by Martha Lagace
All of the most valuable firms in the world today are platforms, starting with Apple, Microsoft, Google and Amazon. But platforms do not evolve in predictable ways, and there is a lot that managers and entrepreneurs can learn about past, present, and future platform strategies.
To shed light on the challenges and opportunities posed by digital platforms, The Business of Platforms: Strategy in the Age of Digital Competition, Innovation, and Power is a new book by Harvard Business School Professor David B. Yoffie and coauthors Michael A. Cusumano and Annabelle Gawer.
Across seven chapters the authors explain the fundamentals of platforms, different strategies and business models, common errors, and platform battlegrounds of the future that involve competing technologies and implications for organizations. There is advice for traditional firms looking to build or join platforms, as well as for entrepreneurs and startups. The authors discuss issues of power and of managing privacy, fairness, and public trust.
Martha Lagace: What trends are you seeing around platforms?
David Yoffie: The first question one must ask is: Are platforms the dominant business model of the twenty first century? Today, the world’s largest taxi company (Uber) owns no cars; the world’s largest provider of accommodations (Airbnb) owns no real estate; and the world’s largest retailer (Alibaba) owns no inventory.
Modern platform thinking has been evolving for the past 30 years. Academic and practitioner interest in the subject was initially stimulated by the explosive growth of the Microsoft Windows operating system. The real value of Windows, we learned, was not about the product, per se, but the applications written by independent software vendors. As a result, most of the early research on platforms was focused on what we call today innovation platforms.
With the emergence of Amazon, eBay, and other firms in the late 1990s and early 2000s, it was clear that a very different kind of technology platform was emerging, which we call transaction platforms. These platforms have their antecedents going back thousands of years to bazaars, but technology has enabled platforms to be globally scalable, which had never previously been possible.
While many people lump innovation and transaction platforms together, we argue that they are very different animals. Furthermore, in the last 10 years, we have also seen a new type of platform company emerge, which we call hybrids. A hybrid is a company that has both an innovation and a transaction platform operating simultaneously within the organization. In some cases they’re deeply linked. In other cases they are simply separate pieces of an organization that operate under a corporate umbrella. For example, Google Search is a classic transaction platform that connects end users to advertisers, and the Android operating system is a classic innovation platform that enables third parties to write new applications that create value for Android phones.
“While many people lump innovation and transaction platforms together, we argue that they are very different animals.”
We argue in The Business of Platforms that hybrids like this will become more important over time. Even if you’re running a pure transaction platform today, you may see the opportunity to create open interfaces—APIs, or application programming interfaces—that enable third parties to build on your platform. Even highly focused transaction platforms like Uber have opened up APIs to enable third parties to add value. This is an inevitable trend of digitization.
Lagace: In your research, you identified 209 platform failures between 1995 and 2015 and about 45 successful firms for the same period. What are drivers of success?
Yoffie: The most important driver of success is network effects. Network effects create the opportunities to scale at incredible speed and the potential for winner-take-all or winner-take most businesses. But network effects are not enough. This is a common misconception, that any business with strong network effects will produce a winner-take-all outcome. The evidence suggests that simply isn’t true. Network effects are a necessary but not sufficient condition.
Beyond network effects, successful platforms make it hard for their users to multihome. Multihome means users can participate on multiple platforms at the same time. In the old days, it was hard to use both a Windows PC and a Mac because of incompatibilities. Whenever there are switching costs, multihoming is hard, which makes it complicated for firms or individual users to switch from one platform to another.
The next criterion we identify is relatively homogenous products and a lack of identifiable niches. The more heterogeneous the market, the more fragmented it inevitably becomes, which lowers the likelihood of a winner-take-all outcome.
Lastly, the truly successful firms in a platform world tend to have meaningful supply-side scale economies. Markets with large economies of scale generally have higher barriers to entry, and they are more likely to tip.
Lagace: Since most platform launches fail, what mistakes should managers and entrepreneurs avoid?
Yoffie: We see four common problems across the data. The number one problem is how to price the product. The vast majority of platforms require subsidies on one or both sides of the platform for some period of time. Failure to get that pricing right inevitably leads to decline. You can see examples in businesses like ridesharing. The first player in the space was a company called Sidecar that never was able to figure out the right pricing to drive market demand or attract new drivers to the platform. Sidecar ended up a casualty of both Uber and Lyft’s aggressive pricing policies.
A second common mistake is the failure to build, establish, and maintain trust. Platforms by their very nature require two unfamiliar parties to do business with each other, which means they may not know each other and have no reason to trust each other. Without trust, many transactions will never materialize. A crucial feature of any platform is enabling and creating a trust environment allowing independent parties to feel comfortable that they’re not putting their business, operation, or personal wealth at risk.
A third mistake is mistiming the market. It’s possible to be too early, which is not often the big problem, but it is more problematic to be too late. This is because of the power of network effects and the power of platforms to scale very quickly. Even if a firm has a better product, if it is too slow in developing customers on each side of the platform, it can still lose out. A great example was Microsoft’s failed efforts in smartphones. Microsoft built a very good operating system for smartphones, but it could not crack the market because its product came out five years after the iPhone and four years after Android. By then, there were already hundreds of thousands of applications written for the other operating systems. Even though Microsoft may have had the best of the three operating systems, it didn’t matter because it was simply too late.
