New Research Explores Multi-Sided Markets

Dating clubs, credit cards, and video games are all examples of multi-sided markets, where firms need to get two or more distinct groups of customers on the same platform. Professor Andrei Hagiu discusses this new field of business research—and why it matters to you.
by Sean Silverthorne

Andrei Hagiu is on the vanguard of a new field of business research that explores the dynamics of multi-sided markets. Although these markets have been around a long time, they have not really been recognized as entities unique from other markets; research into multi-sided markets only began a few years ago.

Most markets are one-sided in nature—customers interested in buying running shoes, for example. But a multi-sided market involves more players—and each has its own interests to be served. Multi-sided markets are also interdependent. Hagiu, an assistant professor at Harvard Business School, defines these multi-sided markets as platforms that serve two or more distinct groups of customers who value each other's participation. Examples: EBay provides markets for both buyers and sellers, dating club clients are both men and women, shopping malls serve both retailers and shoppers, and operating system companies need both end-users and application developers. Increasingly, thanks in large part to technology and the Internet, multi-sided markets are becoming more common.

But how do they work? What are effective strategies for players in these markets? Below, Hagiu discusses his research into this emerging field and its implications for executives.

Sean Silverthorne: How did you get interested in studying multi- and two-sided platforms? Is this a new area of research?

Andrei Hagiu: During my PhD program in economics at Princeton, I always worked summers as an economics consultant with National Economics Research Associates on high-profile antitrust cases: Microsoft and Visa, mainly. Economics consulting is very different from management consulting. It involves coming up with economic arguments grounded in academic theoretical or empirical work to justify or condemn certain business practices. For instance, in antitrust, it is about anticompetitive behavior.

The work I have done with NERA was quite inspiring for my thesis research, and through it I got to work closely with some of the pioneers of the two-sided markets field of research: Jean Tirole (IDEI, Toulouse and MIT), Jean-Charles Rochet (IDEI, Toulouse), and David Evans (NERA, now LECG and University College London). It is indeed a very new area: The first two or three papers on this topic were written in 2002-2003, and I was lucky enough to be working with the authors right at that time. Then I developed and pursued my own stream of research in this field, starting with my PhD thesis.

Q: What are multi-sided platforms and how do they differ from other types of markets?

A: Two- and multi-sided markets are markets in which firms need to get two or more distinct groups of customers who value each other's participation on board the same platform in order to generate any economic value. In traditional one-sided markets, firms serve different types of customers, but they lack the interdependency: For instance, hair salons can choose to serve either men or women or both. By contrast, dating clubs (the quintessential two-sided platform) must serve both men and women.

Q: Please give some examples of these markets.

A: Examples are pervasive in today's economy and range from dating clubs (men and women), financial exchanges, real estate listings, online intermediaries like eBay (buyers and sellers), ad-supported media (ad sponsors and readers/viewers), computer operating systems (application developers and users), videogame consoles (game developers and geeks), shopping malls (retailers and consumers), digital media platforms (content providers and users), and many others.

Q: Why is the study of multi-sided markets such a young science? Are they a relatively new phenomenon? What is driving their growth?

A: Multi-sided markets in general are not new at all: They have been around for ages—think of the dating clubs! However, it is not until very recently that economists have realized that there are interesting common threads that tie together markets which, on the face of it, have nothing to do with each other: credit cards, videogames, and dating clubs. It is also true that certain categories of two-sided markets have become more numerous due mostly to technological evolution: The Internet has spawned many two-sided platforms, and all the software platforms that run our computers, PDAs, and mobile phones have only emerged recently.

Q: You note in your research that in many multi-sided markets, one side of the market is charged little or nothing to participate, while all of the profits come from the other side. Can you give an example or two and explain why these markets are structured like this?

A: Dating clubs usually charge only the men, and credit card companies make their revenues mostly from merchants rather than from consumers. More interestingly, vendors of operating systems like Apple, Microsoft, Symbian, and Palm derive their profits from users through licensing fees and do not charge much to allow application developers to access their platforms. On the contrary, videogame console makers like Sony, which makes PlayStation, and Microsoft, which makes the Xbox, make their profits from game developers through royalties and incur losses on the sale of consoles to users by pricing them below cost.

