فصل 1

Chapter 1

Platform Thinking

Introduction

It is a well-known story. In July of 1980, several executives from IBM visited a young Bill Gates, cofounder and CEO of a five-year-old company called Microsoft. Gates already had a reputation as the best source of programming languages for personal computers, a fledgling new market. IBM was planning to introduce a personal computer for businesses, and it wanted Microsoft to provide the operating system. At first, Gates demurred. He suggested that IBM speak with another start-up called Digital Research. Those discussions failed, so the IBM team returned to Seattle. This time, spurred on by employee Kazuhiko (“Kay”) Nishi and cofounder Paul Allen, Gates decided to accept the job. Microsoft bought a crude operating system for $75,000, fixed it up, and named it MS-DOS, for Microsoft Disk Operating System. The less well-known part of the story is how Gates structured the deal with IBM. The basic contract was fairly routine: He charged IBM a development fee of $200,000 and as much as $500,000 for additional engineering work. Gates also gave IBM the rights to DOS and several programming language products it could bundle with the new computer. But—and this is crucial—Gates allowed IBM to use the operating system with no additional fees or royalty payments as long as Microsoft, and only Microsoft, could license the software to other manufacturers. What was Bill Gates thinking?

Gates was familiar with the “clone” industry that had emerged around the IBM mainframe during the 1960s and 1970s. The clones created a new but relatively small business around software and services for IBM-compatible machines. It occurred to him that, if the IBM PC became popular, then a new mass market would likely emerge for the personal computer. If he alone had the right to license the operating system to companies that wanted to build compatible machines, then Microsoft would be at the center of a whole new industry. Indeed, over the next several decades, this industry—today we would call it an “ecosystem”—attracted thousands of software and hardware companies, produced millions of “complementary” software applications and peripheral devices like printers, cameras, and game controllers, and today still counts more than a billion users.

Gates’s decision to give away the basic software to IBM in return for the right to license it to other companies is now a famous and striking example of “platform thinking.” (This agreement was also one of IBM’s worst strategic decisions.) But how would Microsoft make money if it gave away the software for free to its main customer? Gates realized there would be more than one “side” to this new market. His goal was not to maximize profits from the sale of DOS to IBM as a stand-alone product. Instead, the strategy was to make the operating system into an industry-wide platform—a foundation that many companies could use to build personal computers and compatible software applications. IBM seemed intent on controlling the PC market with production of its own personal computers, using Microsoft’s software as a component. To Gates, however, encouraging many firms to invest in making IBM-compatible PC hardware and software applications would make the personal computer—and especially Microsoft’s operating system—increasingly useful and valuable. Gates would soon enter the software applications business himself to grow the IBM-compatible PC market and take more of the profit, with Word, Excel, and PowerPoint. He did this first for Apple’s Macintosh computer, introduced in 1984, and then for DOS and Windows PCs, bundled in the Office suite from 1990. To encourage other companies to help expand demand for PCs, Gates also decided to give away for free the software development kit (SDK) needed to build applications for DOS and then Windows.

By contrast, Apple cofounder and CEO Steve Jobs did not give away software development kits for free or try to build a broad applications market. Instead, he hired Microsoft in 1982 and paid Gates a $50,000 advance to write applications for the Macintosh personal computer, which was incompatible with DOS. Jobs also charged hundreds of dollars to developers who wanted to build Macintosh applications on their own. The development fee was in addition to large expenses that programmers usually had to incur in order to design applications. Most expensive was the $10,000 Lisa, a failed predecessor to the Macintosh that nonetheless served as a software development platform before the Macintosh was available. Programmers also had to buy some programming language and database products. But, Jobs reasoned, with an easy-to-use graphical interface, the Macintosh was going to be such a great product that companies should pay him for the right to build applications. Partly due to the resulting paucity of applications software as well as the high price of the hardware (about twice as much as an IBM-compatible PC, because Apple was the only manufacturer and there was no price competition), the Macintosh never garnered significant market share. Ultimately, PCs running DOS and then Windows—which mimicked the easy-to-use Macintosh user interface—captured roughly 95 percent of the market for personal computers. Microsoft was thinking platforms. IBM and Apple were thinking products.

The personal computer, like social media, online marketplaces, cloud computing, and smartphones in more recent years, turned out to be a platform business, not a product business. In the case of the personal computer, by this phrase we mean that, unlike in traditional businesses, success did not depend simply on the quality, price, or timing of Microsoft’s operating system as a stand-alone product. Success depended more on complementary innovations that determined what users could do with the product—such as the number and quality of software applications or digital services produced by many companies. These “complements” added significant and even essential value to the core product, what we now call an innovation platform.

