a16z Podcast - a16z Podcast: Competing Against Luck

Episode Date: September 2, 2017

with Clayton Christensen, Marc Andreessen, and Steven Levy In business, mistakes of omission may be just as bad as (if not worse than) mistakes of commission -- simply because of the loss in potential upside: new companies, new products, new opportunities for growth. Or even in the ability to respond to the disruption coming to one's industry and company... if it hasn't already. Sometimes, and in certain industries (such as hospitality and education), it just takes longer to pull off. But it's not like people and companies are dumbly sitting around waiting for disruption to happen. In fact, having read the book on disruption for years -- 20 years, to be precise, given the anniversary of The Innovator's Dilemma this year -- many smart business leaders know it could happen, yet fully determine that it's not going to happen to them... and then, of course, it still happens, observes a16z's Marc Andreessen. Why? Part of the answer, shares father of disruption theory and Harvard Business School professor Clayton Christensen, is they don't have a common language, logic, architecture, way to frame the problem. And that's where other theories and frameworks -- like jobs-to-be-done and modularity -- come in. A theory, after all, though never perfect can help. So in this episode of the podcast -- from our inaugural a16z Summit event -- Christensen and Andreessen (in conversation with longtime tech writer and Backchannel editor-in-chief Steven Levy) share their thoughts for how such theories can play out practically in both managing business, and managing priorities in life.

Hi everyone, welcome to the A6 and Z podcast. I'm Sonal. This weekend's special episode features Mark Andreessen in conversation with Clayton Christensen, father of disruption theory, professor at Harvard Business School, and author of several books, including his most recent competing against luck. The conversation took place at our inaugural A6NZ summit, is moderated by long-time tech writer and editor-in-chief of Back Channel, Stephen Levy, and covers everything from management to product to modularity. But they begin with Clay's book, The Innovators dilemma, which popularize the theory of disruptive innovation and actually turned 20 years old this year. I'm wondering, have your views changed about how companies get disrupted, how big

companies can't stay off innovation? What's, a lot's happened in 20 years, especially in the technology realm. What's different? And have you adjusted your thoughts? It's changed a lot. I think the most important thing I learned is I learned a bit about what is a theory and how is it developed. So theories, when they emerged from the researchers' minds, never did they emerge perfect. They were always just kind of half-baked. And what success was predicated upon is subsequent researchers had to find things in the world that the theory could not explain. And if you can find an anomaly that your theory can't explain, it gives you the opportunity to improve the theory. And so that's what happened to us over the last 20 years. What were some

of the big anomalies that caused you to adjust the theory? Well, one of them is, you know, the diagram of disruption where there are two trajectories. One is the trajectory of improvement that customers can use. And another is a trajectory of improvement that innovating companies provide. And I can't tell you how many millions of slides I projected that shows that trajectory of technological progress

almost always outstrips the ability of customers to use it actually isn't true. And where that came from is one of my students said that hotels aren't disruptable. And in fact, if you look across history, there had been no disruption of hotels. The trajectory of improvement had been flat. Somebody else said a bit later, there hasn't been any disruption in higher education. And so for a while in my mind said, well, I guess what's

that means is disruption applies to some industries and not others. But then Airbnb came along. Now here is a new business model. And all of a sudden, what was flat went boing and disrupted them in a powerful way. If I'm recalling your theory correctly, it's almost a classic disruptor in that it started off on the low end, you know, giving convenience. you know, and serving a market that the bigger place seemed not to be interested in, right? So maybe you were just late rather than it being an anomaly or just the innovation didn't come about yet. So I think what we can say is that the processes and the pressures on companies to improve is always at work. But nobody in the hotel industry thought that anybody could disrupt them, the incumbent.

were so powerful. And it turns out that they're wrong. So you can predict that somebody are going to try to disrupt, whether they have been able to do it is still something that we need to study. So Mark, it was a couple of years before Clay published his book, that you were at Mosaic Netscape, but after the book, you kept disrupting with the Upsware and now the venture capital world. How did you come across Clay's work? Has it shaped what you do or your thinking, or basically just makes you confirm what you're doing along?

