“I’m Free”: Platforms and Antitrust Enforcement in the Zero-Price Economy
Good afternoon. Thank you Phil for that kind introduction, and also thanks to you and the Silicon Flatirons for inviting me back to speak at this impressive conference.
At the outset, let me congratulate you on your election as Attorney General of Colorado. I very much look forward to working with you on the issue we both love, protecting free-market competition on behalf of American consumers and entrepreneurs.
I have thoroughly enjoyed the presentations and exchanges yesterday and today. The growing attention on the policy challenges posed by today’s large tech platforms, as this conference demonstrates, often comes up within a broader debate on the role of antitrust in today’s digital economy. I’ve touched on a number of topics in this area in past speeches, including innovation, data, privacy, the consumer welfare standard, and industry concentration. I try to do my best not to repeat past comments, and today I will focus on a related issue that is relevant to both platforms and the digital economy: how antitrust enforcers should think about “free” – or more accurately “zero-price” – products and services.
The title of this address is “‘I’m Free’: Platforms and Antitrust Enforcement in the Zero-Price Economy.” “I’m Free” is a popular song by the Rolling Stones, written in 1965, just a few years before the DARPA project by visionaries Bob Taylor, Vint Cerf, and Robert Kahn, who actually invented what we know today as the Internet. The song tells us that we’re free to do what we want any old time and free to choose what we please any old time. In today’s digital economy, consumers too can choose and do an increasing number of things for “free”, at any old time. We can search online, connect via social media, play games, write emails, reads the news, listen to music, make reservations, get directions, track our fitness, and a whole host of other activities all without paying a penny.
Yet are these services really “free”? As an aphorism often associated with Milton Friedman goes, “there’s no such thing as a free lunch.” Most firms that provide goods at a price of zero are making money somewhere else, either through different products, different consumers, or at a different point in time. Consumers also typically exchange something of value, such as their attention to advertising or their personal or usage data, for these free services. So when we talk about digital platforms providing “free” services, we are really talking about business strategies where zero is the chosen profit-maximizing price. Setting a zero price ultimately can make those firms more money, for example, by incentivizing balanced participation in a platform, by driving up demand for a complementary product, or by helping a firm break into a new market.
Zero-price strategies have exploded in the digital economy, driven in large part by the Internet’s decreased production and distribution costs and the increase of digital platforms characterized by network effects and economies of scale. We should remember, however, that the strategy of selling a product or service at zero price is not new, nor is it unique to the digital economy.
Let’s start by taking a look back.
Radio, broadcast television and newspapers, all forms of ad-supported media, existed long before the rise of the digital economy, with radio dating back to the 1920s. Audiences have long enjoyed many of these products and services for free, funded by advertisers.
Shopping malls are another early example of a multi-sided platform with a zero-price strategy. Customers can enter and roam around a mall for free, while merchants pay to lease space. Many general payment cards, starting with the Diners Club in the 1950s, have long provided cards for free to consumers, while charging merchants a fee for transactions.
Not only do zero-price strategies predate the digital economy, they also are not all the same. There are several types of zero-price strategies, which vary depending on the industry and context. These differences are important to antitrust analysis. Determining how a firm profits, either today or in the future, from zero-price strategies can shed light on whether conduct in these contexts could raise competitive concerns.
While the antitrust laws are not subjective, looking at a firm’s motivation, in some cases, can be a helpful starting point. Why has a firm chosen a zero-price strategy? Is it a predation attempt to exclude competition and ultimately raise prices? Or, is it a strategy to achieve optimal and balanced participation on a platform? Or, is it a temporary effort to break into a new market and increase competition?
Many multi-sided platforms, especially in the digital economy, follow a zero-price strategy. Ad-supported media is a leading example. Platforms provide content or other services for free to consumers, and make money by selling the attention or information of those consumers to advertisers through ad networks. As such, the users actually become the “input” of the product ultimately sold to advertisers. To attract additional eyeballs to sell to advertisers, some content platforms choose to set the price to view content at zero. Platforms use this strategy in a variety of contexts both online and in traditional industries, such as online search, social media, email, apps, broadcast television, and radio.
