#105 Strategy in the Age of Digital Competition, Innovation and Power feat. David Yoffie

Foreign.

This is Unsiloed, brought to you by alumnifm. Connecting People through stories.

Welcome to Unsiloed. This is Greg LeBlanc, and I'm here with David Yaffe. David is. Has been a professor at Harvard Business School for, I guess, 40 years and has authored numerous cases, like 200 different cases, which is, you know, something that Harvard faculty are encouraged to do. I've used probably a dozen of your cases, David, in my strategy class and other classes over the years. Everything from, you know, really old cases, you know about, you know, Walmart and Intel and Apple to more recent cases like the one on Ripple. But in addition to all of those cases, you've written a ton of books. Strategy Rules, Judo Strategy, Competing on Internet Time, Competing on. I forget, the one on digital convergence. Right. And of course, the most recent one that you've co authored is this one called the Business of Platforms Strategy in the Age of Digital Competition, Innovation, and Power. Welcome, David.

Thanks, Greg. Thank you for having me.

So look, you've been writing about technology, you've been writing about digital for a long time. You've served on the boards of numerous companies that are at the heart of this digital explosion over the last couple decades. But, you know, it's only been recently that we have formulated kind of an understanding of what we now call platforms. I mean, you, you were writing about, obviously digital, you're writing about network effects as early as the 1990s. But, you know, we didn't necessarily use this term platform back in the late 90s. And now I think we're at the stage where sometimes I tell, you know, clients or in talks, like, every company has to think like a platform or has to be a platform. And you talk about the difference between product thinking and platform thinking. And presumably this calls into question the kind of value chain model that has kind of rose to prominence at Harvard with Michael Porter. And, and now, you know, we're discouraged from thinking in terms of industries, we're thinking in terms of ecosystems, we're discouraged from thinking about products. Instead, we think about platforms. So, so what is it about platforms that. That make them unique? What is new here? What is different? Why is it that every company needs to think like. Like a platform? And is it something that is unique to the digital age? Because we see all sorts of corporate forums and business models that look a whole heck of a lot like platforms that go back as far as, you know, ancient Rome and the Middle Ages. So what exactly is new here?

Okay, well, it's important to note that it isn't new. This has been around for a long time. I don't go back to the Roman period. But even in the book, we start with the example of the Yellow Pages, which is a classic example of a platform that's been around for a hundred years. In addition, if you think about the first stage of the digital revolution, Microsoft and Apple created what we call innovation platforms in the 70s and early 80s. That became the foundation of most of our current thinking about what a platform really is. What changed is sort of obvious. It's the Internet. The Internet enabled a much broader range of activity to take place where you could connect two sides of the market together. And it accelerated the opportunities to drive revenue. It accelerated the opportunity to build network effects. And so it led to an explosion of platforms that wouldn't have been possible in the absence of something like the Internet. So even though we've seen them for decades, the last 25 years has seen an exponential increase in the number and the variety and the types of activities. Now, the reason we like to think about platforms as opposed to just products today is that if you think about what we call an innovation platform, where you create a technological foundation, you open up a set of application programming interfaces that enable third parties to contribute, you're essentially leveraging many, many third parties to create value around a product or a service that would never have been possible if you tried to do it all internally. And that's why it's created so much interest, is that the explosion of possible value simply wasn't there before. My favorite example I always like to use is when Steve Jobs introduced the iPhone in 2007. He said, There will never be an app store, that Apple would write all apps for the iPhone, and.

We wouldn't be talking about iPhones right now if he'd stuck with that.

That's right. I always like to say when I give this lecture that if he had stuck to that position, where would Apple be today? The answer is it would be BlackBerry, because. And BlackBerry's market share rounded down to zero about, I don't know, in a year and a half ago. So it would be a completely different world. I mean, there are 2.2 million applications, give or take, on the app Store today. Could Apple have ever, in his Wildest Dreams, created 2.2 million applications? The answer is no. It's not. Just not possible. And yet a lot of the real value that consumers find in the iPhone is premised on having those applications available. So it's that kind of exponential improvement in value creation that was not possible in a pre Internet world. That has made platforms so incredibly interesting and potentially powerful and valuable in the last decade or so.

But if we think back to, say, Henry Ford, I mean, Henry Ford thought that he could build a car from the ground up. And he went and acquired, you know, rubber plantations in Brazil and, you know, timber forests and Wisconsin. And, you know, the idea was he was going to be completely and totally vertical, vertically integrated. And I think if he had stuck with that, you know, nobody would be driving Fords either. Right. And so, you know, he understood that you needed to have independent brake manufacturers, need to have independent tire manufacturers and so forth. Is, Is it kind of. Is that the same idea? I mean, did he create an innovation platform similar to what, what, what Apple did? It just that it was not built around, you know, APIs, obviously, but, but it was built around something kind of similar in that there were these kind of standardized interfaces and, you know, what, what your, your, you know, your, your colleague Kim Clark referred to as, you know, back in an earlier era.

Right, yeah, but you're raising an important distinction, and that's the distinction between a company platform and an industry platform. What Ford did is create a company platform with modular interfaces that enabled suppliers to be able to contribute to the creation of a new product or a service. And the auto industry is actually one of the most important foundations for company platforms as opposed to industry platforms, because they do maintain everything as proprietary to Ford or proprietary to General Motors or Toyota. They aren't open to third parties to contribute. You have to have qualified suppliers. There's a whole process that you go through, and that's fundamentally different from an industry platform which have at least some open interfaces, whether they're APIs or they don't have to necessarily be digital. Right. But they have to be open. That enables people outside of your direct control, which was the way Ford was thinking about it, or Toyota. I mean, they're all basically doing the same thing in the auto industry. You need to, if you want to have an industry platform and you want to have the opportunity to create network effects, you have to do a much more open structure than what the auto industry historically did or what many companies who created company platforms built historically.

