This article aims to answer the question: Is the lean startup method valid? Or is it a scam?

Does it work the way it should?

Or does it not – and implementing it in your business is going to sink it?

The question of the lean startup validation (or, sometimes, “empirical validation”) was posed a number of times by a number of people – most of them entrepreneurs. There’s no doubt that it’s important.

The answer is here.

The article is long – I agree. Believe me: if I could make it shorter, I would. Yet, the matter is complicated. I had to make it long like this.

What I promise you, however, is that after you’ve finished this read, all your doubts will be dispelled: you will know whether or not the lean startup is good for your startup. You will know what to expect.

“Why can’t I jump straight to conclusion?” – you’ll ask. Well: you can. I could give it to you here. If I didn’t – I didn’t for a good reason, believe me.

It’s not the fact that the lean startup is or isn’t valid that’s important here – it’s the reason behind it that is.

Hence, the length.

A Short Word Of Introduction

There is no need to present the lean startup.

Introduced by Eric Ries in his 2011 bestseller, The Lean Startup. How Today’s Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses, the method has soon become the next big thing (replacing Jim Collins’ Great By Choice and Steve Blank’s The Four Steps to Epiphany).

(For a detailed look at the method’s principles, look here.)

A number of startup entrepreneurs have since claimed that implementing the method in their businesses allowed them to succeed where the others would fail.

The others, however, have made an opposite case:

The lean startup method has not been properly validated, they would claim.

Before we implement the method in our businesses, we need to know if it works, if it really works. Otherwise, we may be actually wasting our time and resources. There may be a way to do things much easier.

Their reasoning was that as long as the method hasn’t been properly validated, we cannot be sure if the success of the companies that implemented it was due to it or (who knows?) despite it.

Thus, the validation is important. It makes a real difference.

Among the lean startup’s critics, Ben Horowitz (of Andreessen Horowitz), Peter Thiel (of Palantir, Founders Fund, Mithril Capital, and, formerly, PayPal, which he founded), and Sam Altman (of Y Combinator and OpenAI) stand out as these most famous. Elon Musk is not a fan, either. The names’ list – however – is longer. It includes: Rick Bohan, John Finneran, Mark Graban, Kenneth Grady, Trey Griffith, Daniel Markovitz, James C. Paterson, Debashis Sarkar, Robert J. Trent, Bill Waddell, and so on, and so on.

What connects them all is this basic – but essential – conviction:

If the management method we use is not valid, using it to launch and grow our business can be as dangerous as using no method at all. It might bring us heavy losses. It might sink us – definitely.

And there is a number of people who claim that the lean startup sank their businesses.

The method is a scam, they argue. It has never been properly validated. The evidence that it works is purely anecdotal and has nothing to do with real business.

Who is right?

The ones who implemented the method and succeeded – or the ones who implemented it and failed?

Can both these groups be right?

And what if neither of them is?

And how do we prove it?

The Importance Of Validation


I have spent the past several months looking for the answer.

I’ve poked through evidence. I’ve drawn business plans and logic models. I’ve delved deep into the nuances of the lean startup’s conceptual framework and the principles behind it. I have read case studies – a dozen upon dozen upon dozen of case studies. (I believe I’ve read them all.) I’ve certainly done a lot – a lot! – of research.

Still, the answer was nowhere to be seen. The problem seemed impossible to solve.

Is the lean startup method a scam? – I kept wondering. Is it a scam?

Because if it was, all these companies who were actually trying to implement it were in grave danger: in danger of failing and going bankrupt.

I knew what’s at stake – I knew it too well. Some time ago, in 2015, one of my friends has launched a startup. Michael – because that was his name – was really into the method. He knew all there is about lean startup. So, he drawn his lean canvas and began to iterate.

How could he know he was going to fail? He did everything by the books. By one book, precisely.

And yet, he has failed. On all fronts, he has failed.

First, he couldn’t find people who’d understand what he’s up to. Iterating seemed pointless to them: they wanted to build a complete product.

Second, when they built their MVP, or Minimum Viable Product, they couldn’t find their early customers. No one wanted what they offered. No one wanted to test it.