When I first entered business, I didn't know anything about business because I'd never taken a class on business. Not that I ever could have actually made it into the admissions process of Harvard Business School. So I didn't really know anything. I had a number of coaches and mentors who could teach me things, but I wanted to learn more. And so I went and I read about 1,000 business books and then fed about 998 of them into the shredder because it's just, it's extremely hard to learn business from a book. A big reason for that I found is a lot of business books kind of fall into two traps. One is they extrapolate off the recent past, and so they're very kind of short-term in nature. It's whatever worked in the last five years projects forward.

The other problem they have is I think they confuse correlation with causation. And so they'll take a set of companies that are doing well and an attempt to reverse engineer the causes for why those companies are doing well, when in reality, generally what you're seeing are the effects of what happens when companies do well. There were only a couple books that really stuck with me, and it's really Clay's book. and then, of course, Andy Groves' book. Tons and tons of books will tell you, a company X screwed up because they were incompetent and dumb, to which apparently the lesson from that is don't be incompetent and dumb,

which is, I guess, a useful piece of advice. But the thing with Clay's book that was so striking was the observation that when big companies get disrupted, it's not because they're incompetent and dumb, in fact, quite the opposite. It's because they're very competent and smart. All the existing commitments that they've made, all the existing customers they're servicing,

the existing margin structure that they have. And for me, that was a big mental left turn, which was like, oh, that. explains why all of these incredibly hard-working, sharp people end up in this trap. And of course, what's striking, and I'd love to know actually what you think about this, is you can read the idea, you can know that it could happen to you. You can like fully determine to yourself that it's never going to happen to you, and then yet it will still happen to you.

After 20 years of watching people try to cope with this, do you think it's possible to know that that can happen and then be able to then alter behavior in order to prevent it? Or do you think it's simply fate because of the nature of the fact that the companies end up with so many obligations? So I believe that it doesn't have to kill you, but it'll kill you if you don't understand how the world works. In other words, these processes in the established companies. Can I tell you a story that taught me a lot about this? So in 1999, out of the blue, the phone rang,

and Secretary William Cohen, who was the Secretary of Defense in the Clinton administration, called me. Never met the guy. But he said, I read this theory of you from your work called disruption. And I wondered if you could come down to the Pentagon and explain what that is to my staff. So I described this process of disruption.

And then I said, I'd like to describe how this process of disruption worked its way through an industry that is totally different than yours, just so that you can visualize it in the abstract sense. And so I told him about how disruption occurred in the steel industry, because in the steel industry, the simple product at the bottom was rebar. The high end of the market was sheet steel that used to make appliances and cars. And then the integrated steel companies like US Steel and Bethlehem were making the whole range of these products. But in the late 60s, a new technology called mini mills came in at the bottom of the market melting scrap in electric furnaces. And these mini mills got better and better and better.

all but one of the integrated steel companies like Bethlehem Steel, all but one of them went into bankruptcy and died. And the mini mills now account for nearly 70% of the market. So I described this is the way this had happened. And then a man who was sitting on the front row raised his hand. And he said, Clay, you're clueless about why we're interested in this, aren't you? And I said, I actually am clueless. And he said, well, let me show you. Leave that steel diagram on the screen. And he said, what you call sheet steel,

it's the most demanding market in steel. He crossed that out and he said, we call that the Russians. And what you call the integrated steel companies like Bethlehem Steel that make everything, we call that the U.S. Department of Defense. And what you call rebar at the bottom of the market, we call that terrorism. And what you call the mini-mills screaming up market, we call that non-nation nations like al-Qaeda.

And he said, so that's why we're interested. And then he sat down. And then all of the hands went up. And most of the questions were, have there been any instances of where the incumbent, in leader got disrupted from below, but rather than having it killed them, they went down and co-opted them. And I said, there had been a few examples, but in every case, the leader who caught this next wave had to set up a completely different business unit and manage it independently

with different processes, a different architecture or economic structure. and those have been the only instances where the leader state is the leader. Two months later, Secretary Cohen called me up and he said today we're announcing in the Pentagon that we're setting up a fourth armed of our services. We have the Army, the Navy, the Air Force, and Special Forces Command. And it's organized independently so that we can have different processes and a different structure so that we could go after terrorism we have been trying for six years to figure out how to organize our department of defense so that we could counter terrorism but we just kept

