Back in the 1990s, there was a war being fought for the hearts, minds, and desktops of computer users around the world. Apple had thrust the personal computer into the modern age with the original Macintosh in 1984, introducing users to the now-familiar mouse-driven interface. But the company quickly lost ground to Microsoft. Ten years later, Microsoft solidified that dominance with Windows 95. Two years later, when the United States Department of Justice filed an antitrust complaint against the software firm, the war was already over. The Windows platform had won.

When two platforms—like Windows and Mac OS—are in competition with each other, it seems inevitable that one will win. Yet that is not always the case. Consumers have many different choices when it comes to credit card platforms, including Visa, MasterCard, American Express, and others. Fans of video games currently have three choices when it comes to consoles—Nintendo’s Wii, Microsoft’s Xbox 360, and Sony’s Playstation 3. And people seeking dates have many, many sites to visit on the Internet. In each of these three industries, consumers have enjoyed a choice of platforms for many years. Unlike what happened with Windows and Mac OS, none of those markets has tipped to favor one over the others.

Signs of whether and when a market may tip are closely watched by both economists and regulators. Yet as the PC wars showed, there may be little warning that a market might tip. In fact, tipping may only be obvious when it has already happened. By the same token, no one is certain why markets such as online dating have not tipped. That is in part because tipping is little understood.

“Tipping is really hard to study in the real world,” says Dylan Minor, an assistant professor of managerial economics and decision sciences at the Kellogg School of Management. “We just get a snapshot in time. We don’t know for sure if a market is tipping or actually coexisting.”

To Tip or Not to Tip 

To study the moment when a market tips, Minor and his colleagues Tanjim Hossain, a professor at the University of Toronto, and John Morgan, a professor at the University of California, Berkeley, moved to the laboratory, where they could control conditions that might cause tipping. “A unique feature of the lab is that we could actually simulate and recreate competing market platforms to see when they tip and why,” Minor says.

Even when platforms and their payoffs were identical—which should have encouraged people to stick with their first choice—players still converged on one platform.

In the lab, they first constructed a number of different markets in which participants could choose one platform over another. Four (or eight) players participated in a session and played four sets of fifteen periods each. Each player was assigned to be one of two types, a triangle or a square, and two or four players of each type constituted a market. In each period, the players selected a platform based on their type and the respective payoff each platform offered (e.g., one platform may have offered lower fees and a higher payoff for certain triangles, luring those players to it). Minor and his colleagues adjusted the markets such that half would tip and half would coexist.

Oddly, nearly all of them tipped. Even when platforms and their payoffs were identical—which should have encouraged people to stick with their first choice—players still converged on one platform. “Tipping was so pervasive,” Minor says. “We were very puzzled.” Whatever fosters coexistence seemed to be missing from their experimental markets.

Back to the Lab 

So Minor and his colleagues ran another set of experiments, this time adding elements of horizontal differentiation. Horizontal differentiation is the term for a situation in which one platform appeals to one group more than another. In the lab, Minor and his colleagues rigged the markets so that one platform would pay one set of player types more than another. Yet even then, coexistence was not guaranteed. Tipping only subsided when the type-specific payoffs substantially outweighed the benefits of the higher-quality platform. In economic terms, horizontal differentiation in those markets dominated vertical differentiation.

In the case of Windows vs. Mac, Apple’s vertical differentiation was not enough to maintain a two-platform market. For the majority of users, Macs did not offer a substantially better experience than Windows, and so the operating system market tipped.

In the online dating world, however, coexistence is the norm. “Dating sites are a good example of a very fragmented market,” Minor notes. “But they also have different kinds of dating sites geared toward a different kind of people, as opposed to some other markets, like say an operating system.” So while has the largest market share, it controls less than 20 percent of the market. Broad appeal may be good for operating systems, but not necessarily for dating sites. Niche players like—which caters exclusively to Jewish people—thrive in the dating world by attracting certain people who may be turned off by large sites like or eHarmony.

Online dating is not the only market where competitors coexist. As mentioned earlier, the market for video game consoles is currently split relatively evenly among Nintendo, Microsoft, and Sony. Job market postings are divvied up almost equally among CareerBuilder, Yahoo!, and Monster. Coexistence is persisting in the smartphone market, too, where many observers have been predicting the eventual dominance of one platform over another. Years ago it was Blackberry, with its firm grasp on the business market. Then Apple introduced the iPhone. Though the iPhone was not and is still not available on all networks, Blackberry’s market share waned. Google has also joined the fray with its Android operating system. To date, no platform seems to be dominating. Coexistence in the smartphone market seems assured for the near future, at least.

But in markets where tipping seems inevitable—or even beneficial if the efficiencies of one platform are large enough—regulators keep a close watch. “Indeed, regulators are concerned” about the effects of tipping, Minor says. Where coexistence is possible or beneficial, they may act to preserve competition. But in cases like Windows, where one firm already has control of the market, regulators may act to protect consumers from potential abuse. “That’s where the attention of the regulator comes in,” Minor says. “That’s a big reason why tipping is studied by economists.”

While Minor and his colleagues’ discovery that horizontal differentiation can prevent tipping may not help regulators prevent monopolies from forming, it could help potential market entrants determine the best way to succeed in a tight marketplace. “A new platform entrant is most likely to succeed in a platform space serving heterogeneous customers,” Minor advises.

Related reading on Kellogg Insight

The Superfluousness of Realtors: Homes sold through Realtors don’t garner a price premium over ones sold by owners

Increasing Revenue from Online Auctions: Buyer-seller interactions affect customer value in two-sided markets

What has the Internet Done for the Economy? The puzzling spread of the commercial Internet could explain wage inequalities