The Minimum Viable Product wasn’t the Minimum Desirable Product, apparently.

Then, when they have finally found the customers, they couldn’t get the information they needed from them.

Since the lean startup de-emphasizes the importance of feedback, the developer has to find another way to gauge whether or not his early customers are happy with his products. He has to find a way to gather behavioral information: What are my customers doing with my product? How are they using what I gave them?

Which features are popular – and which ones are not?

It turned out that implementing the system to gather all this behavioral data is actually one hell of a task. Instead of working on the product, the team had to focus on developing the means to monitor how it’s used.

It turned out to be quite expensive as well. Far from everything you may call “lean.”

“I’m so stuck here,” I remember Michael told me on one occasion. He really seemed overburdened. “But,” he added, “there is no way around it, is there?”

There wasn’t – is what he believed.

He kept telling everybody that it’s not the method’s fault they’re running thin on budget and time. That it’s not the method’s fault that there are no customers. “It’s a new thing,” he said. “People do not yet know that they want it. But they want it.”

“We only have to show them.”

But they never did, as soon, the money ran out and they went bankrupt.

They ran split-tests, experimented with the product on as many customers as they could, developed metrics – the metrics did work and allowed them to gather a lot of information – and iterated. They iterated like crazy.

By this time, all of our friends have started to view them as such: as crazy. “Why not simply make the thing and let it find its way to the market?” they asked. “Why do we need all of this hassle?”

But Michael was unmoved. He seemed to know what he’s doing.

Even after they failed, he kept telling everybody that it’s not in the method – the reason they lost it. “We were so damn close,” he told me. But were they?

After they failed, Michael had to sell his flat to pay all the debts he made. He became depressed and broke up with his wife. He has subsequently left the country; he is currently working abroad to pay what he still owes to the bank and loan services.

Was he a poor manager?

Well: maybe. Maybe he was. But even if, his story is not the only one. The Internet is full of stories that resemble it – in detail. Were all those people who implemented lean startup and failed poor managers? No – I doubt it.

Something’s wrong. But what exactly?

Is the lean startup method a scam?

Or is the problem elsewhere?

Who We’re Up Against? The Lean Startup Critics

Some of the method’s critics – e.g. John Finneran – believe that it’s the method that’s faulty.

He proposes another modus operandi for startups:

“First,” he writes in one of his most popular blog posts, The Fat Startup: Learn the lessons of my failed Lean Startup, “pursue achingly high standards in every aspect of your startup. Mediocre execution will slowly murder your startup.”

“Second,” he continues, “narrow the scope of your product until you can develop an extraordinary product. The purpose of your first release (and every other release) is to give your customer immediate value. You are not launching a series of science experiments for you to learn what you should already know.”

“Third, find enough funds for a substantial marketing budget.”

And “[f]inally,” he concludes, “beware of Lean Startup principles, or any other shrink-wrapped utopia offered by the entrepreneurial dream industry.”

Others, however, Michael included, believe that the method is valid. “It’s the execution that’s at fault,” they’ll answer.

Asked about Finneran’s advice, they will tell you that’s exactly what they’re doing (except for the fourth one).

“If I were to do it all over again,” Michael told me, “I’d do it all the same way – except that, maybe, I’d try to secure more funding. But the lean startup – lean startup’s all right.”

And even if the method was “all right” – some argue – so what? Can you imagine building anything that’s truly valuable with it? Can you imagine building SpaceX by iterating and experimenting on your clients?

Certainly not.

That last argument is Peter Thiel’s. Although Thiel is not criticising the method because it hasn’t been validated, he doubts that it’s actually useful – unless you want to build mundane things that won’t really change the world. And the purpose of startups – these most successful ones, he argues – is always to change the way that things are.

Thus, you can go with lean startup. But you won’t go from zero to one this way.

(Going from zero to one is one of Thiel’s most interesting concepts: it emphasizes the fact that the most successful startups always create something new, unique, and strange. Yes – strange. Unlike anything you’ve seen. The most successful startups always go from zero – because there was no such thing before they made it – to one – because now, there is one. The examples are plentiful: Facebook and Google seem to be Thiel’s favorites, with Amazon as their close peer.)

