This is an article I have been sitting with for years. Incessantly pointing out to founders that they’re getting the MVP wrong or that Lean Startup misled them, I want to throw out provocative notion to challenge us all to ask why, with the clarity of what to do in Lean Startup, and so many practitioners and adherents, do overwhelmingly most startups still fail while everyone says “MVP”
The conventional wisdom we all preach is that 90% of startups fail. Now, my first startup was in 2002, when it was also said that 90% of startups fail. One of two things must be true: either Lean Startup is wrong, or founders are doing it wrong.
Eric Ries gave the world a gospel: build a minimum viable product (MVP), validate it with customers, and iterate toward product-market fit. The idea promised to democratize innovation and de-risk entrepreneurship. Yet nearly two and a half decades later, the success rates haven’t changed (presumably), and venture outcomes haven’t improved. The startup graveyard still fills faster than seats in the accelerator downtown. So, is the method broken? Or are the practitioners?
Lean Startup isn’t the problem
Lean has become dogma. Founders are quoting scripture without understanding the language. Most programs are preaching a process, with Lean Startup the Bible, when the only that works in entrepreneurship is systems.
Article Highlights
The Myth of Product-Market Fit
Let’s start with the sacred phrase: Product-Market Fit.
It sounds so elegant that we wouldn’t even question the order of the words and yet the order of the words, alone, has completely upended entrepreneurship: Product first, market second. And that’s the trap. The language itself is misguiding entrepreneurs. When you put “product” before “market,” you condition founders to build before they understand who or what they’re building for.
The truth is, it’s Market-Product Fit. (at least, philosophically)
The market comes first: its behavior, its unmet needs, its price sensitivity, its patterns of adoption. You build only after the market reveals what it will pay for. And yes, I did just say, in flowery language, “what the market will pay for” (not customers) – we’ll get to that.
Steve Blank, the intellectual godfather of Lean Startup, never said “build a product, then validate it.” He said, “get out of the building.” In The Four Steps to the Epiphany (2005), he warned against starting with a prototype. Instead, he taught that founders should start with customer discovery; not interviews, but immersion. Ries codified that work later, but by the time Lean Startup hit the MBA circuit, it had been neutered into process, stripped of philosophy. In nearly every cohort I experience in Startup Development Organizations, founders still are working on their MVP (their solution) having only spoken to, maybe, 10 potential customers. You aren’t listening to us! You didn’t actually read the book. Or worse, you’re surrounded by awful mentors and investors giving harmful advice whom you need to help us remove from the ecosystem (honestly, this wouldn’t surprise me)
The problem is that most founders never truly “get out of the building.” They Google some stats, talk to five friends, and call it customer discovery. Then they burn three months building something no one wanted.
MVP: The Most Misunderstood Acronym in Entrepreneurship
The next misstep is the MVP.
Lean Startup defines the minimum viable product as the smallest thing you can build to start the learning process. That definition was meant to protect founders from over-building, not to give them permission to build badly.
In practice, “MVP” has become an excuse. Founders treat it like a hall pass to push out half-finished apps and call it learning.
If you’re building an MVP to “see if people will use it,” you’ve already missed the point. Your MVP isn’t supposed to prove functionality. It’s supposed to prove demand.
Dropbox didn’t build before testing. Drew Houston made a simple explainer video demonstrating what Dropbox would do, and the waiting list jumped from 5,000 to 75,000. That’s an MVP. It proved people wanted it before they could have it.
Contrast that with the graveyard of “MVPs” that founders build in isolation; small versions of their big dreams, validated by no one, tested with no market.
Academic work has backed this up. A 2018 study concluded that startups using MVPs incorrectly focus on the product rather than the business model, which limits learning and prolongs time-to-market. Translation: you’re building too much, too soon.
Your MVP should be the minimum viable proof not the minimum viable product. Proof of what? That you can acquire paying customers. Certainly not that you can build something (especially these days, a chimpanzee with ChatGPT can build MVPs). Prove you can create demand and a competitive advantage, not deliver a solution.
Customer Validation: The Cult of Conversation
Here’s where Lean Startup really got hijacked: Customer Validation.
Blank’s original meaning was that founders should validate hypotheses through experiments. But MBA programs and startup bootcamps watered that down to “talk to customers.”
So now founders spend weeks collecting polite praise from potential users. “That’s a great idea,” people say. “I’d totally use that.”
You know what that is? Noise.