To turn its product into a platform, Microsoft also had to solve a critical “chicken-or-egg” problem: how to encourage other companies to build the software applications needed to stimulate demand for PCs. It turned out that broad and cheap licensing of the operating system facilitated the production of low-cost hardware by many companies around the world. Then the rising number of PC users using the same technology created demand for programmers to design increasing numbers of compatible software applications. Who won and who lost depended less on product quality or features and more on who could best bring multiple “sides” of the emerging market together and generate positive “feedback loops.” Fast-forward to April 2018. Mark Zuckerberg, cofounder and CEO of Facebook, was on the hot seat, testifying before the U.S. Congress. His company, established in 2004, started out by building a simple personal computer application accessed via the web. By 2018, Zuckerberg’s free software and services enabled more than 2.2 billion people to send messages, share news stories or digital content like photos and videos, organize groups, send money, and do a myriad of other activities with friends, relatives, and acquaintances, as well as business partners and customers. In the initial stages of the company, Facebook users actively brought in their friends, and then friends of friends, and friends of friends of friends, weaving together a connected network of people that quickly spanned the globe. This network, aided by new features that Facebook enabled, made the social network increasingly valuable as a transaction platform for communications, electronic payments, and other purposes, as well as its core business—selling context-specific advertisements. In 2007, following the lead of Microsoft (which, not coincidentally, was a major investor in Facebook), Zuckerberg started to make Facebook’s data on users and other functions available as an innovation platform—a kind of operating system for social media applications. This decision empowered outside companies and independent programmers to design games and other applications that soon came to number in the millions, and made Facebook an even more compelling experience.

But platforms do not always evolve in predictable ways, especially when they are able to add so many new functions from both inside and outside the firm. In 2014 a researcher at the University of Cambridge working with a small British consultancy named Cambridge Analytica (now bankrupt) built one of those millions of Facebook applications. Its main purpose was to track the preferences of users and their friends. The application provided data on as many as 87 million unsuspecting Facebook users in the United States and helped Russian hackers target particular users with fake news stories supporting candidate Donald Trump and criticizing rival Hillary Clinton. The U.S. Congress called in Zuckerberg to explain how his seemingly harmless social media platform had become an instrument of such power for a foreign government. Zuckerberg explained in his written testimony: We face a number of important issues around privacy, safety, and democracy, and you will rightfully have some hard questions for me to answer. Before I talk about the steps we’re taking to address them, I want to talk about how we got here.

Facebook is an idealistic and optimistic company. For most of our existence, we focused on all the good that connecting people can bring. As Facebook has grown, people everywhere have gotten a powerful new tool to stay connected to the people they love, make their voices heard, and build communities and businesses. Just recently, we’ve seen the metoo movement and the March for Our Lives, organized, at least in part, on Facebook. After Hurricane Harvey, people raised more than $20 million for relief. And more than 70 million small businesses now use Facebook to grow and create jobs.

But it’s clear now that we didn’t do enough to prevent these tools from being used for harm as well. That goes for fake news, foreign interference in elections, and hate speech, as well as developers and data privacy. We didn’t take a broad enough view of our responsibility, and that was a big mistake. It was my mistake, and I’m sorry. I started Facebook, I run it, and I’m responsible for what happens here. How Did We Get to This Point?

For anyone who follows the world of business, it is now common knowledge that the most valuable firms on the planet and the first companies to surpass the trillion-dollar mark in value (albeit temporarily) are platforms. If we look at market values in late 2018, the top firms were Microsoft, Apple, Amazon, and Alphabet (the holding-company parent of Google since 2015). Also among the leaders were Facebook, Alibaba, and Tencent. Together, these seven companies at their peak represented close to $5 trillion in market value. Moreover, in a recent list of more than two hundred current and former “unicorns”—start-ups with valuations of $1 billion or more—we estimated that platforms made up between 60 and 70 percent. These were led by firms such as Ant Financial (owned by Alibaba), Uber, Didi Chuxing, Xiaomi, Airbnb, and other well-known private companies (several of which planned to go public in the near future). So, yes, Mr. Zuckerberg (and Mr. Gates)—how did we get to this point? How have a small number of companies come to exert such enormous influence over our personal, professional, and even political lives, as well as the world economy? There is nothing new about marketplaces; they go back millennia. But digital platforms that span the globe are new. How have they come to control the flow of information as well as such a large number of goods and services? In what ways are these new entities different, or similar, to the powerful corporations we have seen in the past? And are there limits to the market dominance and expansion of these digital juggernauts that can leverage user data as well as scale and scope economies in ways we have never seen before?

