 I want to go with basically over three ideas. One of them is that I do think that we're seeing a new form of business model. I think that I'm going to make the claim that platforms beat products in almost every instance where one goes up against one another. And I think the platform business model will be more successful in the long run. The second thing I want to do is to just give you evidence to show you that this is the case. So what might be the data suggesting that platforms are taking place in the economy? And the third thing I want to do is to give you some of the implications. What happens if this is the transformation? What are we witnessing within changes in the economy as a whole? So let's get started. These are the companies that are most famous in the world. This is a survey done by Interbrand. And you can actually see this is the top 30 or so. And if you look at them this way, these are the fastest growing. Again, these are platform companies, Apple, Google, and Amazon. If you simply look at those companies that have ecosystems associated with them, more than a third of the top 30 are, in fact, platform companies. Microsoft has an ecosystem. Intel has an ecosystem. American Express eBay has collections of external buyers and sellers. Again, they have ecosystems participated with them. This is the fame. Let me give you another data point. This is presence on the web. Now I focused just on the United States area. And we can see Google, Facebook, Yahoo, YouTube is them. If you look at any other part of the economy, though, so if you go to Asia, you'll see much the same thing. It will be Baidu. It will be Alibaba. It will be Taobao. Again, platform companies in terms of presence within the digital economy. So that's fame. One's technology. Let's take a look at actual economic value. If you sort the top firms by market capitalization, we'll find that three of the top firms by market cap are now platform companies. We have Apple, Microsoft, and Google, now among the likes of Exxon, Mobile, and Berkshire Hathaway. And what's interesting is that platform companies are now knocking off the traditional giants, finance, energy, and banking. It's quite interesting what's happened. And in fact, this has continued since the turn of the century. If you sort the number of companies by market cap over the last since the turn of the century, the trend has been continuing for quite some time. So whether we look at fame, internet presence, or market capitalization, platform companies are becoming a dominant force. The product business model doesn't compete. We've all seen the Blackberry as a wonderful example that anyone still carry a Blackberry or willing to admit they still carry a Blackberry. Oh, excellent. I'm delighted over on the right. What's interesting about this particular plot is the remarkable change in market share for Blackberry. If you look at 2009, Blackberry had almost one half the market. Four years later, 2013, it's 2%. That's hard to do. If you're a manager, I'm not sure you want to be in charge of a company doing that. There's something I want to draw your attention to, though, which is also the shape of this curve. Notice the convexity. This is actually going to imply a certain amount of feedback that we can pay attention to and try to use this or explain what's happening as one of the phenomena. Here's another example. Apple is so famous these days, we often forget what happened to Apple in the 1980s and 90s. This is two lines. One, the blue is Microsoft. The red is Apple. And here, Apple's almost flat lined. It never got more than a couple percentage market share, down to almost 0% here. And what's remarkable is Steve Jobs got it wrong at that point in time. They had the world's most beautiful, tightly integrated product, and they never opened it. Microsoft, in contrast, opened its ecosystem, got more than six times the number of developers by the time of the antitrust trial. And this blue line, you can see the feedback as they grew their ecosystem and grew it and grew it and grew it over time. Again, we're seeing some positive feedback with the developer ecosystem. In fact, if we look at the flat line for Apple, Michael Dell was offering his investment advice, which was shut it down and give the money back to the shareholders on Apple. Things have changed a little bit in the futures of that company relative to Dell at the time. Indeed, Business Week and others were forecasting the demise and disappearance of Apple back a while ago. And then things changed as they managed to build an ecosystem. We tend to focus on technology companies ignoring some of the other non-technology companies as perhaps platforms. One I'd like to highlight that's again in this top 30 is Nike Corporation. Nike did something quite interesting. It's a shoe company. It's a sporting goods company. They, too, built a platform. And I'm not talking about platform shoes. I'm talking about a business model. And they did it by adding data. We're talking about the initial and the digital economy. That's what Nike did. They added this sensor data to the shoes. They connected it to apps. And then what they do is they connect it to a community. They connect you to an organization. You can benchmark yourself against friends. I can actually compete with my son on how many steps I take in a day. You can go out and build