fighting against each other finally you had a model and it gave us a common language and a common way to frame the problem so we could reach consensus around what otherwise would have been a very counterintuitive course of action and the reason for the long story is i think when the established leaders have a tough time capturing disruption even though they they think they understand it what makes it hard is they don't have the common language in a common way to frame the problem the interesting thing about the department defense story in that case which i hadn't heard before is the idea of special forces was not new right delta force for example been around since the vietnam war delta force was part of the army yes delta force was reassigned to socom when socom was created

along the navy seals and the other and so what you're saying is basically socom administratively was then going to run these units that had previously existed in a different way that's right and the core problem was allocating resources from these guys versus these guys and they could ever put it all together in a way that met the strategy that they needed to implement. Increasingly, there are companies that are saying, oh, we get it, we have the architecture in the logic to frame it. And so it's not as lethal a disease as it used to be. Well, this is fascinating because putting this in a theoretical construct sort of confounds our impression that disruption is just so tied with technology itself, right? And, you know, Mark,

is that at odds with sort of the theory there? Talk about your view of the product cycle theory of innovation. We'll take what we can get from a theoretical standpoint. Anything that makes what we do seem that random is tremendously reassuring and ideally helpful, which disruption certainly is. A couple things I've observed, which is just the way that people perceive disruption is that it's the tech industry attacking other industries, right? And so it's the tech industry attacking education, or it's the tech industry attacking financial services or the tech industry attacking newspapers.

In reality, the main form of disruption that happens in the tech industry are tech startups attacking tech incumbents. You know, you get these giant, you know, hugely successful tech companies, and then you get a thousand startups coming at them. That's one interesting thing. Second interesting thing is there should be some sort of, there needs to be an equation or a constant of the amount of time that you get to call yourself a startup before you yourself become right for disruption.

This is where I connect Clay's work to Andy Groves' work. The way Clay talks about when companies are in a position to get disrupted is when they have an existing business, existing organizational structure. Andy always talked about CEO of failure and success having a lot to do with commitments, which is sort of the human version of the structural argument. The commitments that the CEO makes to those customers or to the investors on things like margin structure or to the employees on things like strategic plans. and Andy always argued that too many CEOs get hung up

and effectively get subject to disruption because they end up basically overcommitted but then to actually account for it requires breaking too many promises and CEOs can't break the promises which is another way of saying they can't re-engineer what they do at the core of the business and so one of the things that I find striking is because hopefully we'll be working with a lot of our companies here in the room for 10, 15, 20 years

about five years in you just start to see startup CEOs that have actually made a lot of commitments to a lot of people and have built a lot of assumptions into the business. And then there will be other startups that come along and all of a sudden don't respect any of those commitments and any of those obligations and any of those structures. So I actually think the disruption cycle is playing out earlier. I don't know if it's earlier than it was in the past,

but I think it's playing out earlier than people believe. I never thought of in those terms, Mark. I think a reason why you see the acceleration is more than I remember, you know, there are two different types of architecture. in products. The closed interdependent architecture is interdependent because if I change one thing, I have to change everything. It takes a lot of time to change a fundamental architecture in a product. The other type of architecture is an open or modular architecture where how the

pieces of the system fit together are standardized. Modularity enables new products to be develop so much faster, the gravity is shifting over to markets that are optimized by modular architectures. And in new products and new technologies, very often, it's still, it's in a stage of interdependence and it's slow. So we can't say that the whole world has flipped. But little by little i think the evidence or the weight is on modularity and therefore fast recycle times that's right right that's right i'm just kind of curious um whether the founder if the founder is the CEO when the five-year twist comes or whatever is the company better repaired generally the thing that founders have is literally the ability to break all the promises the founder can say i know i told you

last tuesday we're building you know widget x and we're doing the following way with his margins but i've got news for you. We're now doing building widget Y in this totally different way. If you don't like it, go get another job. And there's this mystique where the employees tend to give the founder, especially a successful founder, the benefit of the doubt. And they say, you know, basically they say, you know, maybe he or she is wrong, but maybe there's, you know, magic happening here. Now, I don't believe it's just the founder. Like the founder is special because the founder is literally the name of the door effect, start the company. There's also such a thing as a founder mentality. And this is also something I think Andy Grove himself believed as well, which