These are not simply rhetorical questions. The world is full of existing as well as emerging platform battlegrounds that will have great influence on our lives in the future. We can foresee a time when digital platforms and associated ecosystems will be the way we organize new information technologies such as artificial intelligence, virtual and augmented reality, the Internet of things, health care information, and even quantum computing. We can also see peer-to-peer transaction platforms replacing or competing with traditional businesses, especially as the “sharing” or “gig” economy expands and new technologies diffuse. Use of blockchains (distributed ledger technology that is extremely secure though not unbreakable) and cryptocurrencies (digital money, usually independent of banks and governments) may greatly reduce the need for many different services, from traditional banks to supply-chain contracts and monitoring.

Yet another hot topic as we write this book is increasing demand for governments to rethink data-privacy laws, antitrust laws, and other regulations that could rein in the most powerful platform businesses. Platform companies have faced antitrust challenges many times, and the incidents are likely to increase. The European Union fined Alphabet-Google $2.7 billion in 2017 and $5.1 billion in 2018 for anticompetitive behavior involving Google Search (which at that time had about 90 percent of the global market outside China and Russia) and its Android smartphone operating system (which accounted for about 80 percent of the global market). In fact, Google Android has replaced Microsoft Windows as the most popular operating system in the world, with over 2 billion users. If we combine data from Internet searches with Google’s Gmail (which has over 1 billion active monthly users) and Google’s YouTube (which has close to 2 billion users), in addition to individual profiles that Google generated for targeted advertising, then Google probably has far more personal information than even Facebook could muster. Another aggressive platform company, Amazon, was collecting vast amounts of data on its hundreds of millions of users and their transactions, and coming under rising scrutiny in the United States. With more than 500 million individual products for sale, Amazon has disrupted markets such as books, consumer electronics, digital music and video, cloud computing services, groceries, pharmaceuticals, and package delivery. How should government regulators, as well as competing firms, respond to these new centers of power?

These are the questions we tackle in this book. For approximately thirty years, the authors have been studying and working with platform companies that emerged to build essential technologies and software applications for the personal computer, the Internet, and smartphones. Our books include: Microsoft Secrets (1995), Competing in the Age of Digital Convergence (1997), Competing on Internet Time: Lessons from Netscape and Its Battle with Microsoft (1998), Platform Leadership: How Intel, Microsoft, and Cisco Drive Industry Innovation (2002), The Business of Software (2004), Platforms, Markets, and Innovation (2009), Staying Power (2010), Software Ecosystems (2013), and Strategy Rules: Five Timeless Lessons from Bill Gates, Andy Grove, and Steve Jobs (2015). We have also written many articles, including a case study of Apple that, in multiple versions, has sold more than 1 million copies. Most of our earlier work, especially Microsoft Secrets and Platform Leadership, focused on the ability of platform leaders to inspire complementary innovations from third-party firms. But the world has since adopted a much broader view of how digital platforms impact business, politics, and society. This book builds on our prior work (summarized in the Preface) and that of our colleagues. The goal is to help managers and entrepreneurs, as well as policy makers, better understand how to harness the power of platform thinking while avoiding some of the negative consequences.

Most people know the names of companies that shaped the evolution of platform strategies and business models. Intel (established 1968), Microsoft (1975), and Apple (1976), along with IBM (1911), made the personal computer a mass-market phenomenon during the 1980s and early 1990s. A second wave of firms from the mid-1990s built Internet software and services on top of the personal computer, led by Amazon (1994), Netscape (1994), eBay (1995), Yahoo (1995), and Google (1998), as well as Rakuten (1997) in Japan and Tencent (1998) and Alibaba (1999) in China. In the next decade came social media, pioneered by Friendster (2002) and MySpace (2003), and then Facebook (2004) and Twitter (2006). More recently, billion-dollar start-ups, such as Airbnb (2008), Uber (2009), and China’s Didi Chuxing (2012), have brought great attention to the “sharing,” or “gig,” economy. They match smartphone and PC users with providers of rooms to rent or cars to ride as well as an almost unlimited number of other products and services. We now refer to all these firms as platform companies, even though they are not all the same.