jogging groups or archery groups or soccer teams and actually connect the community. The community starts to add enormous value. And if you watch the stock price of Nike, it is actually going up fairly substantially as this has been the case. Although at present, they're moving out of the fuel band. They're now trying to become the operating system of sports. It's an interesting transition in the business model. Let me go one step further. My favorite example of a company that's moved in the platform direction is McCormick Spice. Now, I posed a challenge to you. How in the hell are you going to build a platform around salt and pepper? You probably have someone on your table. What's the platform business model around that? Interesting challenge. McCormick Spice did it in an interesting way by adding information, and in effect, trying to become the Netflix of recommender systems. What they'll do is to allow you to enter your profile so you can enter whether or not you like minty, malted, citrus, floral. They can then use this to recommend recipes for you, of course, using McCormick Spice. And they can connect you then to others interested in the same things. They can match you to the best recipes. Users can download these combinations, modify the recipes, and re-upload them. I'll give you a further example. It's quite interesting. Here is a geometry of flavor. So they make veggie risotto, and suppose you like garlic. You could expand the proportion of garlic over here. Over here in the mango blueberry cobbler, oh, you think it's too sweet. Well, you can cut that down and make it more like your own recipe, but then you can re-upload it. The beauty is McCormick now gets to learn your preferences in a better way. They can also start to track trends and see whether or not cilantro is emerging, or kale is becoming a new ingredient. And of course, it's becoming very popular for their own content. This has become so successful that it's connecting McCormick not just to the consumers, but it's also connecting them to restaurants, to helping them design recipes, and it's also connecting them to the grocery stores which are becoming the channel. The grocery stores want access to this to know what it is that customers want. It's a remarkable development in creating these recommender systems around information building the communities on them. What I'm going to argue is that to be a platform, you need some embeddable feature or function, some building blocks upon which third parties can build. We're going to use the ecosystem to have others build upon your own products and business models. And by appropriating your goods into their goods, they then become conduits and distribution channels for what value you create. So McCormick's spices and ingredients are being appropriated into others. iTunes helps get music onto or data off of the iPod. Products have features, platforms have communities. This is a fundamental difference. You're not just competing on the features of an individual product, but how big is your ecosystem and who else is adding value to your ecosystem? It's a big distinction. Products have features, platforms have communities. So here's an interesting question for you. I want to get to some of the mechanisms of what's happening. How does the column on the left relate to the column on the right? Well, Uber drivers and Uber riders, or Airbnb rooms, Airbnb renters, Android developers, Android users, there's a feedback. One side attracts more of the other. If you can get one side on board, the other side tends to follow and you get some reciprocity. Go back to those convex curves that we highlighted a little bit earlier. Here's an example, when Android launched, they came out a year or two behind the Apple operating system. So to get them started, they offered five and a half million dollars in prizes in order to get developers onto the ecosystem. Well, what did this do? Well, the prize money brought developers into the ecosystem. Developers then brought users into the ecosystem, which brought more developers into the ecosystem, which brought more users into the ecosystem. It created this feedback. Google's been fantastic at building these ecosystems. Google has the largest API on the web. In fact, the embeddable functions for mapping are, Google's are the most distributed of any on the web. Google has these ecosystems for search, advertising, mobile, geo-location. It's remarkable how much they've managed to get others to appropriate the value and distribute it on their behalf and also add incremental value on top of the platform ecosystem. So what does this do? This is gonna help reinforce the claim that platforms beat products every time. Let me see if I can go back to an example. Let's go back to Steve Jobs in the 1980s and 90s and replay the clock in the new ecosystem. When the iPod was first introduced, this was pre-itunes, there's no iTunes at this point. So classic linear supply chain. Users are gathering their own MP3 files or they're stealing them from P2P file sharing systems. And there are no network effects in this model. There's no reciprocity among the ecosystem members. It's product-first thinking. It was the most tightly integrated, beautiful hard drive on the marketplace. Fantastic $400 device. After Apple introduced iTunes, some interesting things happened. They introduced iTunes and