(More on going from zero to one – in Peter Thiel’s book of the same title: Zero to One.)

And even if you could go from zero to one with the lean startup – others still argue – there are much, much easier ways. Going lean doesn’t necessarily mean going lean on expenses. And sometimes going lean and succeeding at the beginning can lead to a failure in the future: it all depends on the competition and what it’s doing.

The importance of this last thing was first pointed out by Ben Horowitz in his 2010 article, The Case for the Fat Start-Up. A number of other articles followed. Their point: overemphasizing the importance of running lean at the early phases of product development can undermine your efforts to take over the market in the future.

This is especially important in the times of crisis, when every company tries to run lean (because their managers think it also means cost-cutting). Following their tracks may save you now – but it will sure put you in danger later. Doing exactly what your competitors are doing may seem like the right idea for the time being – but it’s not the way you win the competition.

So. With so many critics and such strong criticism, is there any way out? Is there any way to save the lean startup?

Is the method valid? Is it useful?

Let me proceed to the answer.

Isolating The Problem

As Peter Thiel put it in Zero to One:

„Every company starts in unique circumstances, and every company starts only once.”

Thus, we can’t simply split-test Ries’ method. We can’t start the same company twice, implementing the method in one instance, not implementing it in the other.

Besides, split-testing would only bring conclusive results if the experiment was conducted in an isolated environment. In this case – for obvious reasons – conducting such an experiment would be impossible. With the number of factors contributing to each business’ success being this large, the business’ world is as far from being an isolated environment as a thing can possibly be.

Not to mention that such tests would be absurdly expensive to conduct.

Hence, if we want to know whether or not the lean startup method is valid, we have to validate it differently.

But wait. Wait. There are people who used it – right? Who started their startups the lean way – right? – and succeeded? What about them? Isn’t it enough of a proof?

Well: it isn’t. Just as Peter Thiel said, every company starts in unique circumstances. It’s hard to tell what exactly (and to what degree) contributed to its success. It’s more like a living organism than a machine. While some things are obvious (e.g. the better funded you are, the better are your chances), others are not (e.g. whether or not using certain method will help you).

To repeat: as long as the method hasn’t been validated, we can’t be sure if the company succeeded due to implementing it – or despite it.

Also, there’s a difference between full-time validation and this kind of part-time, anecdotal validation.

Full-time validation presents (or should present) us with an answer. Either the method works – or it doesn’t.

The fact that we’re not sure if the method works after so many entrepreneurs have claimed that it does signals the need for more thorough inquiry.

But what can we do?

Before I answer, allow me to point out this curious paradox. I have named it „The Lean Startup Paradox.” It should shine some light on the reasons the matters are so complicated here.

The Lean Startup Paradox

So. We have a method: Ries’ lean startup. And we don’t know if it works.

To be more precise: we don’t know if it is good for a company’s founders to implement it in their businesses.

If it works, it’s great. It’s perfect.

If it doesn’t? Well: too bad.

Only that if it doesn’t, we don’t want it in our business – do we? We don’t want to waste our money. We don’t want to waste our time.

Enter The Lean Startup Paradox:

One. We want to implement the method in our startup. But first, we have to have it validated.

Two. To validate the method, we first have to implement it in our startup.

See the problem?

Let’s continue:

Someone may say that this is not a paradox – not really. It would be – if there were no other startups. Or if there were no other startups that implemented the method previously.

There are two answers.

The first one is that they wouldn’t be right.

They wouldn’t be right because – no matter the number of startups that implemented the method and either succeeded or failed – the method still hasn’t been validated. That’s not what we’re asking for.

The second one, however, is that they would be right after all – but not for the reasons they may think they are.

They would be right because – in fact – validation doesn’t work this way.

I will explain it.

What Is The Proper Validation?

This question’s not trivial.

It is one thing to ask for validation – it is another to know what we’re really asking for.

Unfortunately, most people who ask for the lean startup’s validation (or its „empirical validation”) don’t know it.

They all ask for it because they think that’s what they should be asking for.

They’re mistaken.

What they really want is to be sure that the method – when implemented – won’t sink their business.

