Steve Blank is a Silicon Valley entrepreneur, an author and teacher, and the originator of the lean startup movement. He is an adjunct professor at Stanford University, and a Senior Fellow at Columbia University.

In this discussion with Martin Reeves, Chairman of the BCG Henderson Institute, he discusses the lean startup method, innovation, differences between large corporations and startups, strategic ambidexterity, and the disruption and re-imagination of companies in the context of the COVID crisis.

Could we start off with the idea that you’re most famous for, the lean startup method? What’s the essence of the idea?

The essence in one phrase is there are no facts inside the building, so get the heck outside. It started from my observations about startups. Most investors, without ever saying this exact phrase, basically treated startups like they were smaller versions of large companies. And they told the early stage entrepreneurs that large companies wrote business plans, so you should write one. Large companies did five-year forecasts, you should do one. They were assuming that startups were smaller versions of large companies, when in fact they weren’t. Large corporations became large because they already had, and had discovered, a series of knowns: known customers, known pricing, known competition, known business model. Where large companies were executing that business model, startups were actually searching for a business model. And we had built a hundred years of management tools for execution but had very few tools for search.

So is that the essence of the difference between large and small companies, in your mind? One is searching for the big thing and the other one is optimizing the known big thing?

That’s correct. And of course this thinking now applies to innovation inside of large corporations too. Just thinking that you could run the same product management tools that you use for follow-on products or additional features for existing customers to create new segments, and new markets, and new activities was simply incorrect. More importantly, in the last 10 years, as companies started to be disrupted by startups, these tools became more and more important.

So lean startup was the answer to the question, what kind of tools do startups needed? And it consists of three components. One is something to help you get out of the building. I invented a methodology called customer development, which is kind of a formal process for how to discover and validate customer needs, regulatory needs, and so on.

The second piece, which one of my students, Eric Ries, discovered, was that the methodology of building products had changed radically in the 21st century. Instead of building products serially, that is specify all the features, hand them to engineering as a functional spec, wait a year and a half for the product to come out to the other end, in the 21st century, there were enough software and hardware tools to build most products incrementally and iteratively. So you could get feedback immediately with something called the minimum viable product.

And then lastly, Alexander Osterwalder designed something called the business model canvas, which allowed you to specify about 80 to 90% of all the things you need to know about a new venture, whether it’s internal to a large corporation or a startup.

And those three components allowed us to build a very different methodology for developing and testing new ideas, and more importantly, reducing the infant mortality of those ideas, and finding out whether you were right or wrong with minimal resources.

And that changed the shape, and trajectory, and speed of how we build startups. And by extension how we built innovation inside of large companies.

Just to play devil’s advocate: we do have customer analysis in classical strategy, business model innovation would be called strategy in a large company, and in terms of the agile approach, large companies have often experimented with methodologies like process acceleration and time-based competition. What’s the essential difference between your large and small company approaches?

There’s one other part of lean, which is critical, which impacts culture. And that is when you follow lean methodology, if you get enough data to say that some of your assumptions are wrong, you’re allowed to pivot. A pivot is a substantive change to any one of the business model components. Say you discover, we’re talking to the wrong customers. Or usually what happens is we’re talking to the right customers, but they don’t want all 12 features, they want features three, seven and nine. If you have the authority to rapidly change what you’re doing, you can move like a blur relative to competitors. In a startup, the process is run by the founders, who have complete authority and autonomy to make these moves, and in fact have to make these moves for survival, as they search for product-market fit.

In a large company, sometimes you’ve already committed features, or the culture doesn’t allow you to change product design so radically, or strategy is separate from engineering, and separate from all the other groups responsible for supporting product development and the financing of innovation. So lean startup is not only about rapid processes, but about the autonomy to make pivot decisions.

So if I’ve understood correctly, the key features of the idea are: the authority to make radical changes necessary and a development process predicated on frequent change?

Yes. In large companies, if you get it wrong, you might still keep your corner office and secure another project. In a startup, if you aren’t moving fast enough and finding product-market fit, you are out of business. And so that motivates rapid learning, because it is literally a matter of corporate life or death. That impacts how you design your culture. There is no “Mother may I?” or “Can I attend this conference?” or “Can I get on this airplane to go visit a customer?” or “Can I buy this equipment?”.

In a large company, we have rational rules for HR, and finance, and procurement, that make sense for 99% of the company, but which actually inhibit and sometimes even strangle innovation in its crib. The same processes and procedures that work great for execution are often antithetical to inspiring innovation.

Let’s double click on that, the applicability to large companies. So on the one hand, of course, large companies also have to reinvent themselves, also require innovation. On the other, you said that a startup is not just a small version of a big company, so there must be some essential differences. How would you assess the applicability of the lean startup ideas to large companies?