is that it's not just founders who can do this. It's CEOs who have the founder mentality who can also do it. And there are some spectacular examples in the history of business where non-founders running these companies were able to do really radical things that are normally only associated with founders. I agree. You could add one other element that could make a founding management team better or worse at this. In accounting, we teach people that you should ignore fix and sunk costs and only look at the marginal cost and marginal revenue associated with the proposal that you're making a decision about. And that idea that you should ignore sunk costs and just look at the marginal cost creates

real problems for the management team. So by analogy, we wrote a case about this in 1992 and that mini mills had driven the integrated steel companies out of every tier of the market except sheet steel at the high end of the market. The management at U.S. Steel decided that they needed to make

a mini-mill to start to attack against these mini-mills. And then the CFO came into the conference room and he said this makes no sense at all. Because when our

average cost of making steel is $350 per ton. We have 30% excess capacity in our existing mills. If you guys want to sell an additional ton of sheet steel, do it in our existing mills, because the marginal cost is only $25 a ton. And so they didn't do it. And the mini mills kept gaining market share.

last December, U.S. Steel, 50 years later, announced that they were going to build their first mini-mill next in Birmingham, Alabama's regular mill. And then in January of this year, they announced that they weren't going to build a mini-mill. And the reason is that they had 30% excess capacity in their existing mills. And if we're going to do it, let's do it. Let's utilize what we have. And to me, what enables or makes it impossible for management to make a good decision is this marginal versus full cost pressure that forces people to want to use what you have rather than create something new.

And in order to break that, you've got to have the mind. mindset that you're describing. You have to be willing to withstand the attacks that will come from people who will say you're being grossly irresponsible. Yeah. Because how can you ignore this huge install base that we have? How can you ignore this huge customer base we have? That's right. I don't want to go to too much part of returning to the theory that you talk about here, the jobs to be done theory.

I want you to talk about it a little more because it almost seems like if you do this theory right, you're inoculating yourself against disruption there. What does it mean when you say to have the people in a company think of their product, their service as something to be hired and have a resume and apply for a job? Well, what pushed us into this direction is you've seen in, I don't know how many Jillian management meetings where you produce in your best case and we need this amount of money. It's going to hit these milestones this year and next year and so on. And then about 80% of the innovations that get that funding from the corporation, you then deploy it into the market and they fail. And why is it that you can't be correct more than 20% of the time? Is innovation simply a crapshoot that failing is intrinsic to something?

success or not, we taught that we should understand the customers. And if we understand the customers, we'll develop products that they will buy. But the actual reality is that we get it right 20% of the time. So we decided that understanding the customer is the wrong unit of analysis. But rather, jobs arise in our lives. And we have to get these jobs done. And some are incremental, little jobs, some are dramatic. When we realize that we have this job to do, we have to go out and find something to get the job done. Understanding the job is what we need to understand, not the customer, it's the job because the job is the causal mechanism that causes people to buy out and pull it into

their lives and use it. And understanding the job is the critical unit of analysis. analysis. So here's an example. Good friend comes back into town with this spouse and they want to take you out for dinner on Saturday night. And you say, that's a great idea. We'd love to do it. Let me check with my wife. She said, don't you remember we had this commitment on Saturday night? Can they do it on Friday night? So you call them up. Yeah, they're available when? Well, let's do it at 8 o'clock. And then you call back, what kind of ethnic food do you like? And you get there. So let's go to this restaurant. You call them, they don't have space for you. So you call back, is there another time? No, we have to do it then because we have babysitting problems.

And clearly there's a job that had risen and you're doing workarounds to get the job done. And then somebody comes along and said, wait a minute, what's, what's the job? the fundamental reason why we're doing all of these workarounds and they come up with this concept of open table that solves the problem some startups in the valley they get criticized because they don't think at all in those terms they really don't have a customer idea of you know they've they've got a plan they've got a pitch do you buy that is that what you find basically the way i think about it related to this I think about basically a technology change happens, like something happens in the core technology, architecture change.