So I want to go back to this idea of network effects. You know, since you've been teaching at Harvard for 40 years, the courses have evolved, obviously, and at some point, network effects became an integral part of strategy, you know, the way in which you teach strategy. At what point did this happen? I mean, I haven't been teaching strategy as long as you have, but, but I Remember, you know, we, we included it from kind of day one, when I, when I started teaching it. What, when did this become something that every strategist needed to understand? I mean, obviously we would never have had the telephone networks, we never had the railroad networks. We never had any of this without network effects. But it didn't seem to be mainstream in economics. And certainly I think even before we called them network effects, we would talk about increasing returns. I think that was sort of the lingo back in the day. So how did this kind of get mainstreamized or mainstreamified in business education? When did this happen?

Well, a number of scholars started to write about network effects in economics in the 1980s, primarily built around the model of the telephone industry that was the model for thinking about and where it, it moves from, from the academic literature in a fairly narrow business context made by telephones to a more general context. I, I, I would trace it back to Microsoft and, and at least in my own teaching, I did my very first case on Microsoft in 1990, and I had several long interviews with Bill Gates and Bill Gates. There's a quote on the first page of my Microsoft case where Bill Gates introduces the idea of network effects. Number one. I found it amazing that this guy who I thought of in 1990 as the ultimate nerd, had read enough economics that he actually was thoughtful about understanding what a network effect was, why it was important and how it applied to Microsoft. And this was just at the time the Windows 3.0 was coming out. And of course he was thinking of the context of DOS as well in the 1980s. But what he talked about in that interview, and I have this very long quote for him, was that the power of Microsoft was built on network effects. And then he went on to explain what a network effect was and why it was important to Microsoft and how you could then build value of it. Now, when Microsoft becomes the most valuable company in the world in the 1990s, people started to pay attention to these ideas. I also think that the book by Shapiro and Varian in, I think it was 1997, it came out in 98, did in many ways the best job of introducing network effects to a broad audience of MBA students as well as to faculty and to executives in the tech world. So Berkeley was an important creator of the popularity of the concept. It helped popularize it more broadly. And I remember, yeah, go ahead.

I remember when Paul David came out, he came to visit when I was in graduate school, my first year of graduate school, and I was studying history, economic history, and he Came out and presented his paper and that was why I actually bought Microsoft Word. I was like, okay, well I gotta buy the, I got to buy the productivity software that all my colleagues are going to be buying. So I had to kind of think, okay, what are they going to buy? But I didn't, I didn't think to buy Microsoft stock. That, that was the mistake that I made. I drew the wrong inference from, from Paul David's article.

Well, you drew a correct inference, just not the only one I think would be a way to describe it. And, but, but I, my view is, is that it was Microsoft that taught us about the power of network effects. That's when I started to teach it back in the mid-1990s. And again, as more and more got written about in the late 90s and the early 2000s and the explosion of the Internet led to people like Jeff Bezos and then ultimately people like Mark Zuckerberg internalizing the content, the ideas, it certainly produced more interest, more excitement, and as they often say, a lot of misconceptions. I mean, I spend the first four or five days of my course teaching about network effects. And I start and tell the students that I think everything you've been taught is probably wrong. We're going to miss, we're going to, we're going to fix these misconceptions and make sure you understand. Because if you actually look at the history of how people have used the term network effects is that it's led to enormous abuse. They see network effects everywhere. They assume a network effect by itself creates value and that they therefore do lots of crazy things like raise spectacular amounts of money for businesses that actually have no underlying profit bottle that will ever make money. And so, so network effects are critically important and broadly abused.

So I want to dive into, I want to dive into those misconceptions because certainly every venture capitalist I know is looking for, you know, network effects whenever they're interrogating a founder.

Well, and again, this is why in the book the Business of Platforms, we start out by saying that network effects are critical, but not enough. That the power of network effects is the potential for a winner take all outcome. And that's what the venture capitalists are most excited about and most interested in. And frankly, all of our MBA students are similarly hoping for the opportunity to build a winner take all business, whatever that may be. And the power of a network effect is that if it's really powerful, and again, we have to distinguish between same side and cross side network effects. But if the network effects are really powerful. It does lead to very rapid growth and it does lead to at least the potential for individual consumers to want to opt for a single solution rather than multiple solutions. But that by itself doesn't necessarily produce good business. I mean, you know, my favorite example in the book is, you know, Uber is a huge business and they've only lost $30 billion so far and I'm not sure they will ever make money. And the reason is you have to look at the other aspects of a platform that potentially drive value. So you have to understand whether or not you are have switching costs and whether you're locked into the platform or whether you can multi home. If I can engage on multiple platforms at the same time, like Uber and Lyft for example, and switch between them, whether I'm a driver or a passenger, it makes it very hard to raise prices, which is part of the problem that that business has had. Similarly, there does need to be a business model that says that you can earn money from at least one side of a platform as opposed to subsidizing both sides, which is what we see in the Uber and Lyft type of business. And finally, you also have to have some rules about governance, about who can participate and who can't. So it's not as simple as network effects. And that was part of the misconception that a lot of VCs I think have suffered from.