Some people argue that the traditional rules of business no longer apply in this new age of digital competition. Woe to those who do not understand platform strategy and business models, big data analytics, artificial intelligence and machine learning, and what seem to be the new rules of the game. We believe there is considerable truth to this statement. However, we also believe there are several misunderstandings associated with the “digital revolution.” In particular, the path to success for a platform company is by no means easy or guaranteed, nor completely different from what we have seen before. Why? Because many platforms today are not sustainable businesses. To succeed long term, all firms must ultimately perform better than the competition, whether the rivals are digital platforms or conventional businesses. They must be financially viable as well as politically and publicly acceptable lest they become crushed by debt obligations, social opposition, government regulation, or global trade wars. These observations are common sense, but, amid all the hype over digital platforms, a phenomenon we sometimes call “platformania,” common sense is easy to forget.

The issues may be complicated but our argument is simple: Yes, managers and entrepreneurs in platform companies must understand the finer points of digital competition, innovation, and power. But they must also master the fundamental principles of business and good governance that apply in any company and in any era. Platforms will not generate profits simply because of adept use of digital technology, a clever “multisided” market strategy, or classifying all employees as gig-economy contract workers. If sales rely on large subsidies of one or more market sides, and the platforms continually operate at a loss, then the bigger they get, the more money they will lose. In short, managers and entrepreneurs in the digital age must learn to live in two worlds: the conventional economy and the platform economy. What this means and how to do this well is the subject of this book.

Platforms Defined

Before we get into more details, let’s clarify what we mean by “platform.” In everyday conversation, we hear the word used in many contexts, and this often leads to confusion. Politicians compete on ideological platforms—ideas or policies that bring people together for a common goal. People catch trains on physical platforms—designated areas that bring people together to access a shared mode of transportation. Companies create product platforms—common components and subsystems that different engineering groups within the firm and its supply chain (such as automakers or aircraft manufacturers) can use to build “families” of related products more efficiently than building each product from scratch.

Platforms, in general, connect individuals and organizations for a common purpose or to share a common resource. Our main concern in this book is with industry platforms that emerged in the wake of the personal computer, the Internet, and mobile communications technologies. These industry platforms also create building blocks or common functions for use within and outside the firm. However, the platforms function at the level of an industry (or ecosystem). More importantly, they bring together individuals and organizations so they can innovate or interact in ways not otherwise possible, with the potential for nonlinear increases in utility and value. Later in this book we will give some concrete examples of “nonlinear increases in utility and value.” Briefly, this means that the usefulness of an industry platform can grow with the power of the network: Each additional user, at least theoretically, can benefit from access to all the other users and innovations already available through the platform. What we have, then, is practical and economic value that can increase not by simple addition, as in adding one user or innovation at a time, which occurs in conventional business models. Rather, value with platforms can increase geometrically if each additional user can connect to all the other users or benefit from all the other innovative products and services already accessible to members of the network.

What are commonly called “network effects” are positive feedback loops that come from connecting different users and market participants to each other. The feedback loops can extend across entire ecosystems, which are broad linkages of producers, suppliers, users, business partners, and other stakeholders. We agree that building a business around network effects requires a different way of thinking about market dynamics and competitive strategy compared to conventional businesses. Platform businesses also have different ways of making money, since they may not directly sell a stand-alone product or service. At the same time, though, platforms do not all require “revolutionary” strategies and business models that did not exist before the digital age or that make conventional business logic obsolete. Nor is it always useful to think of platform companies simply as “matchmakers” bringing different market actors together, although that is a common function for many platform businesses. We also have argued for many years that, in a platform market, having the best platform is more important than having the best product. Look at the history of the Apple Macintosh. It was a better personal computer than the DOS or early Windows PCs in terms of ease of use and elegance of design. Despite its strengths, the Mac’s market share remained stuck in single digits for the last thirty years because it was not the best platform. The Macintosh was too expensive and more difficult to build applications for, and Steve Jobs did not encourage mass adoption, such as with lower prices or licensing of the technology to other companies.