So there's TCPIP or there's the microprocessor or there's the smartphone. The strength of the valley is you get 1,000 startups that basically swarm all the different possible implications of that change. And so you get 1,000 mobile apps, you get 1,000 devices that use the microprocessor. You get 1,000 internet startups trying to do every conceivable application of the internet. And then, Clay, to your point, it is something like an 80% mortality rate, right? most of them don't work, and then a small number of them do work and find somehow either deliberately or unconsciously or unconsciously find the job to be done and end up becoming

big, important companies. The well-meaning critics are like, geez, that seems like a lot of waste. Why don't you just think through up front what the problem is, right? With the the solution versus product. Why don't you design a solution rather than a product? Why don't you solve a problem rather than just sort of see what happens? So here's the challenge. The historical track record of innovators, technology innovators predicting the consequences of their innovations is very poor. So is everybody else's predictions. My favorite story on this, which I always go back to and I think about this, is Thomas Edison, who with a tremendous track record of success, you know, invents the phonograph. And he brainstorms ideas on the use case of the phonograph.

And, you know, his preferred idea, idea number one, as a 19th century devout white male Protestant, you know, elder of the community. in the three-piece suit was obviously the purpose of the phonographs is so everybody can have a library of sermons at home. Of course, it turned out the killer app, it turned out the job to be done for the phonograph was music. It turned out, by the way, it was music primarily for kids. By the way it turned out it was genres of music that were primarily new, right? Blues and then later jazz and then later rock and roll. And this came as a complete surprise to Edison. He had no reason to like believe or disbelieve this. It's just he had a worldview of what this, of what the job to be done was that was just completely wrong. And then the phonograph

ended up being a transformative innovation for reasons he didn't expect. And so I would say that the assumption the job to be done theory is that it is actually possible to identify a priori what the job to be done is, as compared to let the technology out and then let the customers figure it out. So I believe that if you're not looking for the job to be done, then you will observe something that happens and impute because of this, then that. If you're very careful about watching what people do, then finding a product that gets the job done is faster. If you understand what causes us to buy it, we can develop products that nail the job better and better and better. And that is really hard for a disruptive company

to co-after because they enter the market thinking that if they make better products than the competitors they will win the game so once you get the job done how do you sustain growth from that how do you build from from that so if you understand the job to be done it then allows you to ask another question about it that is in order to get the job done what are the functional dimensions and the social and the emotional dimensions of that job to be done in almost all even for sophisticated high-tech products, they have emotional, social, and functional jobs, elements of the job.

If you understand that, then you can come to the next level, which is, if that's the job, what are all of the experiences in purchase and use that we need to provide the customer so that they will nail the job perfectly? That becomes the criteria by which the customer, chooses to buy your product versus somebody else's. If we define them well,

not only does it help us do the job better, but it defines the business model. And that's what differentiates our product from the competition. And once we understand the experiences we need to provide, then the last layer is, what do we need to integrate? And how do we need to integrate them so that you can provide these experiences required to get the job done.

And that combination of that package of things makes it very hard for somebody to disrupt you. And it defines what you and you uniquely can do. So it's all going to change the pace with a little different question. You know, people know both of you as people who've been right a lot of times about things. I want to ask each of you if you could recall time when you were wrong on something or maybe even somewhat wrong and you wish you knew then what you know now. Where do I start? How long do we have?

So what I found at least in kind of my line of work, either as an entrepreneur or an investor, what I found useful as a framework to think about mistakes is actually two different kinds of mistakes. Very important to be able to differentiate between them. Mistake of commission and mistake of omission. So mistake of commission, right, is I start a company or I invest in a company and it doesn't work. I tried something and it failed. Mistake of omission is I had an opportunity to start a company or invest in a company and I chose not to and it became a huge success and it then tortures me for the rest of my life.

The nature of the economics of the tech industry is that over time, the mistakes of omission end up being much bigger. Mistake of commission is you can lose 1x. Mistake of omission is you can lose 1,000x because that was the upside. you could have gotten. And so all the mistakes that I care about are all mistakes of omission. And I'll just give one example from the past. I don't know whether I would have had the opportunity to invest or not, but Google just struck me as absurd on its face for two obvious reasons. One is everybody knew search was a low-quality user experience. And two is everybody