Now, when some people talk about platforms, they usually are limiting it to entities that have these cross side network effects. Right. So if it's. I mean, I think one analogy you used in the book, which I like was same side network effect is like a gay dating app. And a cross side network effect is when you have sort of, you know, a straight dating app. Right. Could you like elaborate on that? I mean, the same side networking app is presumably, right. Where I mean, same side network effect is when you have something like the phone network or a railroad network, right. Where there is some interoperability and essentially different nodes join and those nodes all establish relationships with all the other nodes. But with the cross side network effect, it means that you have at least two very distinct kind of groups of customers that have to be treated differently, perhaps have different pricing arrangements, offering sort of a qualitatively different type of product. So could you talk a bit about that? And when you talk about platforms, are you limiting it to situations where you have the cross site network effect?

No, no, no, no. We always think about both cross site and same side network effects as being important and incredibly important depending upon the business that you're in. So a big piece of Facebook's business for example, is same side network effects. It's, you know, I, Facebook doesn't have a lot of value if my friends don't join. Well, they're the same side of the market. Now there also are cross side network effects because there are also advertisers or third party applications that are built that as a user of Facebook, if you can take advantage of. So but let's take a step back. A same side networks effect is basically where you, as you said, you have interoperability and communication between people who are on the same side of a market. And the easiest way to think of it I think is, is again the telephone example. If you and I are on the telephone, then we're, we're on the same side of the market, we're on one side of that, of that platform and we're just communicating directly with one another. But across the market says that there are two different sides of a, of a market where you have to connect buyers and sellers, passengers and drivers, freelancers and companies. Or again, let me go back to a pre Internet example. Credit card companies are connecting you and me as consumers to small or medium or large sized businesses. They're enabling us to actually transact with someone who we would otherwise have difficulty transacting with. And so you have both of these types of network effects. And the critical issue here is not whether it's same site or cross site, it's the strength of the network effect that drives both value and growth. Whether it's same side or cross site. Right. In the phone business, we know same side is very powerful and drives very high utilization, which is why ATT became at one point the most valuable company in the world. If you think about cross side network effects, that's let's say Amazon's marketplace. I'm bringing buyers and sellers together and enabling them to be able to transact. One of the things that we also argue is that in many cases the most powerful opportunity for winner take all is when there are both same side and cross site network effects because that strengthens the network effects, reduces the likelihood of multi homing being on multiple platforms at the same time. So these can operate either independently or they can interoperate together and potentially drive even more value.

Now when we look at some of these standards that emerged in the 19th century, so whether it's the railroad standard or the, you know, the telephone standard, this was a product in part of governmental effort, right? So it was very difficult to get all the railroads to kind of coordinate on their gauges. It was very difficult to get them even to coordinate on, you know, their schedules. Right. We had like 50 different time zones in America and certainly with AT&T, I mean this was a government supported monopoly. And yet a lot of these standards that we're seeing emerge or these platforms that become de facto standards in our era, they're all private. The Wintel standard was created by entrepreneurs and private companies. Facebook, Google, Amazon, they're all private companies. They're not government supported monopolies. We'll get into whether they should be government regulated monopolies, but it seems like the drive towards standardization on these platforms is being driven by private companies. Of course, the Internet standard is a modern example of a government driven standard. Is this a fundamental change, do you think? The role of the private sector here? Is this because the opportunities are, are, it's easier to kind of build these things out now compared to what we had in, in previous era.

I, I would distinguish between standards and platforms. Standards have been around for forever. You can go back hundreds of years and think about, you know, the standard for a light bulb for example, or the standard for a, an electrical switch, which of course is different in the US versus Europe versus elsewhere. And many of these standards were created de facto rather than largely by the same kinds of strategic questions that we talk about today. You know, the winners became companies that drove very rapid volume, they got faster consumer adoption, and then other people essentially agreed to follow the same standard, even though they were not dictated by law. So standards have been around for a long time and I don't see that as being really any different today. Now some of the, some of the platforms that you were describing, like the telephone industry that was started as a non regulated business, but then quickly became a concern by governments that it was a natural monopoly and natural monopolies needed to be regulated and therefore the government stepped in. I don't remember exactly when the original government regulation of AT&T started, but it was probably 20 or 30 years after the company had, had begun building out networks. So it wasn't even in the early stages. Now today, you know, it's certainly the case that we built a lot of de facto standards, not de jure standards. And, and as you said, the obvious challenge that's being created for governments today is a lot of governments don't like those standards or, or the, what I would call governance, the rules that are being applied. And that is leading to increasing interest in whether there should be regulatory oversight or direct regulation of some Type.

Now one of the things, one of the lessons of network effects is that they're sticky, right? And that they're, once you do get something established, whether it's a standard or a platform built on a standard, they're difficult to dislodge. And I think that one of your messages is that, hold on, not so fast, right? And I remember when I was teaching strategy, I used to emphasize that, hey, the Wintel standard is there, it's, it's going to, you know, it's, it's owned, it's not going anywhere. And then of course, we had to write its obituaries not too, not too long after it seemed to be, you know, impregnable. So, so what is, is, is that one of the, you mentioned hubris as one of the potential pitfalls of a misunderstanding of platforms. What accounts for the dislodging of these standards? How can one anticipate the dislodging of what appears to be a impregnable platform?