But, since the method is a theory, they ask for its validation instead. Just as we might have asked Einstein – or someone else – to validate his theory of relativity, they want Ries’ – or someone else – to validate his lean startup.

They want a proof that it works – a conclusive, closing proof. Or a proof that it doesn’t.

We will be back to Einstein and his theory of relativity soon. It actually has a lot to do with what we’re discussing. Before we get there, however, I have to point out one more curious paradox. This one is called „The Sorites Paradox” – or „The Paradox of the Heap.” I didn’t conceive it: it’s actually much older. Much older than me. (It’s even much older than Einstein.) It’s classic.

It goes like this:

One. 1000000 grains of sand is a heap of sand. Right? Right.

Two. A heap of sand minus one grain is still a heap.




1000000 grains is a heap.

If 1000000 grains is a heap then 999999 grains is a heap.

So 999999 grains is a heap.

If 999999 grains is a heap then 999998 grains is a heap.

So 999998 grains is a heap.

If …

… So 1 grain is a heap.


But what does it have to do with lean startup?

Actually, quite a lot!

It explains – partly – why it is impossible to validate the method the way most people think we could. (There are also other reasons – I will point them all out soon.)

Assume that we’ve tested the method in 1 startup.

Assume that it worked.

Is it proper validation? No.

Assume that we’ve tested it in 2 startups. That it still worked.

Is it proper validation?


Assume that we’ve tested it in 2 more. Or even in 20 more. And that it still worked.

Is it proper validation?

So. Here’s the answer. It doesn’t matter in how many startups we actually test the method. It still doesn’t give us what we want it to give us – we know it! But what is the reason?

The reason is simple.

Although we can test the method in 100 startups, in 100,000,000, even, it won’t make it certain that it will work in 100,000,001st.

It can still sink.

When we talk of validation, we want to be sure. If the method is validated, we can be sure that it will work no matter if it’s the 101st of 100,000,001st startup we implement it in.

Just as we can be sure that the scientific theories – e.g. Einstein’s – will do.

Hence, no matter how many startups implement the method and succeed, we still can’t be sure that ours will as well. It doesn’t matter because the proper validation is never about the number of tests. It’s about how these tests were conducted.

The difference between the proper validation and some anecdotal evidence is that the former doesn’t really require 100,000,000 tests but still somehow makes it certain that what we’ve tested is valid.

But how can we do that?

To really find out, we have to take a quick sneak peak… at the history of science!

A Quick Sneak Peek… At The History Of Science!

This thing is complicated. I will make it as simple as I can – without oversimplifying.

If we really want to understand the difference between what’s the proper validation and what’s not, we have to have a better understanding of the history of the concept of validation in general. Why? It is obvious: because in Einstein’s times, for example, the proper validation meant something entirely different that it does mean now.

And it’s not only about the fact that the science has changed.

In the end of the nineteenth century – and the first two-three decades of the twentieth – scientists tended to believe that the proper validation is the conclusive one. Once a theory has been validated, they claimed, it is proven – it is true. We can be sure it is and we don’t have to check it again. The tests have to be rigorous, of course – we need to conduct experiments, repeat them, allow the others to repeat them as well. But if everything goes right and the experiments succeed, we can be sure – we can be certain – that the theory is valid.

Those scientists’ standards were high, but so was the prize. To believe – to know – that our theories are true, that they are the theories – what can be better for science?

Hence, for those scientists, Max Planck and Albert Einstein included, the proper validation meant verification: proving – in a series of experiments – that the theory is ultimately and irreversibly true.

But is the verification in this sense even possible? After the crisis of the Newton’s classical mechanics – which everybody believed to be verified and true – and the emergence of Einstein’s theory of relativity (which only a few trusted at first), the concept of ultimate verification broke down. How can we be sure that our theories can be proven if the ones we believed to be proven beyond any doubt turned out to be false? This question made scientists lose their sleep. There were even suicides.