The key reason is that two-sided platforms must solve a chicken-and-egg problem. For example, without sufficient applications developed for it, an operating system has no value for, and therefore cannot attract, users. Without a solid user base, no application developer will be interested in supporting that operating system. If the platform vendor decides to charge positive prices on both sides, it might end up attracting neither. The same goes for dating clubs, game consoles, and so on. So the idea is to subsidize one side in order to attract it more or less irrespectively of the other side and then turn to the second side and charge it positive prices.

Of course, depending on the market, the timing and mechanism of adoption by the two sides varies, and there are interesting differences to look at, but the fundamental chicken-and-egg issue is the same.

Q: Why is it important to understand how these markets are vertically and horizontally integrated?

A: The vertical and horizontal business scope of platforms are critical because they determine the platforms' ability to create viable ecosystems by getting the relevant sides on board, generating interactions among them, and extracting profits.

Vertical scope has to do with the decision of whether or not to integrate into hardware, complements, content, et cetera. For instance, videogame console platforms are vertically integrated into hardware, whereas most operating system vendors for PCs (Microsoft), PDAs (Palm), and mobile phones (Symbian) have preferred to rely on third-party competing hardware makers. The idea is that the latter can expand the market for the software platform much more than the platform vendor could do by itself if it remained vertically integrated into hardware.

The interesting exception here is, of course, Apple. Apple's logic is that vertical integration allows the company to better control design, innovation, and production throughout the entire vertical chain and therefore come up with superior products. One tradeoff would therefore seem to be between large market share and high quality, of which the Microsoft/Apple comparison is a good illustration.

Sony illustrates another dimension of vertical integration. It is a consumer electronics company that is vertically integrated to some extent into the content that runs on their devices. This can help in their current fight to establish the Blu-ray DVD standard, but has been a significant hindrance, because of internal conflicts of interest, in their efforts to enter the digital music media player market, which they have lost to Apple's iPod.

Two-sided platforms must solve a chicken-and-egg problem.

Horizontal integration has to do with the parallel markets in which multi-sided platforms choose to enter. Indeed, many multi-sided platforms are "imperialistic" in nature, in that they have a strong tendency to expand horizontally. Sony has expanded its PlayStation videogame platform into a digital home entertainment platform by adding DVD-playing and Internet-access capabilities; Microsoft has expanded its PC software platform to the increasing variety of mobile devices running on microprocessors; NTT DoCoMo, Japan's leading mobile phone operator and creator of the pioneering i-mode mobile Internet platform, has expanded i-mode from communications and mobile Internet content to mobile payment systems and soon to mobile television.

The idea is to always enlarge the platform ecosystem by adding new sides and platform functions that might be valuable to the existing sides and therefore create positive synergies (and dominant firms!).

Q: You draw a distinction between two-sided platforms and traditional merchants. What is that difference and why does it matter?

A: If we stick to the simplest definition of two-sided platforms—businesses that have to attract two interdependent groups of customers—then many firms would appear to be two-sided. In particular, one can think of retailers and other merchants who need to get a variety of products from suppliers before selling them to consumers.

Take Wal-Mart, for instance. Is it a two-sided platform? Not in its "classic version." If it functions like your average village merchant who wakes up early to acquire produce from suppliers and then resells it to consumers at its local store, then the transactions it conducts with members of the two groups—suppliers and consumers—are largely independent of each other and there is nothing particularly interesting in a two-sided sense.

But Wal-Mart has grown quite sophisticated. It also rents shelf space to some suppliers—Kellogg's for its cereals, Coke for its cans, and Sony for its electronic devices. These suppliers are then responsible for supplying, displaying, pricing, and advertising their merchandise within the space allocated by Wal-Mart, and they receive the revenue from sales to consumers. In this case, Wal-Mart becomes a two-sided platform for interaction between these producers and consumers.