Well, by the way, I would say that Wintel is still very much alive and well in the sense that it does still own 90 plus percent of the PC market worldwide, which is still a, you know, trillion dollar business roughly. So I wouldn't, I wouldn't throw it out yet.

We can go back, let's get, we can go back to Lotus 1, 2, 3, we can go back to.

No, no, but, but, but let me come back to your key point, which is it is absolutely the case that network effects do not guarantee success forever. There are many things that can cause a business with very strong network effects and would appear to be lock in, to be dislodged. Um, I mean, the most obvious thing that they can dislodge, it is first a technological discontinuity of some sort. The threat of, of Wintel is, is actually more about new technologies becoming more important in the cloud and in mobile that, that make it less relevant and therefore not quite as potent. Second, there can be internal or external shocks of various sorts. An internal shock is, you know, we describe in the book as hubris, where companies get, you know, just to say it politely, fat, dumb and happy, they don't continue to innovate and they open up Windows for, for new players to actually take the market. You know, the obvious example that we use in the book is everyone 15 years ago was certain that Microsoft had won the browser wars. It was over and done. They had 90% share and there was very obvious network effects that one could identify. And then they just stopped paying attention to it. They Stopped innovating, they enabled the oppo. They created an opportunity for cyber risk. So there was massive amounts of malware being deployed through an Internet Explorer browser, and that led lots of consumers to defect. And ultimately Google was the one who figured out that there was a real opportunity there. It's interesting that I read an interview with Eric Schmidt where he said when he first heard the idea that his guys wanted to build a browser, he thought it was the dumbest thing he had ever heard. Until he realized that Microsoft, who was screwing up so badly, created an enormous opportunity for Google to capture a lot of that value. And so it's certainly possible that you can screw up in a variety of ways. The third thing that we talk about in the book, and I certainly teach in my classes, is that many people assume that a market that's growing very fast is, has tipped to the player that has 90% of the market. And a great example I'd love to give of that. Well, there's two examples. One would be Netscape had 90% of the browser market in 1995, 96, 97, and obviously they disappeared. Another great example is eBay had 80, 90% of the market for auctions in China. And of course, today, ebay is tiny, and Alibaba has won that marketplace. And what people often miss is that in a very rapidly growing market, you can potentially defeat a network effect by going for new users that have yet to take advantage of the existing network effect. So why did Alibaba win? It wasn't because they took share away from ebay. It's that they found new consumers who are adopting their product or their service first. For the first time, because of very aggressive marketing, they were giving it away for free. Ebay was charging, and they quickly established their own network effect separate from the. The existing users of ebay. The same thing happened with, with Microsoft and Netscape, which is Microsoft figured out that you don't want to take away Netscape users. What you want to do is the market was growing exponentially. If they go to new users who had yet to adopt a browser, then they could build their own network effect. And if they got big enough, they would be able to eventually take away a significant percentage of the market share. So you can see both these external shocks as well as these internal shocks, as well as technological change that can potentially undermine what appears to be a winner take hold position.

Right? But this, this highlights the, the importance again of, of some kind of traditional strategy concepts like, like switching costs. Right? So, so I guess, you know, the, the story you're telling is one where, you know, even if the switching costs are very high, if the kind of existing participants in the platform are not really needed for the new entrants, they're not going to be interacting with them or they'll be interacting completely with other new entrants then, then it doesn't really matter what this, what the switching costs are. But, but to, you know, switching costs are still important to prevent, you know, multi homing. One of the reason things I've always been curious about is, you know, why Uber doesn't make a greater effort to limit the ability of drivers to work on both, both platforms. And I know Uber had some kind of nefarious spyware that they put into their apps that would alert them to, you know, whenever a Uber driver started using Lift. But, but they, but they haven't kind of devised ways of, of deterring or, or discouraging this, this multi homing and I know kind of loyalty, loyalty programs are one way that you can do it. But, but, but I think that's going to be, they're going to have to figure that problem out if they're going to ultimately be a winner take all.

Except Uber has got a problem, which is Uber is, is dedicated to the idea that they are a platform and not an employer of their drivers. And therefore if they restrict what the drivers are able to do, it increases the probability that regulatory authorities would see them as an employer and say that they then have to provide benefits and provide minimum wage and provide a variety of other benefits that are simply uneconomic for Uber to deliver. And so that's the fine line that Uber's walking, which is if they go the way you're going, which you say is that's what a traditional strategic approach would say, then sure, they would increase the probability of being a winner take all. They would make it hard to multi home for both drivers as well as less so for passengers. But really drivers, and yet if they do that, they run the risk that they suddenly are no longer a platform but an employer of, of people. And that obviously drives up their costs and makes them a taxi company and by the way also makes them a regulated entity because most taxis are regulated. So they're trying to walk a very fine line. And they've obviously struggled at least in terms of their profitability. They haven't struggled in terms of their market valuation because they've been rewarded for doing what they're doing. And as long as capital is available, their incentives to go the direction you're describing are still not that intense.

Now you and your co authors make the distinction between kind of a transaction platform and an innovation platform. And you mentioned these hybrid platforms which kind of combine elements of both. Could you articulate the difference between the two for us?

Sure. I mean both types of platforms have been around for a long time. So Yellow Pages was a transaction platform, right? I mean transaction platforms are just often called marketplaces. And we actually like the term transaction platform better because I don't really think of Yellow Pages, for example, as a, as a marketplace, even though again, you could, you could define it in that way. But, but basically what a transaction platform does is enables people to interact or transact with one another, bringing multiple sides of a market together.