then Apple moved to absorb this retail function. In some sense, this was the starting death knell of Tower Records. They absorbed the flow of information and content on this ecosystem, but with the absorption of the information flow, followed the absorption of the monetary flow. This remade the channel in from a linear relationship more to a triangular relationship where they're doing matching on top of the platform. Users are matched to content, content matched to users in here. You've removed one step in efficiency and Apple now owns this choke point. For every app that goes through, they can tax that 30% for the value they don't have to create. It's a remarkable innovation in the business model that they have before, plus they have network effects. With a large user base that got for this fantastic device, lots of content providers want to go put their music on there or their video on there or their apps on there. And the more content is available, the more people want that. Again, creating these feedback mechanisms that cause the ecosystem to take off. Compare that to other products. Compare it to the Nokia Lumia, the Sony Personal PlayStation or the Microsoft Zoom, another MP3 player at the time. My favorite example is actually Sony Personal PlayStation. It was a better gaming device than the Apple iPhone. In 2007, when the Apple iPhone was introduced, it had minimal features and almost no games on it, Sony was introduced that same year, in September, only two months later, and Sony stock price went up. It was absolutely a better product. After Apple introduced iTunes and got the ecosystem going, what happened? In addition to, Riccardo said you can make calls with it almost by accident, but you could also do MP3 play. You could play video, you have games, you have web, you have ebooks, all of this on the same device. They get this amazing reciprocity between content that gets users, we get content that gets users in the ecosystem. What's fascinating, Sony could have done this. Sony had an MP3 player. It had an e-reader. It had cameras. It even had Sony Time Warner for crying out loud. They had the ecosystem play, but they didn't manage to do it. Google is not making this mistake with Android. As we just saw a moment ago, they built a phenomenal ecosystem of different components actually constructed together. We're doing very well together. So what happened? Well, after Apple introduced the iTunes ecosystem, Nokia stock price has gone from $33, so $30 a share to $8. Sony is from $53 a share down to $30. Microsoft Zoom gone as a product. Microsoft stock price has been flat only recently starting to come up, but they hadn't actually budged even though they had this large monopoly in operating systems. And we don't have to stop there. Polycom speakerphone, you just add Skype to your iPad. Cisco flip camera, you can do video on this already. How about HP calculator? That one's amazing. In fact, I have a perfect emulator on my phone already. I don't need to buy it, I can't calculate it. It's perfect. It is the HP calculator complete with reverse notation. It's remarkable. Don't have to stop there. TomCom geolocation services also absorbed into the ecosystem. Flickr photo sharing service, that's just iPhoto. RIM Blackberry, again, another product that's been crushed by an ecosystem play as opposed to a product play. We need to think of business ecosystems in competition with one another as opposed to products in competition with one another. And oh, by the way, as of today, we get watches. So I know there's data on the prices. I have no sales until actually today. So I think as of today is the first time you can actually order one. And that's now extending into the wearable space as well. Again, attaching a new section of the ecosystem. Amazon is not making this mistake either just like Google. Amazon has Amazon Prime. They've got a huge number of merchant services built on top of the system. In fact, Amazon Web Services has these merchants building stores. They have the market share of the next five cloud providers combined. Again, a phenomenal ecosystem play that allows them to observe everything that's happening among the ecosystem's competition. A remarkable effect. The point I want to make here is that in a market with network effects, our attention has to shift from inside the firm to outside the firm. The reason for this is you can't scale network effects inside the firm the same way you can outside. If there are more users outside, that's where you're gonna get scaled. That's where the value that is dependent on other people's participation is going to grow at a faster rate, which shifts the focus of attention from inside the firm to outside the firm. Let me see if I can give you more examples of why I think this is so different. Think of the giant firms at the turn of the previous century. We have the second machine age. Look, what was the original machine age? We have some fantastic examples. In electricity, there were two major players, Westinghouse and Edison. Again, for the giant supply side economies of scale. In this case, there was Henry Ford and Alfred Sloan. Again, one of the beneficiaries of the MIT Sloan School, phenomenal. This was in automobile production. Again, a supply side economy of scale. How about steel production? Carnegie are only a handful of steel manufacturers. Again, supply