My work ended up on the cover of the Harvard Business Review in May 2013, in a piece called Why the Lean Startup Changes Everything. From around that time large companies were are being disrupted by events outside of their control — the emergence of China, huge capital being poured into startups — so they started looking at tools like the lean startup. And incubators and accelerators started popping up into every large company. We have created a ton of innovation theater, but I’m not sure we moved the needle on increasing revenue and profit in the line with all those investments.

And is that because people have misunderstood the ideas, or because the ideas require modification for application in a large company context?

The latter. A large company is not a bigger version of a startup. In a large company, more than ninety-nine percent of your people come into work to execute repeatable tasks, whether they’re physical tasks on a manufacturing line, or executive tasks in sales or software, or something else. But they’re not there to create new products and services, because to do that, you require different organizational structures. You and others articulated the notion of an ambidextrous organization — one that can execute and innovate at the same time, and it was kind of a “nice to have” theoretical activity for most companies, but in the 21st century, I’ll contend we really need to not only think about it, but embed it in corporate culture.

And that takes major changes in leadership, C-suite design, in process design and for the rest of the company. You just can’t assume that you can set up an incubator inside and then watch other divisions adopt its output. In fact, most of the stuff that comes out of corporate incubators just falls on the floor, because you didn’t do the hard job of wiring the rest of the organization to embrace its output.

So let’s talk about ambidexterity, the ability to both explore and exploit, or innovate and execute at the same time. That’s clearly a very valuable thing to be able to do, that’s the essence of reinvention, but why is it very hard for large corporations to do that? What are the typical failure points?

We see all the time in large corporations that a group that comes up with something new and disruptive and the CFO says, “Well, give me the five-year forecast.” That’s like a divide by zero problem. That market doesn’t exist. So the response is, “Well, we can’t fund anything like that.” Okay, but you might have just decided not to fund antigravity!

If you fund projects like that, the only structural way to deal with innovation is to acquire it expensively outside, which might be a perfectly fine model if you can admit that you’re not capable of engineering your internal processes to deal with disruptive innovation. It’s the companies that basically don’t understand that innovation doesn’t come with a memo that says this will be a radical change in five years, that are going to run into trouble.

The people who do innovate have need to have a different culture from the people who execute. That sets up a whole set of internal issues inside of large corporations. A hundred percent of a startup are crazy people in the early days. They all share a view that they’re going to change the world. That’s not true in a large corporation. While you might share the vision and mission within the company, you don’t want 99% of your company doing crazy things. But if you don’t have 1% of them doing that, at least in most markets, you’re going to be out of business.

So let’s talk about strategies of ambidexterity. At one extreme, the strategy could be don’t even try. Large companies exploit, and that’s why we have startups to develop new ideas. A slightly less extreme version of that would be don’t try to do anything radical within the core of the business but start a new subsidiary or a new venture. At the other end of the spectrum a more idealistic solution would be to have everybody be ambidextrous. In practice, from your work with large corporations and governments, what are the practical solutions to getting to ambidexterity if you don’t have it?

Yes, there are a couple models and you might need more than one. To extend a current business model, to create additional features for existing product lines, there should be a continuous innovation process that exists inside existing functional units or P&Ls. It should be a formal innovation pipeline, which allows innovation to occur not only through the traditional engineering methods, but might allow that division or functional unit to acquire.

But something else that is important for long-term success for companies, is to be able to create disruptive ideas that don’t fit into existing P&Ls. And that could be through acquisition, but it could also be through standalone groups that are basically run out of corporate. And that’s what can confuse people — that you may use a similar innovation process, but depending on the horizon or time period, they either go inside of existing P&Ls and functional units, or go inside of corporate. And again, ideas could be either internally developed, or they could be externally acquired.

In either case, innovation needs to be an end-to-end process, parallel to execution, whether it’s inside a division or being run out of corporate for something longer term that’s disruptive.

What you’re saying is clearly desirable, I’m just wondering how feasible it is. So let’s take, for example, disruption by digital platforms. In theory, we could say large companies should be embracing self-disruption and building digital platforms and avoiding disruption by attackers. In practice, the majority of the dominant digital platforms are owned by digital native companies. Do you have any good examples of companies that have come from a traditional, non-digital world, and a mature business model, and who have embraced ambidexterity and digital business models successfully?

The classic example that’s now a couple decades old is Apple. That was a pure hardware company and invented a series of marketplaces — not only embraced them but created iTunes, and created the iPhone, and created a whole set of digital activities — not just adopting them but leading the world.

I work with different branches of the US government, and there’s no optionality to this. They need to move past a hardware world, where software was an add-on, to now a software world, where hardware’s an add-on. And that’s occurring in every platform they have. I think we’re seeing the same with automobiles. People have been looking at Tesla as a car with a different type of engine, but in reality, a Tesla is a computer with wheels. I think we’re seeing the same with Elon Musk’s other company, SpaceX, as people look at it as a reusable rocket, and they’ve kind of got it backwards. It’s basically software embedded in hardware. You couldn’t land those things on a drone ship without embracing digital and computing from day one. And the hardware’s wrapped around those insights.