In 1939 – and then, subsequently, in 1959, translated to English – Karl R. Popper published his now-famous book, The Logic of Scientific Discovery. The Logic of Scientific Discovery is a funny name: the book has nothing to do neither with logic nor scientific discovery. Not in the traditional sense, at least. Still, Popper’s work is now widely considered to be one of the most important books published in the XX century – and certainly the most important book on the topic of “What is science, anyway?” published ever. Why? Because it allowed those sleepless scientists to finally sleep at night?

No, not really.

Instead, it made them sleep even less. It has shown that everything they’ve believed the science to be was, in fact, a false assumption. No theory – scientific or else – can ever be verified. Can they be validated? Yes – to a point. But they will always be susceptible to subsequent falsification. Someday, somewhere, someone may always prove they are false. There’s no escaping.


Because – Popper tells us – this is simply how things are. Theories are hypotheses. And hypotheses may be proven false.

Even if 100,000,000 tests have already proven your theory, 100,000,001st may disprove it.

This is simply how things are.

For example:

The hypothesis is: “All swans are white.”

We can find a thousand swans – each of them white. Will it prove our theory? It will – limitedly. (It will “corroborate” it – “corroboration” is the term Popper uses to describe this kind of limited validation.) But will it verify it? No – it won’t. It’s impossible.

There’s always a chance to find the black swan.

Although the universe may be limited by space and time, Popper argues, we can never be sure what awaits around the corner. It may be a white swan – one more in the series. But what if it’s black?

And what happens then?

The answer to this question is nuanced at its simplest. Let’s say that we’ve found a thousand white swans. So far, our hypothesis was corroborated. But then, we’ve found a black one. The finding clearly falsifies our hypothesis. Do we now hold on to it – or do we toss it aside as disproven?

Popper’s answer is that, although scientists rarely do abandon their theories simply because they were proven to be false, this is what they should do. Instead, however, they set on to “repair” them: they develop the additional “ad hoc fixes” that are to save their theorems. (Only that, in most cases, they don’t save anything but make things messier.) They try hard – they really do. Much often, it’s easier to hold on to a theory we’ve sacrificed half of our lives to develop than to admit that it’s false and move on. “No Nobel Prize for you, Jack. Maybe try another life?”

Of course, it may all sound pretty weird. How can those people not see that they were mistaken? If they see a black swan – it is obvious, isn’t it?

It may be, but in reality, scientific theories are – mostly – much more complicated than this “white/black swan” example. Hence, it’s not about the scientists being unfair – it’s rather that the thing is problematic. (If it wasn’t, why bother writing so many books about it, after all?)

This shift of perspective is of utmost importance. But most people don’t know about it. Most people don’t know “what’s science, anyway,” and believe that the proper validation means verification – or at least corroboration (or some kind of corroboration).

In reality, only these theories that can be proven to be false – that can be falsified – are scientific.


Because if you can’t prove or disprove it – you can’t design an experiment to prove or disprove it.

And if you can’t design an experiment – how is your theory even scientific?

(There’s a number of theories that people assumed to be scientific when, in fact, they weren’t, Popper writes. His favorite examples are Marxism and psychoanalysis. He calls them pseudo-scientific.)

Summing up:

A good theory is the one that can be proven to be false.

It doesn’t mean that it will be proven false. It only means that it can be: that we can test it experimentally.

Now. Can we test the lean startup experimentally?

Testing Lean Startup… Or Can We?

Some people believe that methodologies can’t be tested. That we can’t design an experiment to validate them. It’s not true. Sometimes, we can.

Can we do it in this case?

As I’ve mentioned, we can’t split-test it – not really. We also can’t test it in an isolated environment: no business can succeed in such circumstances. No business can even be started this way.

What can we do, then?

Well: we can analyze the case studies. It’s a lot of work – I’ve done it. If this still nets us no proof – and in this case, it didn’t – we can look for other answers.

But do we have to?

This is where Popper’s theory comes into play.

A number of startups succeeded because their founders implemented the lean startup’s principles – but a number of others have failed. Right?


Is it enough to negatively validate the method (to invalidate it)?

It’s not. It could be – but it’s not.

This is where most people are getting confused.

If we’ve launched our startup according to what Ries’ tells us to do and it has failed – can’t we say that the method’s invalid? We’ve found our “black swan” – haven’t we?

No – we haven’t.