side economies of scale. Or I like this particular one. Vanderbilt was a colossus of railroads. Another enormous supply side economy of scale. What's happening in the network economy? It's really interesting. Economists have another name for network effect. The alternate name is a demand side economy of scale. This is the other side of that profit equation. This is value being created on the demand side as opposed to value being created by lower cost supply. Network effect is a demand side economy of scale. Think of the giants in this area also are large corporations, windows. Also a demand side economy of scale with a network effect. That was Bill Gates. How about this one, micro blogging? Again, very large network effect, a large economy of scale. Or how about Alibaba? Jack Ma, now one of the richest men in the world thanks to a merchant that is the size of Alibaba, I'm sorry, the size of eBay and Amazon combined. Again, another series of external marketplaces. Or Facebook, you'd almost look at the colossus of social networks and roads. It's really quite remarkable. Each of these ecosystems having an interesting side. Again, arguing, this is a network effect on the demand side of the economy rather than supply side of the economy. Again, why do you think this might be happening? So what changes in this? What are some of the other changes in the economy? Well, we teach in business schools lots of different practices. One of the things I think that changed are logistics and supply chains. In this case, platforms don't own all the factors of production. They tap new sources of value, and spare resources, and user communities. So favorite examples, how many of you stayed in a hotel locally before showing up here? Fair number. All right, sometimes they have wonderful chefs. They have fantastic old buildings. What happens if one of those rooms goes unrented for the evening? The hotel bears that cost. Airbnb, in contrast, has no such cost. They can scale arbitrarily. Airbnb is an amazing example of this externalized marketplace. Really Rides is another wonderful example of this. They will borrow your car and then rent that out. If we go to a park in Boston, for example, there's a $25 fee to go park your car per day. Really Rides will borrow your car, insure it, clean it, and pay you $10 a day. $35 net difference, so $25 out versus $10 in in order to use your vehicle, and they'll rent it out to people who just come out of the airport. Another one, Instagram sold for a billion dollars. Is it because of the software of the 13 employees or the contributions of the 30 million people who'd added value to the ecosystem? To actually see it, you can see another example of how much things have changed. Anyone recognize this photo? Anyone recognize this photo? And it's only eight years difference. It is excellent, the Pope, right? Very good. This is Pope Ratzinger and Pope Borgoglio. It's the exact same scene, only eight years difference. There's an enormous amount of additional information captured. If you can position your platform at the choke point of that information reign and you can collect the taxes on the activity that are taking place there, you may be able to create an interesting business model. Another fun example of the same phenomenon. Here's one. He goes to the moon and takes five photos. She goes to the bathroom and takes 37. There's a little difference here. You can actually see what's happened with technology over time. There was a beautiful quote describing what happens in the web fairly recently. Tom Goodwin made this observation. In 2015, Uber, the world's largest taxi company owns no vehicles. Facebook, the world's largest media owner creates no content. Alibaba, the most valuable retailer, has no inventory. In Airbnb, the world's largest hotelier has no real estate. Think of all that value being created externally. In business school, we teach just-in-time inventory. I like to say this is not even mine inventory. I'm selling someone else's stuff. Another example. What changes also is marketing and prices. There are whole books now written on the power of pull as opposed to push. Advertising used to be pushing advertisement out. Now, it's customer self-service as another way to do it. Another one is monetizing platforms using free pricing. We get enormous amounts of value from Google and Airbnb and others by allowing us to do these free activities. Now, was Google monetizing its developers by giving them free stuff? Was that an advertising model? It's actually a different model. Let's see if we can explore what might actually be happening. So I'm going to argue that your intuitions in these marketplaces are probably pretty good. Here's another kind of matching market. Suppose you're opening a bar. Which side of the market do you subsidize? The guys or the girls? I hate to tell you guys it's not us. It's the women. There's going to be a ladies night where they get in free and you charge the men. You would have a very different kind of bar if you priced the other way around. This particular bar, the Brass Monkey, has another version of the same pricing model in Taiwan that I really like. This is English-speaking night, where they give a discount to one side of the market and they charge more to the other side of the market. The rule of thumb is to subsidize the side that creates more