This is not just an academic exercise: The distinction is quite important as the two models (merchant versus two-sided platform) are associated with different business strategies. One analog of the Wal-Mart in the digital world could be iTunes. Today Apple operates iTunes as a merchant that buys music rights from publishers and then sells it itself as it pleases to users. However, it is easily conceivable that at some point Apple may wish to transform iTunes into a two-sided platform, i.e., a portal through which music publishers can sell directly to users at prices they choose individually. Clearly, an iTunes portal would differ significantly from the current iTunes music and video store.

Q: In attempting to provide a general theoretic framework on multi-sided platforms, you begin by looking at the functions they perform rather than their nature. What are these functions, and what do they tell us about multi-sided markets?

A: In my opinion, the early attempts to overview two-sided platforms and markets have struggled because researchers were attempting to classify platforms by their nature. Instead, I think it is much more productive to first isolate the fundamental functions platforms perform, of which there are three: reducing search costs (this helps in matchmaking contexts: men and women, buyers and sellers, etc.); making audiences (this is essentially the function of advertising platforms); and saving on shared costs (i.e., providing an infrastructure that can be used for many transactions between the different sides of the platform). All multi-sided platforms perform at least one of these three functions and many perform more than one. Credit cards and all other payment platforms, for example, save on the shared costs of buyers having to compensate sellers for a product or service. Operating systems save the shared costs of providing the low-level functionalities that all applications can use. Shopping malls save on shared costs by pooling together many retailers in the same place and thereby also reducing search costs. EBay creates value by reducing search costs for buyers and sellers, and at the same time it saves on shared costs by offering PayPal as a convenient payment platform. And in performing these services, it is also creating audiences for advertisers.

Of course, there are intermediaries other than multi-sided platforms that perform similar functions. For example, retailers save the shared costs of suppliers individually selling to consumers and on the search costs of consumers looking for a given product. Retailers are usually not two-sided platforms: They are merchants who simply buy and resell products; they generally do not have to solve a chicken-and-egg problem at any moment.

With that caveat in mind, I think that understanding the fundamental functions multi-sided platforms can perform dictates competitive advantage in multi-sided markets.

Q: For the business practitioner, what are the practical aspects of your research?

A: There are many important implications for practitioners. Summarizing just a few based on the above:

  • The pricing structure—that is, how much to charge to one side relative to others—matters; and in order to determine the optimal pricing structure one needs to carefully analyze the relative interdependencies among the multiple sides as well as their willingness to pay and join the platform. More often than not, one has to subsidize the participation of one side, i.e., forego profits there, in order to derive them from other sides.
  • The scope of the platform is also a key decision variable: More than determining competitive advantage under an exogenously given industry structure, many platforms are in a position to change the structure of their industry by their decision of how much vertical and/or horizontal integration to have. Vertical disintegration and licensing can help expand the market, but vertical integration allows for tighter control and perhaps higher profit-extraction power. Horizontal expansion can bring in new sides and create more sources of profits by unleashing new virtuous circles of value; but at the same time they bring the platform in competition with new competing platforms from previously separate industries.
  • Managers should be clearly aware of the fundamental economic functions performed by their platforms—these functions are the main sources of competitive advantage—in order to be able to make the best decisions on what new functionalities and services to build and enhance. They should always look first for those functionalities/services that enhance and enrich the core functions they already perform as they offer the highest synergies and payoffs, rather than adding functionalities related to fundamental functions they do not yet have, which could prove distracting and counterproductive.

Q: What research are you working on now?

A: I have just finished a book on software platforms, Invisible Engines: How Software Platforms Drive Innovation and Transform Industries, with David S. Evans and MIT Sloan School of Management Dean Richard Schmalensee, which will come out from MIT Press next fall. I am working on several papers attempting to provide theoretical frameworks for studying some of the issues mentioned above, such as platform scope, fundamental functions performed by platforms, and two-sided platforms versus merchants, as well as three related HBS cases.