Tipping the Balance

In markets where the value to consumers of a high-tech product depends on the actual and expected availability of software titles, the winning product can take all.

The drawn-out fight between high-definition videodisc formats Blu-ray and HD-DVD ended in early 2008 when Warner Bros., a big Hollywood studio, announced that it would no longer release movies on HD-DVD. Warner Bros. felt it had to make a decision because the dueling formats were holding up the market. No one wants to own a video player that can't play the movies they want to watch. As a result, customers were on the fence about which standard to buy until they could see a clear winner.

In many high-technology markets, the success of a product largely depends on the availability of software titles that complement that product. When one studio after another said that it would stop producing movies on the HD-DVD format, the HD-DVD player became virtually worthless. This "indirect network effect" also appears in markets for video game consoles, devices that play movie downloads on television sets, and computer operating systems. One of the reasons why Microsoft's operating system holds such a commanding share of the market is the large number of software applications that can run on its platform.

When people care very much about how many software titles are available, it can set off a series of reactions, from consumers to software developers to hardware manufacturers that, makes one standard "tip" or pull away from the rest of the competition. For instance, the supply of software increases in the number of consumers that have adopted the hardware, but consumers will likely adopt that standard only if it has a bigger library of software titles. Thus, a small initial advantage can grow much larger over time. Far from being the result of "bad acts" by a company, a product can grab a very large share of the market simply because of indirect network effects.

The tipping phenomenon may be easy to spot, but it can be tricky to measure. In a new study, "Tipping and Concentration in Markets with Indirect Network Effects," University of Chicago Booth School of Business professors Jean-Pierre Dubé, Günter J. Hitsch, and Pradeep Chintagunta propose a modeling technique to analyze how indirect network effects can lead to success in hardware and software markets. One of the key findings is the crucial role that expectations play: Consumers' beliefs about the future evolution of the market can determine which product will emerge the winner.

The Fight for the Edge

The market for gadgets that allow consumers to watch movies downloaded from the internet on a television set can be used to illustrate how indirect network effects lead to tipping.

Consumers have a choice between two popular platforms today: the Apple TV or Netflix's player by Roku. If Netflix starts out with a larger movie library than Apple TV, then more consumers may buy the Netflix player because they value variety. But that's not all. If more people adopt the Netflix device, then film studios will have a bigger incentive to issue more movies for that specific platform. As a result, the Netflix library grows even larger and its player will attract more customers as it approaches the tipping point. "It's a feedback loop that makes an initial advantage grow stronger over time," Hitsch says.

In this example, one standard starts out with a small advantage. But what if both platforms have the same number of movie titles to begin with? The market can still tip in favor of one device, as long as consumers believe that one of the standards will eventually surpass the other.

For instance, if consumers think that Apple's recent string of high-profile successes with the iPod and the iPhone means that the Apple brand is going to be dominant in the long run, then it would be quite reasonable for them to believe that Apple TV also will be a success. Apple TV becomes more desirable than Netflix's device because consumers expect that more titles will be available for Apple TV in the future. Movie studios soon follow and the feedback loop sets in. "It's essentially a self-fulfilling expectation," says Hitsch. In that scenario, the sheer power of consumers' expectations drives the market to tip in favor of Apple TV.

While the story works out well in theory, the challenge lies in developing an approach that can compare what a product's market share would be with and without indirect network effects. No two products are exactly alike, and one can have a larger share of the market for many reasons. Thus, simply observing actual market shares does not reveal the importance of tipping.

To solve this problem, Dubé, Hitsch, and Chintagunta built a complex model that mimics the way firms compete and how customers adopt products in a market with indirect network effects. This model allows them to "shut down" the two main avenues of indirect network effects–the current and expected future supply of software–and then to look at the predicted market shares.

What sets this study apart is the fact that it is the first study to take into account that consumers are not myopic. Indeed, people form expectations about the future, whether it is the price of a device or the availability of titles when making a decision on which standard to adopt. "It is crucial to consider that consumers are trying to predict the future evolution of the market," says Hitsch. To assume that customers are not forward-looking may lead researchers to the wrong conclusion.

Sony vs. Nintendo

To demonstrate how their model can measure tipping, the authors used data from a standard example of a market with indirect network effects: the 32/64-bit generation of video game consoles of the 1990s. Sony launched its first-generation PlayStation in 1995, and the CD-ROM console became an instant hit. Competitors Sega Saturn and Nintendo N64 also entered the market, but because of Sega's early exit, the market was left to Sony and Nintendo.

Sony's strategy from the start was to provide as many games as possible. In order to stimulate game development, Sony asked for only half of the game royalty fee that rival Nintendo charged. Nintendo, by contrast, maintained strict conditions in licensing games and insisted that it would compete on quality, not quantity. As a result, by June 1997, the N64 had just 17 game titles compared to PlayStation's 285 titles. In the end, it was clear that PlayStation's dominance was mainly due to its vast library of games. Its game proliferation strategy suggests that tipping must have played an important role. To find out how much, the authors turned to their model.

They first looked at what the model would predict if they vary the strength with which consumers care about the number of games currently available, but assumed that consumers always place a high value on the future supply of games. Consumers obviously value the number of games that they can play on their machine today, but the authors wanted to know how strong this effect should be in order for tipping to arise in this market.

The authors found that if this preference is somewhat weak, Nintendo's N64 is able to get a bigger share of the market after several time periods. However, when game variety matters very much to the consumer, the market suddenly tips in Sony's direction with PlayStation taking about 83 percent of the market. The reason for the switch is that while Sony has more software titles than Nintendo, the N64 console is priced lower than the PlayStation. This gives Nintendo the edge as long as the indirect network effect is not too strong. However, as this effect becomes more pronounced, Sony's advantage prevails.

Similar results were found when the authors vary the strength of consumers' expectations about the future supply of games. Sony ends up with a much larger share of the market when consumers display a very strong preference for the availability of lots more games in the future. Otherwise, the market will not become as concentrated.

Comparing the predicted market shares with and without indirect network effects, the authors found that the degree of tipping is relatively large: Market concentration increases by at least 23 percentage points. "Indirect network effects are a very strong determinant of success in this market," Hitsch says.

When Market Concentration Does Not Always Equal "Bad Acts"

The study provides firms with an understanding of the rewards of being first in the market and the value of changing consumers' expectations about the future of a product. After all, a slight initial advantage can potentially grow into a bigger one over time. "If firms can make consumers confident about the adoption of their standard and confident that they will be around for a long time, then that may lead to large differences in the market share that firms can gain," explains Hitsch.

The authors' approach to measuring tipping also should be of interest to policymakers who may be concerned that one firm dominating the market is bad for the consumer. Markets can become "naturally" concentrated, that is, without any violation of antitrust laws. "Looking at market shares alone need not mean much in markets with indirect network effects," Hitsch says. A market share of 80 percent may set off alarm bells. But to fully understand why this market is concentrated, one must also look closely at the importance of tipping in markets with indirect network effects.

"Tipping and Concentration in Markets with Indirect Network Effects." Jean-Pierre Dubé, Günter J. Hitsch, and Pradeep Chintagunta.