Because we don’t know if the reason behind our company’s failure is the method – or maybe something else.

We could have known it if we tested it in an isolated environment. But this is impossible here.

(To stick to the “white/black swan” analogy, we don’t know if what we found is a swan. Or, maybe, it’s another kind of bird that simply resembles a swan so closely that it mistaken us, initially, but now we have to do some more research?)

Hence, we are left with no answer. Or maybe – just maybe? – there is an answer, hiding in plain sight somewhere? At least some kind of answer?

There is. Of course. Of course there is. The first important thing that we learn here is that these failures don’t falsify the lean startup. No matter how many startups implement the method and fail – the method can still be all right.

And this is something. (For most scientists, truth be told, it would be enough to feel satisfied.)

The con side is that it also works the opposite way. No matter how many startups implement the method and succeed – the method can still be broken. But this, actually, is less of a problem: since we’ve already established that no theory can be proven ultimately and irreversibly true – there’s no love lost here.

So what’s left?

We are left with a method that can neither be verified nor falsified. Not in the proper sense.

It can’t be verified because nothing can be verified. There’s no problem.

It can’t be falsified because it can’t be split-tested in an isolated environment. To put things in more technical terms: there is no way to design a potentially falsifying experiment.

Is this a problem?

Well: it can be. But if it is, it would be more of a problem if the method couldn’t be tested at all – and this one can be. As long as we can implement it in our startup, we can test it as well.

Only that our conclusions will have limited importance.

Can we possibly extend the importance of our tests so that they will appear as conclusive and falsifying?

It depends.

Let me explain.

(We’re nearing the end.)

One More Sneak Peek At The History Of Science

If you’ve been reading carefully, you already know the answer.

If not, let me help you.

We’ve said that no theory can be verified.

We’ve also agreed that some can be falsified. And if they are, they should be abandoned. Yet, they rarely are, as most scientists grow attached to them and try to repair them, sometimes at all costs.

We’ve agreed that this is perfectly normal and understandable. It would take a sociopath to simply abandon his life’s work – at least to abandon it in this manner.

Hence, we know that no matter how many experiments we conduct, we will always be able to cling to our theory, even if it’s falsified.

Besides, in many cases, it’s not only understandable, but rational not to abandon the falsified theory. Scientific theories, as some scientists argue, are “born” falsified: they are born imperfect and full of mistakes. And yet, they prevail – while their better corroborated competitors fail.

An example: Einstein’s theory of relativity was – at first – deeply flawed.

It took him ten years to perfect it. It took him more. And then, even more to make it even better.

And even then, there were those who would doubt it. (One of them was Niels Bohr.)

Why is that?

Obviously, because it takes time to come up with a theory that is able to explain as much as the theory of relativity and make it one hundred percent correct and coherent.

Less obviously, because our human minds are imperfect – and thus, so are our creations, scientific theories included.

But even less obviously, and this is the answer, because no theory needs to be 100% correct and coherent to do what it was designed to do.

And none is – nor ever was. Nor will (probably) ever be.

But what is the reason to choose these underdeveloped theories over these fully-developed and well-corroborated ones? What was the reason to choose Einstein’s theory over Newton’s one? While Newton’s – although visibly flawed – was still holding pretty strong?

The answer?

It’s the potential.

Although there were people who opted for Newton’s, some saw Einstein’s theory’s incredible potential – and decided to give it a try. Just as we may see that a startup has the potential to change the world (and thus, include it in our incubator’s program). It is never without risk.

But what we have to understand here is that those weren’t random people. Those were the most clever, intelligent, well-educated, far-seeing people on the planet. Nobel Prize laureates.

Their reasoning was as follows:

The Newton’s theory is well-developed, and the Einstein’s one is not. Yet, the Einstein’s nets potential to explain the things that Newton’s one never will. Will it explain them? Who knows. But its competition won’t, and that was what they knew for sure.

As for the theory, it might be flawed and falsified already, even before it’s fully developed, but it can be made better. Much better.

Besides, we’re choosing between two flawed and falsified theories here, right? The Newton’s mechanic is flawed as well. And it has been falsified, too.

Let’s give it a shot.