And think about a transaction platform is it kind of blurs the boundary between firm and market. Right. The traditional distinction that we use in kind of organizational economics. Right. So if you are, I don't know, I mean pre ebay, you're managing a swap meet, right. You know, you have a bazaar like the one in Istanbul, right. Where people have stands in there. I mean that is the bazaar itself, when it rents, stands out. Is that a firm, is it a market? Like what exactly is it? Or New York Stock Exchange, kind of same thing. It's a transaction platform, right, exactly.

So I completely agree with that distinction and I think that that's a good way to think about it. Innovation platforms are what we just talked about when we talked about with Microsoft, or again many others that we can identify where what you're building is a technological foundation and then you're enabling people to innovate around that foundation so the consumers are able to connect directly to that innovation. And you could think about Windows, I mean take the obvious ones, or Windows. IOS and Android would be the most familiar to everyone. And a hybrid says what an individual company is trying to do today is increasingly doing both innovation and transaction within the same company. Now that almost never happened 15, 20 years ago. And again, this is related to digital technology and the Internet. That's what's changing it. But you know, if, go back in your early days of the PC, if you bought a third party application like Word for example, or let's say WordPerfect, which would be now didn't deliver by Microsoft. Where did you buy it? You didn't buy it at Microsoft. You went to a store and you bought it. So it was a classic innovation platform. Microsoft opened an API, WordPerfect then enabled the application to run on work on, I'm sorry, either DOS or then Windows. And then people went to A marketplace or a store to buy it. Today, almost everyone who is thinking about platforms is doing both. That's why we argue in the book the future is about hybrids. So if I'm doing a transaction platform today, most likely I've also opened up APIs and enabled third parties to innovate on my platform. If I'm doing a innovation platform like let's say Android, I also create the Play Store and create a distribution mechanism to connect buyers and sellers directly. And even if I'm Uber, I mean, Uber's opened up APIs and enabled third parties to create applications that, that work with Uber. So if you think about almost any major platform today, they're all trying to do both, even in the B2B space. So you know, Slack, for example, has about 3,000 applications. They've opened an API, they have 3,000 applications written for Slack and they have their own transaction platform on the Slack site that enables people to buy directly through the Slack transaction platform. So whether we're talking about B2C or B2B or B2B 2C, you know, all of them are basically moving in the directions of hybrids today.

Now, I think when you open the book, you talk about this decision that Bill Gates made in the very earliest days when he essentially allowed IBM to license DOS for free, but he reserved the right to license it to other OEMs. And you say that this was very kind of foresightful, right? You know, he really, he really under understood that. But in, in those cases, as with say the Nintendo case, which you also talk about, which I actually have taught many years, it, the pricing is, is key, right? Understanding, you know, which side of the market, which customers are going to be the ones that are paying, which ones are going to get the product for free, which ones might have to actually get, get paid. Right. So you know, one example I use in class is the nightclub where, you know, you might charge the men and, and have free admission for the, for the women, maybe even give free drinks to the women, right. And understanding that these, these different groups require kind of different, different pricing and if you get that wrong, then, then the whole, whole model model crumbles. Is that one of the common mistakes is, is the, or is that, you know, that's, that's a variation on the go to market, right? So go to market is super important. Have to solve the chicken and egg problem. Do you find that when people wave the magic platform wand, they haven't thought through very carefully what exactly the go to market is, how they're going to get that critical mass, at least on one side of the market, that's going to attract the other side.

The single biggest problem for any platform is which, which side of the market is the profit making side and which side is the loss making side and how do you make that equation work? And it seems like it should be obvious, but it turns out to be incredibly difficult. I also use the example of you're creating a dating site and do you charge men or women? And I've given this lecture all over the world and generally you get the same message no matter where you are in the world. And of course the question always is, well, okay, well, why do you charge the men and not the women? And it's, you know, a general perception, whether it's true or not, I can't, I can't say it's that men, men are more desperate, they care more and therefore their, their willingness to pay is going to be higher. And, and that therefore if I, if I can attract the women for free, I will get a paying side from the.

Well, actually, actually in, in the case of thoroughbred breeding, it's the opposite.

Okay, well let. The other example that I teach in my course is, is Upwork. Are you familiar with Upwork?

Yeah, I saw you had Stefan Castriel give a blurb on the, on the back of the, on the back of the book. Right.

And of course he's no longer there. But, but you know, the, the reason that's such a great example is they've been around for 20 years and they're still about a break even company and they made the decision early on to charge freelancers and not charge companies. Now why do they do that is because it's easy to get freelancers, freelancers are more desperate and it's hard to get companies. But it turns out that that actually is highly problematic because the freelancer actually has very strong incentives to circumvent the platform and to negotiate directly with the company once the connection made.

And so even though that's what happened with homejoy. Right, so homejoy, I think fell apart because people were just disintermediating them after the initial intro.

It's a constant problem. So, you know, again, what does the platform do? It connects two independent parties. Well, now the question is why do I need the platform in the middle once I've been connected? And what is interesting about Upwork is on the chat boards on Upwork they say, why are we doing this on this platform? Let's go off and do it independently. So, you know, upwards even knows it's happening, they can see it happening and they try to fix it. But of course part of the problem is you have to sue your customer in order to actually fix the problem. And they've been struggling with this for, for two decades. It's, it's a non trivial problem. And again, there are a variety of mechanisms that platform theory has begun to develop on how you can solve this. But, but it is one of the, the critical issues is trying to figure out which side do you subsidize and which side do you get to pay. And you want to avoid in most circumstances what Uber has done, which is subsidizing both sides, which is why they have a hard time making money. Right. They discount for the passengers and then they provide these enormous incentives, bonuses for drivers to come onto the platform and that then produces the $30 billion in losses. Right.