That isn’t validation, it’s social politeness. Humans hate confrontation, so they rarely tell you that your idea is dumb. The result is a false sense of traction.
True validation only happens through transactions.
When someone gives you an email, that’s interest. When they click “buy,” that’s validation. When they pay and return, that’s product-market fit (er, market-product fit).
This isn’t just opinion. A 2020 analysis by CB Insights on startup post-mortems found that 42% of startups fail because there’s “no market need” – not because the product didn’t work. In other words, founders keep building before marketing.
If you think validation means conversation, you’ve confused empathy with evidence.
Lean Startup Didn’t Reduce Failure Because It Was Never Used Properly
Let’s put data to the claim. The global startup failure rate remains roughly nine out of ten within five years, nearly identical to pre-Lean Startup eras. So let me be clear in establishing that while I’m defending Lean Startup given my point of view, the fundamental fact must be that either it’s wrong, or you are.
I don’t think it’s that the framework doesn’t work; it’s that no one follows it correctly. Instead of “build–measure–learn,” founders are doing “build–launch–hope.”
You treat Lean like a checklist, not a scientific method.
A 2019 Harvard Business Review critique, What the Lean Startup Method Gets Right and Wrong, observed that most teams “fail to apply the hypothesis-driven approach rigorously,” relying on shortcuts that invalidate the learning cycle. In plain English: the method doesn’t fail; the execution does.
The consequences of this misunderstanding go beyond startup failure. Economic development agencies and venture funds have built entire infrastructures around “Lean” principles, incubators, accelerators, government innovation hubs. Yet those systems also misapply the framework. They fund product-builders, not market-finders.
That’s why so many startup programs produce pitch decks instead of businesses. It’s all process, no market. It’s step-function, not system.
We’ve created a culture that celebrates the building of startups, not the selling of them.
If you want evidence, just walk through any coworking space. It’s filled with brilliant prototypes and empty wallets. Founders are still confusing invention with entrepreneurship.
Doing It Right: A Market-First Framework
If you want to do Lean right, you have to start before you build anything. The real sequence should look like this:
- Discover the market. Who has the problem? What have they tried? What do they already pay for?
- Prove demand. Build the smallest, cheapest test to show people will pay: a landing page, a pre-order, a prototype with a Stripe button.
- De-risk before you develop. Use those insights to remove your biggest assumptions.
- Then build the product that fits.
That’s the inversion most founders can’t make: marketing before development, demand before design.
Marc Andreessen put it, “Product-market fit means being in a good market with a product that can satisfy that market.” The order of those words was never meant to be sequential; it was descriptive. But if you start with the product, you’ll never get there.
So, Is Lean Startup Wrong?
Let’s return to the question. If Lean Startup was supposed to change everything, why hasn’t it?
Maybe because Lean was never supposed to be a “method,” it’s a mindset. It’s about humility before the market, not hustle before the investor.
But today’s founders aren’t humble. They’re pitching before they’re listening. They’re coding before they’re marketing. They’re mistaking validation for affirmation.
So no, Lean Startup isn’t wrong. It’s just being done wrong, and worse, preached wrong, by nearly everyone.
If Lean Startup really worked, we’d have more success stories to show for it. If it didn’t work, it wouldn’t still be the foundation of every accelerator and innovation program on Earth.
That paradox should make you uncomfortable.
Lean Startup is either wrong or you’re doing it wrong. But one of those must be true.
If you’re brave enough to assume it’s you, not the method, then flip it: start with the market, test demand, sell the story, and then build.
If you’re not (brave enough), if you’d rather keep building in a vacuum, waiting for your epiphany, the startup cemetery is still plenty big.
Yes!!!
There’s also a contrast in venture scale MVP vs normal business. To have something venture scale, you likely need something that’s 10X better in at least one of cost, performance/appeal, or distribution network.
Absolutely right Paul
Brilliant take!
Cheers Neil, hope you’re doing well my friend
Interesting perspective in the age of AI and vibe coding. I’d agree with this, especially as virtually anyone can put together what used to be called an MVP. Today, we have to be a bit more ruthless.
Good reminder too: no one has any excuse not to put together an experience to prove you can generate demand for what you’d like to offer.
Interesting points, Paul!
How I see it, the fact that failure rate hasn’t changed doesn’t mean lean startup is good for nothing. Point of lean startup is to make those failures quick and cheap, instead of planning for months and then meet a big fat nothing in the market.