value or is the stronger attractant. You have two different sides of the market and you differentiate the pricing in order to make that work. To show you how it works, I mean, at MIT, we developed the math behind it. I promise not to give up any equations for lunch. But here's a graphic representation of how this actually works. If you have those multiple sides of the market, this would be those two columns I showed you earlier. You could price to either one, and you would sell this much at this price and this quantity to this market, and this price and this quantity to this market. What happens if you give one of them away? If that's the case, your consumption goes from here to here, but with a network effect. You've driven and changed the size of your market over here so you're now selling this much as opposed to just this much. Of course you're not making money over here anymore, but you're now making money in spades over on this side of the market. That's how the business models work. That's the financial mechanism that makes this stuff viable. Most folks think of free as either advertising or they think of the razors and blades or they think of cell phones in minutes. That's not the same thing. For razors and blades, the same person pays for the razor and the blade or the same person pays for the cell phone in the minute. That's only one individual or one portion of the marketplace. In these two-sided markets, eBay will let one group on for free and they can charge more for the others or Uber might give you free rides to get you on the platform and then be able to charge more through the pricing mechanism through on the divers side. You subsidize one part of the market in order to charge more of another side of the market. That's different than razors and blades and cell phones in minutes. It's a fundamentally different economic model. For work on the sizes of these firms, notice the large antitrust mechanisms. We saw this earlier in the turn of the previous century. We're starting to see it now. Google has 91% market share here. Uber has a dominant form in the taxi market. The Nobel Prize in Economics was just given to Jean Tareau on regulation of such markets, market power in these two-sided platforms. It's really quite remarkable how much these issues are coming to the fore. Another element that changes. Again, so many of the different things that we do in business schools change. I want to argue that finance changes as well. Here's another fun example. And if you see the debate that took place on how to value Uber, this is a particularly good one because there was an interesting question that happened only June of last year, less than a year ago, in which Aswath Damodaran, a very famous finance professor, was wondering how in the hell is venture capitalists are pouring $1.2 billion into a corporation with only a few hundred million in revenue? This is mind-boggling. It makes no sense at all. He used the classic tools of finance. He estimates the size of the global taxi market. He gives them a 10% market share. He does a risk-adjusted series of cash flows and he considers what are their proprietary methods or barriers to competition? How easily could someone else come into this marketplace? He then does the risk-adjusted cash flows and comes up with a value of 5.9 million. As a marvelous fellow, to his credit, one of the things he really did nicely is he put his spreadsheet up on the web so anyone could check the numbers and see whether or not they agreed. One who took up the challenge was Bill Gurley, one of the investors in Uber. And he said, oh, by the way, those numbers are correct, but you're forgetting something. You're forgetting network effects. To use this, he puts up a chart that's not even his own. It's from David Sacks, the CEO of PayPal and the current CEO of Yammer. Hey, he says, the more riders you get, the more drivers you get, the more riders you get, the more drivers you get. There's a feedback effect in here. And you can even get some geographic effects where the price falls down, you can actually increase the size of the market. Once that happens, you're not talking the taxi market anymore. Now you're starting to encroach on the rental car market. After that, you start to encroach on the second car replacement market. After that, you can start to encroach on the logistics and the delivery market. Wait a second. We're no longer talking about movement up and down a demand curve. What we're doing is we're fundamentally moving a demand curve. We're reshaping an entire marketplace. With that, we get an estimate of 17 billion. And oh, by the way, if you focus on just the data in San Francisco where Uber operates, they had already expanded the taxi market by a factor of three, so 5.9 times two. You're pretty close to 17 right there. And by the way, did anyone know what the market cap of Uber is today? This was June last year. What is it now? 40 billion, 40 billion. It wasn't even 17. It's remarkable how much of these things make a difference. The last idea I want to give you is actually on R&D and innovation. You might originally have thought about internal R&D labs. One of the things that you can do is capture ideas from other sources. A good advocate of this is Mark Andreessen who argues an attribute of platforms is that they allow them to be adapted to countless needs and