But I think two strategies you mentioned, one is where you have sort of a freemium, although I don't think you referred to it specifically. You know, the idea that you get people in and then you offer them enhancements and then the other are compliments. Right? So if we think about Taobao, which was initially, you know, initial listings are free, you can pay for kind of upbringing and you can pay, you know, obviously Alibaba makes money off the payment services, right? So, so really understanding kind of complements. And so I think this is when I think about venture capitalists now when they, they're, they're funding these startups, they're like, look, just, just get user traction, make something that's indispensable, make something that people want, and then we'll kind of figure out how to monetize it later. We don't know what the compliments will be, but we'll figure out something that seems to be the strategy.

And that's absolutely one way to solve it is with compliments. If you can create a set of compliments that create the switch of costs that we talked about earlier, then you increase the probability that there'll be an ongoing willingness to pay and you won't circumvent the platform. Some of the obvious complements are payment systems, but the problem is payment systems have become commodities today. And again, those are getting easier and easier to circumvent with an independent payment system without having to use the platform. So you still have to find creative ways to build those switching costs to reduce the likelihood of multi homing and to reduce the likelihood of circumvention.

Now you mentioned that some of the four pitfalls you talk about Kind of mispricing. You talk about mistrust, right? You talk about hubris and talk about kind of mistiming. You know, do you think that, are these things that you can evaluate at the startup level? Right. So if you're, if you're kind of someone's thinking about doing series A, can you really kind of, are these, are there a series of systematic questions that you can, you can ask of the startup to try and make sure that they're on the right track?

You know, I think so. Again, it's obviously industry dependent and context dependent on the competitive environment in that space. But, you know, let's just, you know, take the example of mispricing. You know, we use the example in the book of Sidecar going up against Uber and you know, the CEO was financially disciplined because he wanted to make sure that everybody was paying. When Uber said, no, no, no, that's not the point. The point is to build the network effect first and then you worry about the willingness to pay and as long as the capital is available, that obviously proved to be a very good strategy. And the pricing that SignCar was, was trying to use to, to try and get everyone to pay their fair share missed the logic of the network effect and missed the logic of building the kind of rapid growth and adoption that's required to build a successful platform. You know, similarly, the idea of mistiming is a question of, I mean, this is the hardest question. How do I know went up too early or went up too late? And I always think that the too early problem is when firms try to build a product or service based on rocket science where the only way they're going to win is by accomplishing some spectacular breakthrough in innovation and science that may happen. It's not impossible, but it has a very low probability of occurring. And that's often what leads, at least in the tech world, to being too early. It's that desire for rocket science. And when we wrote the book on Netscape in the 1990s, that was one of our criticisms of Netscape is that Marc Andreessen had some wonderful ideas, but they were all rocket science back in the 1990s. They actually have emerged, but it took 20 years. And so a lot of his ideas have actually met with reality and come to fruition. But it took 20 years and, and obviously Netscape didn't survive that. Being too late is a little bit easier because the too late question is how strong has network effect been built by our competitors? And is it possible to be able to go to the new customers who haven't yet adopted the particular business or product or service and potentially build an independent network effect. And that is a case where data actually is available. You can figure out what is the total available market size, what is the market share of the existing players, how rapidly are they growing, what point would you need to have an intercept in order to be able to potentially mitigate the existing network effects from the existing players? So those are a little bit easier to identify. The rocket science one requires people with deep technical expertise to be able to make judgments on whether or not people are stretching too far in terms of what they hope to accomplish.

Now you have a whole chapter on can old dogs learn new tricks? And I certainly spend a lot of time with old dogs and they're often trying to figure out what to do. And it's easy enough to say, well, you know what, you got to become a platform company, right? Or at least you got to somehow become part of some ecosystem. And it's a whole lot easier said than done. And there are examples, and I think you mentioned in the book, and the book's a couple years old now. And so one of the examples was Walmart acquiring Jet. Another example is GE and their development of Predix. And I'm not sure either one of those has really panned out as expected. What are some of the challenges in trying to rebuild a business into something that's more platform like? I mean, can you, is it simply just about having something so Accor Hotels, right, is another example that's tried to do this. You know, is it, is it possible? I know we want to hold out hope for it, but, but realistically, what are some of the things that are, are there any, are there any easy, low hanging fruit opportunities for, for large companies to be more, more open to allow kind of third to work within their, within their new aspiring ecosystem?

Well, remember in that chapter of the book we talk about three different strategies which we call build, buy or belong. So first, come back to the belong piece of it. For many companies, particularly traditional companies, the answer is not to build a platform and not to buy a platform, because they will never succeed. And they're either culturally not suited to understand how it works and they simply don't have the resources or capabilities to build or buy their own platform. But they can belong to platforms. You do have literally hundreds of thousands of companies who belong to the Amazon platform and for many of them they actually do quite well. It's not that you necessarily have to be independent, but Toys R Us didn't.

Fare so well, on the Amazon platform.