To be sure, not every industry lends itself to a platform strategy. Often, a stand-alone product or service is the best way to beat the competition or earn the most profits. However, a platform strategy should prevail over a stand-alone product strategy when (1) there are opportunities to tap the innovation capabilities of outside firms to enhance value; and (2) it is more economical to enable transactions rather than to own assets and deliver products or services directly. Managers and entrepreneurs need to understand how product and platform strategies and business models differ and when to use each approach. In many though not all industries, platforms can create significantly more value than conventional businesses and traditional supply chains. In some cases, such as personal computers, smartphones, video game consoles, or even social media, innovations from outside firms can determine the success or failure of the platform business. Here is a quick summary of what makes industry platforms unique.

ENGAGE MULTIPLE SIDES OF A MARKET

First, industry platforms deliver products or services by bringing together two or more market actors or “sides” (e.g., buyers and sellers, or an operating system maker with users, application developers, and hardware producers) that would not otherwise interact or easily connect. The platform company may begin by targeting one side, such as buyers or users. Over time, though, platform companies usually link actors who want access to another side of the market, such as sellers or producers of complementary innovations. For example, Facebook began in 2004 by connecting Harvard College students to their classmates. It expanded quickly as friends brought in friends and acquaintances. Fairly soon, the company identified another set of market actors: advertisers. These companies had goods and services to sell, based on what people were communicating about. Then Facebook opened up its platform to a third side of the market: application software developers such as game producers or companies seeking to understand user behavior like Cambridge Analytica. And then came a fourth side: content providers, such as online newspapers, magazines, news sites, music sites, and others. (See Figure 1-1.)

GENERATE NETWORK EFFECTS

Second, as industry platforms connect users to other users or to other market participants, they generate network effects. The unique feature of network effects is that the value one user experiences potentially increases as more people or organizations use the same product or service and as more complementary innovations appear. Network effects can be strong or weak, positive or negative. When they are strong, the results are nonlinear increases in utility and value. These powerful feedback loops enabled Facebook to grow from two users to over 2 billion users in just a few years.

Network effects may sound like a vague concept, but we can explain how they operate with some concrete examples. With only one user of a telephone, fax machine, personal computer, or messaging app, these platform technologies generate no network effects and therefore have little or no value. Two people make these innovations more useful, three more so, and so on. More users encourage more users, which is a positive feedback loop. The user-to-user attraction is also an example of what we call a “direct” or “same-side” network effect. Similarly, a new smartphone operating system that has only a few users is not likely to attract many advertisers or developers of complementary software applications. However, more advertisers and developers are likely to appear if the number of users grows.

These examples illustrate why economies of scale are intimately related to network effects. In addition, negative network effects, such as declining user numbers or poor user ratings or too much advertising, can lead to increasingly rapid declines in usage. Friendster and MySpace in social networking as well as Nokia and BlackBerry in smartphones all experienced negative network effects and rapid declines in their businesses.

When one side of the market, such as users, attracts another side of the market, such as sellers or developers of complementary products and services, we refer to this type of network effect as “indirect” or “cross-side.” What is especially interesting here is that different market sides offer the potential for a platform business to generate revenue without directly building products or delivering services themselves. Furthermore, platforms can access different market sides to substitute for contracting with traditional suppliers, investing in internal firm capabilities, or directly owning critical assets. For example, Apple, Google, Microsoft, and Facebook did not have to build their own engineering teams or write contracts and pay third-party suppliers to create all of the millions of software applications that work on their platforms, even though they built some applications themselves. Similarly, Uber, Lyft, Didi Chuxing, and Airbnb did not have to own any of the cars and homes that their users accessed when they got rides or rented rooms, even though they may someday decide to own or lease vehicles and buildings.

SOLVE A CHICKEN-OR-EGG PROBLEM

Third, in order to link multiple market players and get the network effects started, industry platforms all must solve a chicken-or-egg problem. This means that one market side usually needs to come on board first and provide something that attracts another side. The dynamics vary with the type of platform and the specific business. Sometimes two sides need to come on board at almost the same time and grow together in a kind of zigzag fashion like credit card users and merchants. Nonetheless, the business challenge always remains the same: where and how to start, as well as how to get enough momentum and then scale. Solving the chicken-or-egg problem and then generating strong network effects can be very difficult if one side of the market realizes value only when another side is fully engaged. This is usually the case. As a result, platform companies must decide which market side to line up first: drivers or potential ride-sharing passengers, people with extra rooms to rent or potential renters, or smartphone handset makers or software app developers.