The fourth common mistake is hubris. When firms are very successful in the early stages of a platform they often think the market has tipped and they don’t need to worry about competition and new technology. They lose their paranoia. The reality is, even in markets with strong network effects, it’s possible for competitors to overturn a leader’s advantage.
Lagace: What advice do you have for conventional firms looking to explore platforms?
Yoffie: There are three potential strategies. You can belong to an existing platform, buy a platform if time to market is critical, or build one if you want to control your ecosystem.
None of these strategies are without risk. Belonging to a platform is a way to quickly participate in a platform market. The challenge is to avoid the problem of “hold up” by the platform itself—that is, how do you prevent the platform from extracting most of the value? How do you prevent it from observing what you do and then simply copying it? Nonetheless, belonging to a platform is a way to engage in a platform market quickly at relatively low cost and learn the tools of the trade. In the book, we explore a number of “belong” strategies which have delivered excellent results.
Buying a platform is a higher risk, higher cost strategy, but it accelerates engagement in platform businesses. For some companies, it may be the best way to get the talent and culture required to operate a platform. The example we use in the book is Walmart trying to compete with Amazon. Walmart largely failed in the platform retail business for 20 years. Its acquisition of Jet.com, however, let Walmart aggressively expand the top line of its platform revenue and bring in a team that understood platforms. Although still far behind Amazon, Walmart’s acquisition of Jet.com finally turned Walmart into a serious online competitor.
Lastly, the hardest problem in responding as a conventional firm is trying to build your own platform. To be honest, very few firms have been successful. A lot have tried. It is something large, established firms under the right circumstances need to put on the agenda as an option they might pursue. In the book, we discuss General Electric’s efforts to build their Predix platform. Predix was clearly going to be a multiyear, maybe multibillion dollar investment and one that was increasingly difficult for GE to afford. But it’s a great example of recognizing the potential that platforms can create even within a very traditional, industrial business. We think the basic design of Predix was heading in the right direction, but the challenges in execution have been severe.
Lagace: A central concern of The Business of Platforms is the double-edged sword nature of platforms and its impact on business.
Yoffie: Platforms really are double-edged swords. They are some of the most valuable, efficient ways to organize commerce, and they are also a potential source of violence, disinformation, antitrust abuse, worker abuse, racism, misogyny, and the list goes on. In other words, the fundamental challenge we see in platforms is that they are vehicles for good as well as evil. And the vast majority of platforms in the last 10 years were only focused on the good and not on the potential for evil.
There are two theories about how to think about platforms today. We clearly are in one camp and not the other. One theory is that digital platforms are like the public square. Anybody should be allowed to do anything: it’s a world of free speech, and let the buyer beware.
The other theory is that these are environments that reflect the values and philosophy of the companies themselves. They are not a public square, and companies therefore have roles and responsibilities to their communities. Platforms have to maintain the efficiency on one hand and reduce the potential for evil on the other. This is a controversial position.
We write about a number of different elements of this problem. One, obviously, is the antitrust problem that has become a big issue in the current US elections. We argue that once these firms become dominant, many have engaged in bullying behavior. Some of these actions were perfectly legal when the companies were small. But these same actions became unnecessary (and in some cases illegal) when the big platforms became dominant. Nonetheless, leading platform firms have been slow to adjust to nonbullying behavior. A lot of the antitrust problems could go away if they could internally think of themselves as a dominant player, rather than continue to operate as if they were entrepreneurs trying to build the business from scratch.
“Twitter, YouTube, and Facebook … are going to have to find ways to curate, ways to say certain kinds of activities are unacceptable, and if users are not happy, they should just go to another platform.”
Probably the most controversial thing that we discuss in the book is the question of whether platforms need to be curated. Curation would try to eliminate disinformation, fake news, promotion of violence, and so on. Is it up to the platform to prevent that activity from occurring? Should it be done by third parties or by members of the platform community policing themselves? Facebook, for example, has about 30,000 people working on curation of the Facebook platform. The problem is there are more than 2.5 billion users. Thirty thousand people, even with the help of AI and sophisticated big data algorithms, simply are not able to catch up.
We argue that curation is going to need to become more important. There will be costs associated with it, but ultimately it’s about maintaining and establishing trust. If the platform loses the trust of one or both sides of the market, it will disappear. And ultimately it’s the platform’s responsibility to ensure that their users can trust the platform itself.
Yes, it is a philosophical change in approach. These are companies that were built as the town square, with the philosophy of “we don’t need to worry about this misuse because the ecosystem will take care of itself.” Potentially restricting free speech is anathema to many of the users as well as many people inside the company.
It is a wrenching problem. If you look at Twitter, YouTube, and Facebook, the three biggest platforms where this has become a serious problem, all three of them are struggling to address this challenge. Our argument is that they are just postponing the inevitable. They are going to have to find ways to curate, ways to say certain kinds of activities are unacceptable, and if users are not happy, they should just go to another platform. But that does mean giving up some potential revenue, which is also difficult for a public company to do.
About the Author
Martha Lagace is a writer based in the Boston area.