We talk about that, right? We talk about the way they contracted. They made some big mistakes on how you belong. So the details of how to belong is as important as the whole strategy of belonging. So you can be subject to holdup. And that is what happened to Toys R Us with, with Amazon. So again, you've got to understand what belong means, how far you can take it. We talk about some interesting examples in the book about companies that have in fact figured out ways to avoid Amazon, essentially capturing everything that they do. I would actually argue that the Jet.com acquisition by Walmart, maybe it hasn't delivered what, what Walmart had hoped, but Walmart is the number two E commerce player in the United States today, which wouldn't have been possible without Jet.com they have over $60 billion of e commerce revenue and I think if you look at their, their growth rates, it has averaged 30 to 40% every, every quarter over the last four years since they've owned Jet.com and the key thing with Jet was Jet brought the people the talent. It was really an aqua hire. It wasn't buying a platform. I mean Jet was, we always thought Jet was a terrible platform, but the people who were running Jet understood how platforms operated. They came from Amazon, they knew, they knew how E commerce worked. And Amazon made the right decision. I'm not sorry, Amazon, Amazon, Walmart made the right decision of giving all of their E Commerce to the Jet.com team. And that's why it's grown so quickly now. It still is very small compared to Amazon, compared to Amazon. But again, we have a wonderful case that we teach on Amazon versus Walmart. And the question is again, if you're Amazon, if you're Walmart, what do you do? Do you go head to head with Amazon? Do you stick with brick and mortar? You try and do a hybrid strategy. And of course, at least what I argue is in the short run you have to do a hybrid. You've got to find ways to leverage brick and mortar with E commerce because that's where most of your assets are and your capabilities are. But long run, if you don't go head to head, Amazon's going to be the dominant retailer in the world and you're going to be playing second fit. And so it's critical that Walmart figures out how to actually run a platform and they've gotten a lot better. You know, I come, I use the example of before they bought Jet.com, it took Walmart six weeks to qualify a vendor on the platform, six weeks. How long did it take? Amazon? 24 hours. Now why was that? Because Walmart thought of the business of a platform being like before did. It was about the supply chain and it would take Walmart six weeks to qualify a supplier. And of course the, the, the other side of the platform is not a supplier. That and, and their whole thinking was wrong. And the folks at Jet figured this out and they solved the problem. And, and so they were, and they, they enabled Walmart to get into the platform space. Now they should have done it 15 years earlier, but you know, the fact that there's $60 billion in revenue today compared to, you know, I don't know, 3 or 4 billion before they bought Jet is a pretty, pretty reasonable trajectory for a company in five years.

Now of course, this issue of openness raises its own problems, right? So you know, if you have a passive platform that is simply open to anyone, to join, to conduct transactions and so forth, and you have little in the way of curation, this obviously raises a number of problems. And also this idea that once you achieve a certain market power, you are indispensable and you're behaving more or less like a public entity or public utility, which then creates some additional need for curation. How can these platforms think about the trade offs between openness and the need for curation? How can they regulate themselves? New York Stock Exchange is a great example of a platform that was self regulating for most of its history. And Even now the SEC enforces a lot of its rules through these SROs. And so how can if, particularly if you think about Facebook, I mean Facebook is in the news quite a bit for this difficulty they're having with this trade off. Can they escape regulation? Can they escape the new Brandeisian antitrust folks through some policies that are carefully selected that will maintain the trust of the people within the platform without necessarily alienating the people who are going to be excluded by this curation.

You're getting at one of the most critical problems that platforms have today. We talk about this in, in, I think it's chapter six of the book where we talk about platforms is a double edged sword that while they create the possibility for spectacular growth and spectacular value creation, they also have this problem of potentially enabling very bad actors, enabling bad things to happen on their platform. And Facebook is the poster child for that. Now the way we think about it is when you're building a platform in the early stages, you can't worry about curation. What you're really focused on is building network effects and driving growth. But as platforms get bigger and bigger, then you have to start worrying about the clarity of the platform, the cleanliness of the platform. And increasingly, my view is that as volumes and revenues increase with the size of the platform, you increasingly have to try and look at curation and find ways to keep the platform clean. And that's been Facebook's problem, which is it was philosophically not what Mark Zuckerberg believed in. He believed in letting a hundred flowers bloom, and he had this very naive theory that the good would basically force out the bad. And of course, it appears that the opposite has happened, which is that the bad forced out the good. So it was a very naive view. And by the way, it's not just Facebook. You're seeing today the same problems at Amazon, where there's a growing problem with counterfeit goods, with illegal goods from the platform. You do you have lots of electronics that don't work?

No.

You have 500 million SKUs. You have a problem because there's going to be a lot of SKUs that are not going to work. And now the good news is, when you get to be of a certain scale and size, you have the resources to fix it. I mean, Zuckerberg has even said this. I don't know if he said it to my students or he said it publicly. But, you know, Facebook has the resources to curate, whereas a lot of the smaller platforms they compete with just don't. And so, you know, if. If they wanted to. To curate it again, there are many ways you can curate. You can curate through artificial intelligence, through machine learning. There's a variety of problems with how far that takes you if you wanted to curate with human curators. You know, my rough estimate is to really curate the platform, the Facebook platform, you probably need somewhere between 50 and 100,000 curators. And so that's somewhere in the neighborhood of a $5 billion hit to the P and L. But guess what? Facebook actually has the resources, should it decide to do it, that they could actually curate the platform, even with human curators, if it was necessary. Whereas, let's say Snap couldn't do it, Twitter couldn't do it. Many of their competitors simply wouldn't have the resources to make that investment and make it occur. So, increasingly, this is going to be a subject of firms either have to do it themselves, they have to either find a way to be socially responsible in what they're doing, or they are going to get regulated in ways that might be much less good for their business.