Traditional companies generally have more direct influence over how customers perceive their stand-alone products and services. The companies may depend heavily on suppliers but not so much on third-party firms to voluntarily make complementary investments. Many new platforms also fail to grow because they misjudge which side of the market is most important. This is one reason why the Apple Macintosh never achieved more than a modest share of the global personal computer market. Other new platforms never get beyond the initial stage because they require too much money to subsidize one side of the market and run out of cash or venture funding before generating large enough scale economies and strong enough network effects to be profitable and survive.

Platform Business Models: Two Basic Types

Platforms in the digital economy can do many things, and we might construct a complex typology based on the huge variety of applications. However, to make things simple, we divided digital platforms that emerged with the personal computer, Internet, and smartphones into two basic types, depending on their primary function. (See Figure 1-2.) Chapter 3 discusses this typology and the different strategic and operational challenges in more detail, but here we offer a brief overview.

The first type we call innovation platforms. These platforms usually consist of common technological building blocks that the owner and ecosystem partners can share in order to create new complementary products and services, such as smartphone apps or digital content such as from Apple iTunes or Netflix. By “complementary,” we mean that these innovations add functionality or access to assets that make the platform increasingly useful. The network effects come from the increasing number or utility of the complements: The more there are or the higher quality they are, the more attractive the platform becomes to users and complementors, as well as other potential market actors such as advertisers (and investors). Microsoft Windows, Google Android, Apple iOS, and Amazon Web Services are commonly used operating systems and cloud computing services that serve as innovation platforms for computer and smartphone ecosystems.

The second type we call transaction platforms. These platforms are largely intermediaries or online marketplaces that make it possible for people and organizations to share information or to buy, sell, or access a variety of goods and services. The more participants, functions, and digital content or services available through a transaction platform, the more useful it becomes. Again, it is mostly the digital technology and scale that make these platforms unique and powerful in today’s world. Google Search, Amazon Marketplace, the Facebook Social Network, Twitter, and Tencent’s WeChat are examples of transaction platforms used by billions of people every day. Credit cards such as Mastercard, Visa, and American Express, as well as catalogues such as the Yellow Pages (think of this directory as bundled with the telephone), are transaction platforms that originated before the digital era.

There are important strategic differences between the two platform types. Innovation platforms usually create value by facilitating the development of new complementary products and services, sometimes built by the platform owner but mostly by third-party firms, usually without supplier contracts. Firms often capture and deliver value (monetize the platform) by directly selling or renting a product. In a few cases where the platform is free (e.g., Google Android), firms monetize the platform by selling advertising or other services. By contrast, transaction platforms usually create and deliver value by facilitating the buying and selling of goods and services or facilitating other interactions, such as enabling users to create and share content. The firms that own this type of platform primarily capture value by collecting transaction fees, charging for advertising, or both.

Some firms start with one type of platform and add the second type, or mix and link the two. We refer to companies that support both types of platforms as hybrids. Some people use the term “hybrid” to refer to companies like Apple, Oracle, SAP, or Salesforce that emphasize a combination of product and platform businesses. In this book, however, as we discuss in Chapter 3, we use “hybrid” to refer to the combination of innovation and transaction platform strategies within the same company or within the same platform infrastructure.

What the Data Says

Clearly, public and private capital markets have placed enormous value on platforms associated with personal computers, Internet services, and smartphones. Yet no one has done a systematic analysis of the performance of these relatively new companies over long periods of time and compared them to firms in the conventional economy. To fill this void, we analyzed two decades of company performance. We started in 2015, when we began writing this book, and then went back twenty years to 1995, when the mass-market Internet first exploded onto the scene with browsers from Netscape and Microsoft. Using the Forbes Global 2000 list, we identified forty-three companies in 2015 that laid the foundations for the new digital platforms—eighteen for innovations and twenty-five for transactions. (For the list of companies, see Appendix Table 1–1.) Our basic rule of thumb was that a company had to have at least 20 percent of its revenues from a business dependent on network effects to count as a platform. Note that we excluded older platform companies such as telephone and telecommunications operators as well as credit card companies, although they remain in the Forbes Global 2000 sample and continue to be important examples of platform businesses.