While demand drives the success of many products undoubtedly and that’s always a good place to start with, I think your “demand first, product later” framework assumes you can separate them cleanly. You often can’t validate demand for something novel without building enough of it to make the value proposition tangible.
I do agree that founders can get stuck in endless loops of iteration and lose time in this framework.
What do you mean by MVP became a hall pass to build too much, too soon?
Consider your last question with your first point. IF Lean Startup works (or is applied properly) then we can’t have as many failures. Just speaking from experience (running incubators for decades, I’ve seen more pitches than a VC), there are both very dumb ideas and exceptional startups. Lean should eliminate the former and ensure greater likelihood of the later. We’d see as many bad ideas, and we certainly wouldn’t have as high a rate of good ideas failing.
Make sure you note the article, I actually didn’t say it’s good for nothing; I’m actually very supportive of it – I’m pointing out that founders don’t practice it despite saying so, that most mentors are awful, and most investors are worse.
And that is the answer to your question (or, my point). Founders use “lean startup” as permission or forgiveness to just build an MVP as though it’s telling us that you build an MVP to validate customers and find product-market fit. That’s not only wrong, it’s ignorant and, frankly, just idiotic.
That, we can validate and idea, find and convert customers, and even prove the ability to grow, without the thing. The point of marketing (and talking to customers) is to only build WHEN you have determined that it’s worthwhile… an MVP exists to further prove that you scale and compete. Lean isn’t permission to build an MVP because you’re ignorant and think you’ve been told to do it to prove you can (to investors), but that’s what founders do.
This is so true.
But somehow, “build something people want” is good advice in an accelerator. There’s a lot of blame to go around.
Paul O’Brien, doing well thank you. I’d like to give you an update sometime soon.
Most MVPs today look like group projects turned in at the last minute. No wonder the market hands out failing grades. Selling before building would save a lot of egos and a lot of cash.
Mark, what you’re pointing out is why I increasingly find it inexcusable if Startup Development Organizations or Mentors are disregarding this fact, or worse, failing to help a founder do it.
Ouch! This one landed. I realized I’d been building for an observed pain, founder cognitive overload, instead of an expressed one.
After our earlier exchange, I flipped the process: paused the “elegant solution” to test the problem people actually pay to solve, board-deck drag. The Board Deck Generator became my minimum viable proof: can it save a CFO at least one revision cycle?
Market-first thinking turns out to be liberating: less energy spent explaining, more energy finding proof. Thanks for being so vocal and opinionated! We need that clarity in the ecosystem.
“can it save a CFO at least one revision cycle?” this is smart Amanda Ross. Simple. If Yes = Money.
This is one of THE biggest mistakes I see Founders make consistently.
-> “customer discovery” = talking to five polite friends.
-> If you don’t know your customer,
-> you don’t know what pain point your solving for > you don’t know what to build to address the pain points
-> you don’t know how to communicate with your customer in order to gain adoption
If you don’t interview customers and create feedback loops you are building a business in a black hole.
re: “The truth is: it’s Market-Product Fit”
First, LOVE this ^^^. I want whatever it is you’re having for breakfast. Startup Wheaties??
Now that you mention it, and this could be to your point, one of the things that kinda bothers me about “product market fit” and “startups” is that – to me – it feels like it dumbs down ambition, in a way.
That is, if startups, by definition, are supposed to do the unthinkable, break the rules, etc. then – again, to me – the emphasis on P-M-F kinda implies, “Go big but not that big. Not too big.”
Fit, by definition, kinda implies a sense of normal, does it not?
Maybe I”m wrong?
Mark Simchock well, hate to pass on what might turn into jealousy to all my friends north in November, but it’s not what I’m eating rather, where ??
Love your point. It’s better put: Market Solution Innovation
Innovation is by definition, applied invention, meaning it works (or fits)
Part of the challenge is the name.of the learning loop. Build, measure, learn implies the first step is to build what you think is right. This creates a firm anchor point for you and your customers. An anchor point most founders latch on to for dear life. It is sacred. Customers that don’t like it don’t get it and are wrong.
The actual process starts with research and learning before building. Creating the first hypothesis. Which is a term I hate. It makes what you are going sound mature, established, with deep rigor. It is a guess and an opinion. Often a bad one based on limited data and your personal beliefs. Calling it a guess and an opinion makes it easier to evolve and for others to challenge it safely.
Matthew Burris I’ve explored and pushed hard hat even the “idea” is a horrifically misleading direction to advise.