And this really requires, as I call it, an innovation doctrine. The ability to integrate all these strands together. But in some companies, the board and the C-level staff are spending most of their energy fighting off activist investors. So while all this is nice, you’re busy and occupied doing other things. And that’s the first thing you need to check, is do we have the bandwidth to do this, or are people looking for immediate results that pump up the stock price?

Let me ask you about that. In a sense, you’re talking about the allocation of attention. If you don’t pay attention to innovation, then it probably won’t happen. One surrogate for that is what you measure. So many large companies are focused on financial outcomes, and productivity ratios, which of course measure what happened in the past, not the potential for prospective growth. What should companies be looking at to see whether the volume and the quality of their innovation activities is up to the task?

It used to be that startups followed large company trends. Nowadays, it’s the other way around. One of the things I’d be asking C-level executives to talk about is, tell me all the startups that are being funded in your area? And where are investors putting early-stage capital? And if you don’t understand that, you’re missing important early warning signals.

So share of innovation capital flow , or share of voice in new investment, or something like that?

Sure. Maybe the people outside the building have some insight about why they think they’re going to disrupt you. You ought to be able to have that conversation. And I’m not suggesting you follow them, but you ought to be able to understand what they see, and decide whether they’re right or wrong.

And that brings up other things you ought to be looking at besides financial metrics like the rate of change of technology in your industry. And I don’t mean just for technology companies. What are some enablers that might change how your customers will buy, how will they hear about products? If you have physical locations, are those going to change? Are the demographics changing, or is the geographic customer base changing? And by the way, I think the pandemic is going to create a lot of lasting change.

Let’s go there. Most of what we’ve talked about so far has been technological disruption, we now face another type of disruption, and we’re even seeing companies talking about reinventing themselves. How do the lean startup ideas play into the disruption and re-imagination of the company in the context of the COVID crisis?

I think it’s an important tool. It’s obviously not the entire tool set, but it’s a key tool to rapidly test new business models. But I think before you do that, you have to have some key assumptions about what has changed, and be able to test those. I think, at least for me, for most companies, it’s pretty obvious that there are some real demographic changes. People over 60 or so, who are more likely to get sick and die of the virus might have different behaviors, even in the recovery, until there’s a vaccine. People in urban areas are going to be avoiding dense gatherings, not all of them, but enough to significantly to affect your business so I’d be thinking about demographic and geographic changes.

And then I’d be running experiments to see whether can we offer different services depending on some of these findings. Where you offered the same thing in every city or country, and we value the fact that we were uniform, we might be thinking about changing that.

I think the other thing about the pandemic is that it exposed, excuse the bad pun, billions of people to the future in a way that we didn’t expect to happen for a decade. Meaning almost everyone who does business or talks to their kids has now been on some type of video conferencing software. Lots of businesses have now discovered that, even for the first cold call, you don’t need to go physically see someone; a Zoom or Microsoft Teams call is just fine.

I think with people still wanting to work at home, offices are going to be reconfigured and commercial real estate will take on different meaning. I think medicine is changed forever due to telemedicine. We could go on and on. I just think that every corporation ought to figure out how much of this is going to stick, and how do you take advantage of it as the economy recovers.

That’s a very important question you just touched upon there, which is of course we see a lot of transient change — the acceleration of demand, the postponement of demand. But there are some more persistent changes too. How do you distinguish between the two?

I think running some experiments offering subscription services for remote x and y, and seeing if people will sign up for anything long term is a good way to test this. You can run some very inexpensive experiments about some hypotheses you have — and here’s a key idea -they don’t need to be across your entire business. Let’s design an experiment, let’s run the experiment, get some data, derive some insight, and either modify, validate, or invalidate the hypotheses.

Let’s wrap up with a very provocative statement I read in one of your articles recently. You said that a company’s survival in a downturn can be captured by a very simple formula. Survival is the product of your speed of understanding the situation, the magnitude of the pivots that you make, and the speed with which you’re willing to make the required changes. Could you expand on that a little?

In March as we were shutting down, speed meant your ability to constrain costs and expenses, as revenues were dropping, in some segments radically, and now during the recovery, I think speed works the other direction. How quickly can you recover in the areas that people are beginning to buy, or come out to shop, or whatever. More importantly, how quickly can you take advantage of those changes in customer behavior that will be sticky, or that you could make sticky, that people realize might be more convenient?

Whether that’s in a B2C business, or B2B business, this allows you opportunity to lead rather than follow. And to do that, you could run some very rapid experiments to figure that out without moving the whole company and lurching in one direction or another. And the good news is, no one knows the answer yet. The even better news is you could be the first one to find out!

Thank you so much for spending your time with us today Steve, and for talking through your ideas on entrepreneurialism, innovation, ambidexterity, the rejuvenation of large corporations, and finding opportunity in the COVID crisis. I’m sure there’s a lot of very valuable grist for thought here for our audience, so thank you again.

Thank you for having me, this was great.