Our first finding was that publicly listed companies associated with personal computer, Internet, and smartphone platforms are relatively rare; we counted only forty-three out of approximately 2,000 firms. We also compared these firms to all the Forbes Global 2000 non-platform companies in the same industry categories. The companies in this industry control sample were similar in terms of revenues to the digital platforms, with median sales of about $4.8 billion compared to $4.3 billion, respectively. (See Table 1–1.) After using a variety of statistical controls, we found no meaningful differences in sales between platform companies and the non-platform control sample. What we did find is that, despite comparable revenues to other firms in the same industries, our sample of platform companies had about half the number of employees, much higher operating profits, and much higher market values as well as higher ratios of market value to sales. The platform companies spent significantly more on R&D and other expenses related to sales, marketing, and administration, but they also grew faster in revenues and market value. Similarly, the platform companies were more productive (in terms of sales per employee), much more profitable, and much more valuable than conventional public firms in the broader world economy, as represented by the Forbes Global 2000 list. To summarize, by almost any measure, the forty-three digital platforms in our sample performed extraordinarily well. We also checked our analysis by removing the largest companies (Apple, Amazon, Microsoft, and Google) and ended up with similar (and statistically significant) results.

*** = p < 0.001 for Industry Control Sample vs. Innovation & Transaction Platforms comparison using two-sample Wilcoxon rank-sum (Mann-Whitney) test.

Mkt Value-Sales Multiple = ratio of market value compared to prior year sales.

S&M + G&A/Sales = Sales and Marketing Expenses + General and Administrative Expenses divided by Sales.

Growth numbers refer to prior year data.

Observations refer to the number of years of data for each firm, which depended on when the companies went public. We had on average 13 years of data for the 18 innovation platforms and 5 years of data for the 25 transaction platforms.

This data also allowed us to compare innovation and transaction platforms. The sample is small, but the two types of platforms clearly seemed different from each other, even though all relied on network effects to drive at least part of their businesses. We see in Table 1–1 that innovation platforms, compared to transaction platforms, were four to five times larger in terms of median sales and employees, and had median market values about three times higher. Innovation platforms also spent relatively more on R&D as a percentage of revenues and less on sales, marketing, and general administrative expenses. However, transaction platforms were growing faster in revenues and market capitalization, and traded at higher ratios of market value relative to sales. In other words, investors considered transaction platforms more valuable relative to revenues compared to innovation platforms.

We also analyzed the annual reports of the 43 publicly listed platform companies going back several years and counted 209 direct competitors (public and private) that were platforms and went out of business. In effect, the 43 companies in our sample were the survivors of this competition, and only 17 percent (43 out of 252) remained in 2015 as independent public companies. We discuss this data further in Chapter 4.

We can make one additional observation, looking at the financial performance of individual firms: It was often unclear if a company was successful primarily because of its platform strategy or its product strategy. Take Apple as an example. Is this company so profitable and valuable primarily because of strong network effects and the multisided market strategy surrounding the iPhone, which now accounts for about 60 percent of revenues? Or does a large part of Apple’s market value come from its design skills, brand, and ability to charge premium prices for the iPhone and some other products and services? We can ask similar questions about other companies: How much of the success of Cisco, Oracle, SAP, or Salesforce came from platform strategy and network effects versus the excellence and brand value of their products and services? Amazon is even more complicated because of how it integrated and mixed platform and non-platform businesses. How much of its market value derives from scale and scope economies associated with the online store and its enormous physical warehousing system as compared to Amazon’s “pure” platform businesses, Amazon Marketplace and Amazon Web Services (AWS)? In 2017, the online retail store accounted for about two-thirds of Amazon’s revenues and Amazon Marketplace about 17 percent. AWS accounted for less than 10 percent of revenues but 60 percent of operating profits, and more in prior years. Amazon did not always break out these numbers, which made it difficult to do historical analysis.

With these caveats, however, we can say that most platform businesses that survive to become public firms have been highly successful enterprises. This success was due at least in part to their platform strategies and business models, as well as to their strong product or service businesses.

Overview of the Following Chapters

We devote each of the following chapters to a specific theme and set of guidelines. Based on our experience and research, we believe the discussions will help managers and entrepreneurs understand how platform markets really function and how to build platform businesses viable for the long term.

Chapter 2 examines the fundamental drivers of platform markets and dynamics of a “winner-take-all” or “winner-take-most” outcome. In order to achieve a dominant market share, a company needs to master several dimensions of platform competition. First and foremost, the company needs to encourage and take advantage of network effects. We use the historical example of the telephone and Yellow Pages to illustrate how both same-side and cross-side network effects work, even without the benefit of modern digital technologies. But we also show why network effects are not enough to dominate a market. For example, sometimes users participate on multiple platforms for the same purpose, a practice called “multi-homing.” Dominant platforms usually make it difficult or unnecessary for users to multi-home. In addition, successful platform businesses reduce the impact of niche or differentiated competitors, which can further weaken market shares and network effects. And all companies need to build significant barriers to entry. We conclude this chapter with a discussion of how digital technology is impacting each of the four market drivers as well as platform competition more broadly.

