coding for a week and go talk to twenty potential customers. Whatever you learn will be worth more than whatever you would have built.
I’ve watched dozens of founders go through this. Some do the work, face the data honestly, and build companies on a foundation of evidence instead of hope. Others skip it, build for a year, launch to silence, and circle back wondering what happened. Hope is great fuel for motivation. It’s a terrible basis for strategy. Will more founders start taking validation seriously? Will that 73% number ever drop? Hard to say. The tools keep getting cheaper, the playbook keeps getting clearer, and yet people keep falling in love with their ideas before checking if anyone else will. Maybe the next generation of founders will be different. We’ll see.
(0) Comments