Any half-experienced person can tell if an idea is a good one, so what the hell are we doing validating the idea. Besides, a startup isn’t an idea, it’s many ideas that together solve a problem in a way that creates value.
An idea (to your point) gets built so you can prove it.
Ideas need to be discarded quickly so as not to waste time and resources.
Hey Paul!
I propose that there is a two-headed monster that is driving this behavior: entrepreneur (and VC) impatience and the runway problem. Whether a startup is self-funded or VC-funded, there is huge pressure for “results”. In the back of everyone’s mind is the “fact” that there is no company w/o a product.
Unfortunately – as you correctly point out – “MVP” perpetuates this “fact”, focusing on the wrong results: a product, not transactions. (I had to remind my current team this just last week, for my new project).
Solutions?
–> We should be seeing more UX researchers in founding teams.
–> And more shared UX research resources in incubators.
–> And more emphasis on JTBD (streamlining the work of Ulwick and Christensen).
–> And more funding decisions based on “transaction metrics proof” (like Dropbox’s video, not prototypes.
Cheers,
John
John! How are you my friend?
To be fair, I’m actually a fairly strong opponent of the notion that it’s transactions. I think that too have gone of the rails, largely by investors who shouldn’t be in startups (who don’t know how to assess a startup and so they push a focus on customers).
Startups, in their distinction from new businesses, can not yet know the business model – the point is the invention of a new business model. And that being the case, customers & revenue are indeed an indicator, but are just as likely to be misleading, wrong, or informing and enabling competitors. Prove demand and the ability to compete and scale.
I’m confident that the answer is an actual marketer (and that’s part of the problem too: most founders, investors, and mentors, have no idea what marketing actually means)
Reality bites. There’s still too much ‘drive-by’ advice on social media platforms from so called pundits who’ve built well known communities and followers. Please keep it coming Paul. The industry needs more content like yours.
Reality bites. There’s still too much ‘drive-by’ advice on social media platforms from so called pundits who’ve built well known communities and followers. Please keep it coming Paul. The industry needs more content like yours.
This is excellent Paul. I’ll just say I wish I could have read this years ago.
Hm, I’m a bit confused by your stance here.
Lean startup has nothing to do with failure rates. It’s not an idea generation mechanism, but idea testing mechanism. Hence it has nothing to do with failure rates. It doesn’t make startups protected from failure nor does it give framework for better idea generation. But it does make failure less catastrophic. It allows founders to spend €10k before realizing it doesn’t make sense rather than €500k. So if anything, it should increase failure rates because failing is less costly and risky than ever.
I haven’t read the full article, so apologies for that. I’ll just argue the points made here.
I’m confused on this permission thing. Who do they need permission (and even less so forgiveness) from and why would building an MVP to test demand be idiotic?
I frankly disagree that you can remove uncertainty by “talking to customers”. This is something academia would try to teach you (from my own experience). People are 1) bad at interpreting other people’s ideas; 2) bad at giving advice with no skin in the game; 3) bad at predicting their own behaviour; 4) have bias to confirm beliefs.
The only real way to see what demand your product drives is to have people put their money where their mouth is, i.e pay for “something”. Yes, some products might be validated super quickly with a waitlist, but hardly anymore. You’re expected to ship in weeks nowadays. So I don’t agree with this. Obviously, you can’t just build blindly either.
Your claim that you can “validate and idea, find and convert customers, and prove ability to grow without the thing” might work for a small percentage of products. For most B2C and many B2B products, you need something concrete because people can’t visualize abstract solutions well.
I do agree that most founders should build MVPs that are a minimum artifact to test their riskiest assumption, not their full vision of the product.
Fair points. And while you might be confused, from what I read, we actually just disagree on a lot.
Paul I love that you’ve addressed how a truly helpful tool and principle has been twisted and misapplied by the community at large. It’s like the old game of telephone where the final message is nothing like the original. Thank you for calling this out ?
The dogma part hit home. Like Pilates, true sistems matter.
same. There is no right answer.
This is a great piece. The ideas behind Lean Startup are as valid as ever, but very few startups execute them properly.
Thank you for catching and sharing that, I’ve had a few detractors push that I’m wrong, that “Lean Startup is right” (they clearly missed my point)
Most startups (and even their incubators and investors) start out assuming that the idea is valid, and they only need a tiny bit of “validation” to move forward. They think “lean startup” is just a way to decide what features to build.