Chapter 3 explains how strategies and business models differ for innovation platforms compared to transaction platforms. Both types follow the same steps to build their businesses, but they do so in different ways. Transaction and innovation platforms both need to identify key market participants and solve their distinctive chicken-or-egg problems (i.e., assess which side is more important in order to trigger interest and attract the other side). Both need to find business models that generate value and translate into revenues and profits. Both face similar governance challenges. However, the two platform types must engage different kinds of market participants, solve different launch and monetization problems, and preside over different types of ecosystems. We also discuss the trend toward hybrid strategies and the advantages they afford some of the leading platform companies. Innovation platforms can add transaction functions to help them distribute complementary products and services, as Apple, Google, Microsoft, and Salesforce have done with their app stores. And transaction platforms such as at Amazon, Facebook, Snapchat, Uber, and Airbnb can add innovation platform functions to help them add new features and services from third-party firms, with minimal in-house investment. Hybrid companies also differ in the degrees to which they connect or integrate the two types of platforms; some are highly integrated while others look more like modern-day “digital conglomerates.” Chapter 4 looks at the reality of starting a platform business and the mistakes that managers and entrepreneurs commonly make. Because of the importance of network effects, many people assume that first movers have the advantage in platform competition. Our data suggests the opposite. Being first can sometimes be an advantage in the conventional economy as well as the digital economy, but most first movers in platform markets have failed. More often, fast followers have come to dominate their markets. In all cases, though, timing and ongoing innovation are critical. Managers and entrepreneurs should hesitate to enter a market after it has already “tipped” toward one platform because network effects make it difficult to take away share from the platform leader. However, if the platform leader becomes complacent and stops innovating, or if other opportunities for differentiation and niche marketing appear, then it is possible to enter a platform business late and compete successfully.

Chapter 5 turns to the dilemma that conventional companies face. Has the digital revolution made old-economy firms the equivalent of dinosaurs, doomed to extinction? Or can traditional firms adapt to the new technology and new rules of digital competition? Clearly, the history of the personal computer and the Internet suggest that some traditional businesses that did not embrace a digital presence will be unable to adapt, such as bookstores, department stores, travel agencies, or brokerages. However, we have identified three strategies for “old dogs” who want to learn new tricks: Build, buy, or belong to an existing platform. Traditional companies can use these approaches to fend off new entrants and compete more effectively if platform challengers invade their space. To illustrate the opportunities and obstacles, we look at London’s black cabs, Walmart’s acquisition strategy, and General Electric’s attempt to build a platform for the industrial Internet of things.

Chapter 6 explores platform governance and the legal, political, and social challenges that digital platforms often encounter. Many people believe platform companies and digital technologies are, in general, good for individuals as well as the world economy. They seem to represent technological progress and efficiency as well as facilitate the global flow of information, technology, and capital. Other people argue that platforms pose a threat to how markets and societies should function because they too often restrain competition, sometimes violate the law (e.g., with regard to taxation, labor, or sectoral regulation), and may invade our privacy and abuse the data we often unknowingly provide. Our view is in between these two extremes. We see platforms as “double-edged swords,” with good and bad potential. We argue that platform companies must regulate themselves as well as build trust with users and ecosystem partners. Perhaps more importantly, we think it is possible for managers and entrepreneurs to anticipate and mitigate threats from antitrust, labor, and data privacy litigation.

Chapter 7 summarizes key points in the book and then looks to the future. How can we evaluate if an emerging technology might become an important platform over the next decade and beyond? Of course, no one has a crystal ball. Nonetheless, we can use the guidelines outlined in the previous chapters to identify platform potential and evaluate different future scenarios. We discuss several examples of ongoing platform battlegrounds, such as self-driving vehicles and how they are likely to affect ride-sharing platforms, and the “voice wars” competition around home digital assistants using artificial intelligence. Then we turn to examples of fundamental technologies that may evolve into new platform battlegrounds over the next decade or two, with their own technological, regulatory, and ethical challenges: the race to commercialize quantum computers and ongoing efforts to apply and build ecosystems around CRISPR technology for gene editing. We conclude that the age of unfettered, open platforms is largely over, and that platform businesses need to self-regulate or “curate” in order to remain socially, politically, and economically viable.