niches that the designers didn't anticipate. You get third-party ideas. Here's a wonderful example of that in previous history. What was one of the first truly modular products in the physical space? How about the Ford Model T? Users could disassemble and repair themselves and make new things out of them. So here are other things that users happen to make. Ford Model T hay carrier. Flower Mill. Race car. Mobile church. Snowmobile. Sawmill. And goat carrier. All of these incremental innovations adding value to the ecosystem. Third-party examples of this. Do these things matter? Here's a wonderful chart of the market accesses to MySpace and Facebook, both of them being social networks. Now you might think with large network effects that MySpace being the leader in this would have had a first mover advantage. MySpace is in red. Facebook is in blue. Over the period 2004 to 2008. And who knows what this first bump is in MySpace? I mean, sorry, in Facebook market access? Ah, close. Hang on to that one. You're very close. You're on the better target. It's not the games yet. This one was opening to .com. So this was .edu to .com. The thing that really mattered was what you mentioned in the audience. Opening to the games, exactly correct. Once you got developers in the ecosystem, you got mafia wars. You got Farmville. Folks are now spending hours on the system and inviting their friends in, which causes more developers to want to be on, which causes more people to want to be on the system. You're getting the convex growth again. That's when Facebook managed to take over MySpace. You don't have to take my word for it. You can actually listen to read the words directly the founder of MySpace, DeWolf, who said, we tried to create every feature on our own and said, we can do it. Why should we let third parties make any money? We should have picked five to 10 key features, totally focused on that, and let other people innovate on everything else. He made exactly the same mistake Steve Jobs made in the 1980s and 90s, all over again. You need to open to third party innovation. I'll give you an illustration of this. And actually, this is the long tail diagram. Eric did some marvelous work on the incremental value that Amazon brings to the marketplace by allowing folks to indulge in the long tail. We can apply those same ideas here. If this is Windows, or this is a mobile phone and it comes preload with the main apps, this should be Word, PowerPoint, and Excel. What's interesting is you don't need, this is the stuff out on the long tail, and you don't need to own this. Apple didn't build angry birds. They didn't even conceive of angry birds. The other thing I want you to notice is what's the proportion of value in this portion of the chart relative to the portion of value integrated over this portion of the chart. It's a tiny fraction when I'm over here, but the integral of all that value is enormous. It dwarfs the value created by the firm itself. For an R&D group, you can even imagine, this is the R&D value add of an individual firm. This will be the internal group. What's the pace of innovation? If you open to external innovation, you start to look something like this. Even if you start from behind, you add the incremental innovation to your own and you can get a higher rate of slope. So even if you start from behind, you have to overtake the leader. This is another reason why I argue platforms beat products every time. If you can harness third-party innovation, your rate value accretion has to enable you to grow faster and better over time. There's also another really neat trick. What's your cost of failure? If other people are doing the experimentation, there's almost zero cost of failure. What's your upside? 30%. That's a hell of an innovation trick. If your R&D department can actually shed the cost of failure but capture 30% of each of the successes, that's a really neat innovation trick. It's another reason why I would argue that platforms tend to beat products over time. Just taking a look at where this might go, we could also consider a couple of other industries. One would be city as platform. This is a chart provided by some other colleagues at MIT in the architecture school where they're creating a data layer looking across the city and you can actually look at weather patterns, taxis, shipping. You could even reroute the taxis in this area to the rain that's taking place in that area and make everyone better off. You're creating apps on top of that data layer and creating whole new environments. Another one, energy as platform. You might think that heavy industries would not necessarily be a place where you could do this kind of thing, but no, there too, it's also possible. One of the most difficult things in energy is what happens when you hit peak capacity. Bringing on new capacity is extremely expensive. Firms like EnerKnock are doing demand management and having individual consumers or businesses turn off their demand in order to keep it below the threshold required to bring on new forms of capacity. Another one, affecting all of us, education. What happens, MIT is now sponsoring edX, MOOCs, massive open online courses. Now it might be possible for a kid in Bangladesh or in Romania to take classes from MIT on the MOOC or some of the best content from anywhere in the world. Even places like Harvard could