That’s, basically, their reasoning.

What’s the lesson here for us?

The Lesson Learned

The lesson is simple.

Imagine that you have a car.

Every day, you have to get to work. You go there by car.

Your car is broken, it’s barely running. It’s so frustrating to even drive it. How you wish to finally trash it!

But can you trash it? Now? Right away?

No – you cannot. You still need it. And so, you drive it.

You could trash it – but you’d need to have something to replace it with. A new Tesla, possibly, or Audi at least.

The way things are, you’re forced to use it. It’s not that you can’t abandon it – it’s only that it would be much more of an effort to go places without it.

You’re stuck with what you have. Until you have something else.

The same applies to the scientific theories.

We may know that our theory is invalid. It might have been falsified a number of times. But – as long as there is nothing to replace it with – you’re stuck with it. And looking for another.

The theories may be falsified and, subsequently, abandoned, but scientists generally try not to ditch them until they have an alternative.

The more promising the alternative, the more eager are the scientists to abandon their old ways and invest their time and careers in the new ones.

In cases such as this, when there is a theory that clearly overshadows its predecessors in terms of potential, scientists may extrapolate the results of one of the falsifying experiments they conducted and make it conclusive.

It’s always a process. (And, in part, it’s always arbitrary.)

When this happens, and it never happens overnight, the importance of the test is extended and the test is deemed conclusive, the theory – proven false.

If we’d want this to happen to the lean startup, we would need an alternative.

One that would be clearly better.

Yes, there are alternatives. There is Agile, there’s Design Thinking. But all these methods are fallible as well – and besides, they’re surprisingly similar to the lean startup.

(They are so similar that people still argue if the lean startup is – possibly – just a version of Agile.)

If we don’t want to start our business the traditional way, then, we have to use one of these.

And, for startups, the lean startup stands out, and not only because of its name. Just as Design Thinking stands out for creative projects, Ries’ method simply “fits.”

It is not flawless but it has a lot of potential. And it certainly provides its users, all these startup entrepreneurs, with a framework that allows them to better understand the startup community: the way it works, the way it doesn’t.

This is where it’s better than the traditional management theories. And this is the reason for its appeal.

The Conclusion

The conclusion is of double nature.

One: the lean startup method has flaws. But we won’t be able to eliminate them simply by asking for the method’s validation. No amount of validation will give us what we really want: certainty that our business will succeed. No amount of experiments will take us there – and even if split-testing the method in an isolated environment was possible, it’s doubtful that the results would be anyhow satisfying.

This answer may seem trivial, but, in fact, I find it deep. It shows that many of us don’t know what we’re asking for – and yet, we’re eager to blame the method when we don’t get what we wanted.

It also shows how the problem was misstated. If nothing else, now it’s been re-established, and properly.

Two: the lean startup method is not universal. It may be better not to implement it if we’re launching our business in the times of crisis. And it certainly won’t help us to beating the competition. It probably won’t allow us to save the world the way Peter Thiel would like startups to try – by going from zero to one and inventing something new, something that will serve the betterment of all of our lives. Here, however, we can’t be sure. Thiel will tell you that you cannot build another SpaceX while going lean – but Ries’ followers will answer that’s just what Musk did. (And that especially Tesla was running pretty lean at its beginnings.)

So. Is the lean startup method valid?

It is – but not in the way we’d like it to be. But then, nothing is – so how big of a problem can that ever be? One day, some better method will surely replace it. For now, it’s what we have. And it’s not bad.

And it certainly isn’t a scam. While it has its flaws, it also has certain advantages, and you may want to use them (depending on what you’re building). The evidence provided by Ries’ may indeed be anecdotal – but there’s much more evidence outside of his book, blog posts, press articles, and keynotes that shows that it works.

Yet, does it work the way it should?

Often, it does. (And when it doesn’t, there’s probably a reason.)

Well then, finally, one last thing: can implementing it sink you (altogether with your business)?

Yes, it can. Of course it can.

But then, what cannot?

Starting a startup, you know it is risky.

And Michael knows it as well.

One day, when he’s back, he told me, he is going to start anew, to try again.

He already knows the method.