Mm.

Michael Kusmanu, Anna Bagauer. And I have just finished an article which is going to be published in a refereed journal on this whole idea of self curation. And the argument is, is that if you are a forward thinking platform today where with significant market presence and market power that either you should be curating, you should be making the investment to clean your platform. And if you don't, the consequences are much more likely to be detrimental to the long term health of your business.

Yeah, I mean even setting aside regulation, I mean if you fail to curate then people will start looking elsewhere for, you know, people will start multi homing and they'll look for the platform that is better curated.

Let me make one point of that which is critical element of every platform is building trust. And the, and the problem again doesn't matter whether you're Facebook or Amazon or Uber, it doesn't matter if you lose trust. What ultimately remember what a platform is doing is connecting two sides of the market that don't know each other. If you lose trust, you will actually destroy the underlying premise of the platform itself. And by the way, it may not only be for you, it may be for everybody in that segment of the industry. So we think of trust as a common resource that can be destroyed if you enable or don't prevent bad actors from engaging in bad behavior on the platform.

Well, this takes us to education. I mean we can wrap up with a discussion of education because you've been in the education business for a long time, as I have, and we kind of use some of these concepts of the platform business model in what we do. We certainly, you know, Harvard Business School Press is using cases that are written by people at other universities and published by other universities. And then of course all the other universities use Harvard Business School school cases. We, you know, we don't necessarily make all of our own products. But still, I think the typical university, typical business school is very much of a product type company. What would it look like for say Harvard Business School to be more of a platform? What would that look like? And when we think about post Covid and we think about how new forms of educational distribution have been demonstrated to be somewhat effective, can you see the scaling of a place like Harvard Business School to a level where it becomes more of a platform and less of a, of a, of a kind of product oriented entity?

Well, let's distinguish between education and Harvard Business School. We may not be totally aligned on what the broader educational world should do versus what Harvard Business School should do.

Well, if we think about, I mean if we think about A business school, you've got, I mean, you've, you've got your recruiters, right? You know, you've got, you've got your, your students, you've got your, your alums, you've got your, your faculty. And, and in a way, they're, they're all different customer, customer groups, right? Or different, different groups that are being brought together. I always tell my students that they're not the customers, they're the product. Right. And so, so in that sense, you.

Know, they've always been.

They'Ve always been a platform in that sense. But, but in terms of, you know, the, the raw materials that go in it, it's still more product oriented.

Well, in the beginning of our conversation, we talked about the difference between a company platform and an industry platform.

Yeah.

What most schools are today are company platforms. In other words, we, we are building our own capabilities and we are then trying to leverage it through the various parties that you just described. The exception at Harvard is the case distribution system. And by the way, that wasn't always true. And it used to be originally only HBS faculty and then publishing, which is by the way, quite independent from the running of the day to day educational mission of the Harvard Business School. It's off literally a mile away from the campus. They realized that that was a potentially losing strategy, that it was important to, to get other educational, educational institutions to contribute to the platform. I mean, if you think of Coursera is, is a platform, right. It is an educational platform that tries to attract faculty from every school and then connect them with students who are interested. But today, MBA programs, I don't, you know, with few exceptions, let's say that the top MBA programs are all company platforms with very limited industry platform ambitions. By the way, at Harvard Business School we have been discussing the idea of trying to be a more open platform and enable the broader distribution of content to tens of thousands or millions of people rather than our existing audience, which are 10,000 executive participants a year and 2,000 MBAs. So there's a lot of debate and discussion about that. Part of the question is, are we as an institution or as Berkeley as an institution, are we trying to serve an elite group of students and executives or are we trying to serve the whole world? And that's a debate. And that debate is an important one because if we want to serve the whole world, then we want to be an industry platform which attracts the best talent from wherever they happen to be. We want to curate it, make sure it's high quality and can potentially make it available to everybody in the world at a very modest price. If you're a premium educational institution like Harvard or Berkeley, you know, you may be trying to appeal to a segment of the audience rather than the whole audience. In other words, we want to be more like Apple than we want to be like Android. We want to, you know, apply. We want to appeal to the people that have a high willingness to pay, who are looking for higher quality, not necessarily to everyone in the world. And I would describe that as one of the central debates that will happen at every major business school in the next year or two as we come out of COVID trying to figure out what is the right strategy in a post Covid world. Is it to be that industry platform, or are we trying to serve a more segmented audience that has a higher willingness to pay, higher capabilities in terms of their ability to understand the concepts and questions that we want to answer, or do we want to really serve the whole world?

Right. And I definitely look forward to seeing how these conversations play out. It's certainly fascinating. Education is one of those industries that, you know, hasn't changed a whole lot in the last couple centuries, but I think is changing very, very rapidly now.

Yeah.

David, thank you so much for joining me. Business of platforms, just kind of the latest version. And I don't know when it comes to cases, like people in business school read cases, but no one outside of business school reads cases for fun and pleasure. But, but I encourage all of you who are not business school students or have never been through business school, hey, just, just go to Harvard Business School website, download some cases, read them, read them for fun and discuss them with your friends. That's some, some fantastic writing that David has done and his colleagues. Thank you so much, David.

Thanks, Greg. It was fun. This is Unsilo brought to you by alumni FM connecting people through stories.