be added to the platform and then made available to everyone. We're creating this long reach of these platform ecosystems or another one is healthcare. Nike has this an example, but if we expand it to a whole ecosystems of doctors, insurers, medical products manufacturers, you create entire ecosystems that can innovate and then actually create and capture additional sources of value. So the ideas I wanted to bring you are I really do think that platforms are building products. It's a fundamental difference in business model and there's some evidence to suggest that these forms of business are becoming dominant. The drivers are communities and network effects. Again, in economic speak, we think it's the demand side economies of scale not just the supply side economies of scale and this changes any number of different things. It changes supply, it changes prices, financial indices, innovation, strategy, all different kinds of elements of the business model. So if you think that there's a value in people connecting and building things, this is a possible business model that I would encourage you to embrace or whatever industry you might be in, think of how you might convert that or how it might affect your industry going forward. And with that, happy to take any questions on the nature of platforms, businesses or what we can do next. Thank you. Marshall, we have a question from the cube crowd chat. Does an open platform strategy leave you open to being substituted or vertically integrated into something else? Great question. All right, so what's the optimal level of business openness, shall we say? You can go, openness helps you and bring in third party innovation but it has two risks once you go out. One is that you can lose control of the ecosystem and two, you can lose control of the money. Great examples of that are the Linux operating system. It could have moved over to mobile ahead of any of the other ecosystems but it didn't have sufficient control. Another interesting example is Android is now gone perhaps a little too far open. Android has 80% market share in mobile devices but it's forked. There's now the Google controlled version and there's the Android open source version that's now growing in Asia and they've lost control of it. In order to avoid that, you don't go all the way open, you maintain a portion that's closed, a portion of the property rights that you own, you maintain some proprietary complements that you can still charge for and that will actually help you continue and sustain the business model. You can go too far in the openness. It's a wonderful question. I think we have time for maybe one more. Yes, I know we have time for a few more. Thank you, Marshall. Don't take this personally but how do you see with all this your role changing as a professor, both in your role and in your environment? So I actually, my biggest fear as academia becomes increasingly MOOCs or massive open online courses is that I think professors like myself will write the lectures and Morgan Friedman will deliver them. I actually think it's a great question. I think this is one Dean Schmidt line and others should be thinking about over the long haul. My speculation is actually two-fold. One, I think that the commodity content, the stuff that's really established, Econ 101, CompSci 101, will be taught as well as possible in just a couple different versions. There are 3,000 colleges and universities in the United States. There's no reason why there need to be 3,000 versions of Econ 101 and CompSci 101. Those are very standardized classes and I think those will move increasingly online to the best of breed, best pedagogy possible. That being said, I also think that there's a huge reason even why we're here. I think the best research is not commoditized yet. The best ideas, the novel ideas, the stuff that Eric and Andy are doing, the stuff that Ricardo is doing, these things are novel and not get commoditized and that's I think where the research institutions have an incredible role to play and where we still educate doctoral students, where there's a chance to be near the leading edge, will change a lot. I do think that there will be some shakeout in academia. I do think that there will be fewer schools than there are now and I think there's gonna be a greater need for differentiation. The thing that's protecting academia mostly is still the credentialing and certification that goes in place but I actually think there are likely to be some big changes. It's again, very interesting question, very interesting dilemma to looking at for the long run. And again, we'd love to have your participation in the redesign of academia and indeed the redesign of research. Yes, we've got a couple. One of your examples towards the end was about creating platform economics and healthcare. But how do you see the challenges of creating a platform business model in a regulated industry like healthcare or banking compared to sort of frivolous things like Farmville? So one of the greater difficulties, there are two sets I think that would make a big difference in healthcare. One of them is the increased regulatory hurdle relative to other industries. So for example, in the United States there are the HIPAA requirements about maintaining privacy of your own health data or who has access to those health data or the certification