knew that search was a cost center and out-of-profit center. Those were two widespread prevailing views. And I, in that case, I didn't do enough thinking from first principles and didn't really work my way through it and it wasn't open enough to the idea. And so that's one that always resonates in my head. So when you have an error like that, how does that change your thinking? Were there times later that you felt, oh, wait a minute, Google, and then you did something different than you would do otherwise? Yeah, so every VC has a bunch of theories on why we make the investments we make. And then I like to do what I, you know, what they call the back test, which is like, okay, you know, if Sergei Brand and Larry Page walked in the room and you applied

your fancy criteria to their, you know, at the time would look like total craziness, like would Would they have gotten funded? And the nature of history, right after the fact, these all looked totally obvious. At the time, these all looked incredibly uncertain and incredibly tenuous for reasons that we could talk about at length in each case. Naturally, as we go through life, we accrete beliefs, beliefs about how the world works, beliefs about causes and effects, beliefs on patterns that we've seen in the past, either successes or failures, either, you know, this actually worked or massive scar tissue

on something that didn't. I try as hard as I can to be as ruthless as possible at shedding the old beliefs, at leaving them behind. are so rarely actually predictive of something new. By the way, it's very difficult to do because it's a fundamental attack on the ego, right? Just think of if somebody does that to you. It's like, no, no, you're wrong on X. Like, your natural reactionist, no, I'm not like, you're wrong.

You're an idiot. And so to do that to yourself, it's like a, it's a very unnatural process. Let me give two. So before I became an academic, I started a company with several MIT professors to make a set of products out of advanced materials. I was the business guy. My partners were the technical brains. And in the early 80s, we raised $60 million, which is a lot of money back then.

And one of the biggest investors was BMW, because you could calculate that if you used silicon nitride instead of steel in making the engine, you could save a big number in efficiency costs because you wouldn't have to cool it. So anyway, they decided that we would start. by using silicon nitride to make piston pins that connect the piston to the connecting rod and they gave us the specs we tested them out the gazoo went over to Munich in triumph and they had this test engine laid out already to put in these different piston pins and we turned it on and it It shaked uncontrollably. And it turns out that many of the other elements in the system

were designed to accommodate the up and down movement of steel. And you couldn't take that out and put in a different material with a different ratio. In order for BMW to use our piston pins, they will actually need to redesign the engine block to reassemble or redistribute the weight. And if they did that, then they'll have to redesign the car. So we came home with our tails between our legs. There's a theory about innovation that we call interdependence and modularity, and it did not exist at the time.

We did something that was impossible to do. And had we had a theory, we could have predicted. What's scary is even BMW's engineers didn't know that there was this interdependency. But that, I think, Mark, that's exactly what you're talking about. And in my language, I spent my life trying to provide more and more theories of causality so that people don't have.

have to win by trial and error. I want to end with a book you wrote called How Will You Measure Your Life, in which you took some of your strategies and applied them in a non-business sense to people's lives and goals. Tell us what you found. One day I was driving to work about six in the morning on Huron Avenue in Cambridge. and I realized out of the blue that God doesn't hire accountants in heaven. And as I tried to understand where did this idea come from,

I realized that what's going on is you and I have finite minds. And because we have finite minds, I can't keep track of all of the invoices that come in and go out, but rather we have to have an accountant to add. up all of the numbers and all of those invoices, and I can then look at the sum of all of the detail. Because of our finite minds, we have to aggregate things. We get a sense of hierarchy in the business world. So people who preside over bigger numbers are more important to mankind than people who preside over smaller numbers. We have this sense that if we can go up this hierarchy and

reside over bigger numbers, my life is worth meaning. But then I realize God actually has an infinite mind. And therefore, he doesn't need to aggregate up above the level of individual people in order to have a perfect understanding of what's going on in the world. And what that means is that when I die and I have my interview about whether they're going to let me in or not, it's not going to say, oh, my gosh, the famous Clayton Christensen of the Harvard Business School. What God's going to do instead is say, Clay, I put you in this situation. Can we just talk about the individual people who you help to become better people? because you used the talents that you have.

And then, lest you forget, we gave you five wonderful children. What did you do to help them become magnificent people? And the reason that God doesn't need to hire accountants is he will only measure my life by the individual people who I helped to become better people. And I was grateful with that. insight because being a teacher gives me opportunities to help individual people. But even more than that, I realize that management is a noble profession because every day a good

manager goes to work and does her best to help everybody in the organization to succeed more productively and rewardingly than they were able to do that before.