of doctors and able to explore that. There's also the number of different participants in the ecosystem that I think are increasing the frictions in the transition. But I think that the efficiencies are sufficiently great and the innovation potential is sufficiently great that we will see this increasingly in healthcare. And in fact, I'll give you a wonderful example of how this is actually playing out. Think of an MRI machine. Multi-million dollar piece of equipment sold by major corporations, Siemens here in Europe, Phillips, GE in the United States. The utilization rate of this multi-million dollar machine is around 47%. There are becoming new firms analogous to Airbnb for MRI machines in order to increase that utilization rate. It makes perfect sense once you see it, but you need to get the mechanisms in place to make that actually happen. I'm actually talking with a startup, Coheelo, that's trying to do exactly that kind of thing. So I think there are some regulatory hurdles we overcome. I think there are multiple players that need to be worked with, but I think the efficiencies to be had make the transformation of that industry almost inevitable. Yeah. Do you think markets are becoming winner-ticket-all markets, like monopolies in a sense? And do you see platform having this power to expand the segments of market? That's a perennial issue that I think we're going to see again and again and again. I like to use the historical precedent as examples of that. At the turn of the previous century, industrial era, again, I argue that the winner-ticket-all markets were driven by supply-side economies of scale. The dominant factors that lead to winner-ticket-all markets are supply-side economies of scale, or within the other side, there are the network effects. There's also multi-homing. It's the ability for users to cross different ecosystems. So for example, you can carry a master card and a visa both pretty easily because your multi-homing costs are quite low, but it's very difficult for you to carry an Android and an Apple at the same time because those multi-homing costs tend to be rather high. You have monthly recurring charges. The higher the multi-homing cost, the greater the tendency toward market conversion. Or switching costs is another one. Again, if we look at network effects as another demand-side economy of scale, it's almost certain that we'll be seeing a large number of these firms that are approaching monopoly status within their marketplace. Google has 91% market share search in Europe. It's quite remarkable how that's the case. Alibaba has by far the dominant market share in retail, in China. Again, I do think that we will see a lot of the regulatory issues of the previous century transported to this century, but in the demand-side context. So yes, I think we'll see a lot of the one to take all effects taking place. The Uber store was very interesting in the valuation example. It's not a very complicated story. So what's the barrier to entry? Why can't I just go do the same thing? Because it's pretty simple for me to find Google Mac, maps, and start a network and compete. So why is evaluation 40 billion and not some other number? Lower. Okay, let me see if I can paraphrase the question again. So why would some of these industries where there are apparently no switching costs or very low switching costs have such A, high concentration, B, high valuation? Let me give you two examples. So we'll take the Uber example as one, take Google as another example. One of them, demand-side economies of scale actually are kinds of barriers to entry. Imagine that you have managed to sign up millions of drivers. The startup, then you may have millions of riders that come with it. The startup is not having to produce simply the app. You also have to recruit the drivers and the riders in order to join the ecosystem. To give yourself another intuition, imagine trying to start Facebook from scratch. Now the physical assets are small, but you'd have to transport all of those octopus tentacles of social networks into the new network. In the same way that if you were to try to displace Uber, you would need to transport the drivers over to the other ecosystem. So there's some protection in that network effect that's taking place. Another element of it is the data-driven feedback loops. So when you do a Google search, your results and what you click on are then used to inform my, what's presented to me next. And when I search it's presented to the next person. That data-driven feedback loop helps to make the service better and better and better so you get almost a data barrier to entry because the service is improving at high rates. Again, it's really easy to click over to Bing or Yahoo or other search engines, but the search isn't quite as good because they don't have as good data because they don't have as much participation. You've got a chicken and egg problem getting the participation in order to get the data and that's another reason of it. Both of those things scale really well. I'll give you another example in the same space. Do you remember that Yahoo when it launched was categorizing the web manually by employees? That can't scale in the way that algorithms with Google can. So they got scale, they have network effects, they have data-driven feedback loops. Those things tend to make a difference.