Exchange platforms that connect groups of “buyers” and “sellers,” like eBay or OpenTable, also use zero-price strategies. In order to strike the right balance of buyers and sellers, platforms may decide to charge only one side of the exchange for its services. Traditional shopping malls, as well as digital marketplaces like eBay, Etsy, and Amazon, typically are free for consumers to use, but charge merchants to participate. Restaurant reservation apps like OpenTable and Resy provide another example, allowing diners to book tables for free while charging restaurants. Different business models are experimenting specifically in this restaurant reservations industry, and the proverbial jury is still out as to which model or models will ultimately succeed.
Transaction systems, such as payment cards, also must balance user and merchant participation. As the Supreme Court noted in the recent Ohio v. American Express case, “[t]o optimize sales, the [credit card] network must find the balance of pricing that encourages the greatest number of matches between cardholders and merchants.” In some cases, that chosen optimal balance leads to setting the consumer price at zero, or even below zero with loyalty points and free loans, with merchants paying a fee on each transaction.
Another zero-price strategy is to tie or bundle a “free” product with a complementary or related positive-priced one. This strategy can be profit-maximizing if the offer of a free product increases demand for the positive-price good.
For example, cell phone service providers may offer a free phone with the purchase of a service plan. Another example is Gillette’s strategy of providing some consumers with a free razor, hoping it will create brand loyalty and lead to future razor blade purchases.
Some businesses provide a basic version of their product or service for free and a higher quality version for a positive price. This “premium upgrade” or “freemium” strategy attracts consumers to test out the free product, with the hope that some consumers will upgrade to the paid version. The sale of the premium product, often to a small fraction of the users, subsidizes the provision of the “free” service to everyone else. Adobe software and Dropbox are examples.
Firms also use a hybrid strategy, offering a “free” ad-supported version and a positive-priced premium version. Music streaming services like Pandora and Spotify are examples. Amazon’s new IMDb Freedive service is another example of such an offering that gives the consumer the choice to watch premium movies with ads inserted rather than renting it on Amazon Prime or on Apple iTunes.
Finally, businesses may temporarily offer zero-price products or services to promote their brand or break into a new market. In such cases, a firm will lose money in the short turn in order to attract customers with the hope to turn a profit in the long run.
Walking by a cosmetics counter in a department store, you are likely to be inundated with free perfume or cologne samples. Vineyards often provide free wine tastings to visitors in the hope you’ll like what you try and buy a bottle or, even better, a case. The same is true with services. Summer interns often will work for free in the hope of securing permanent employment or to develop marketable skills.
In the digital space, platforms also may offer a new product for free during its beta-testing phase to help measure consumer demand and improve the product before its official launch. Some of us may remember the AOL discs sent in the mail for 50 or 100 hours of internet access, a foreign concept, I am sure, to the students in the audience.
A less benign zero-price strategy is to provide a product for free in order to drive out rivals from a marketplace and then recoup the cost through monopolistic pricing at a future date. Predatory strategies, however, tend to be difficult to undertake in practice. Once consumers expect products or services to be given away for free, it becomes hard to impose even a small positive price. For example, one study, which we have not scrutinized, estimated that 84-87% of respondents would leave Facebook if it started charging $1/month. We should also distinguish between penetration pricing, which can enable a new entrant to challenge incumbents, and is the heart of competition, and predatory pricing.
Consumers often benefit from these different types of zero-price business models I’ve just described. Zero-price strategies bring goods and services to consumers who otherwise would be priced-out. Consumers can exchange their data or attention in lieu of money to receive valuable services. Studies demonstrate that many consumers prefer ads to paying for certain services. In the study I just mentioned, 79% of respondents represented they would choose an ad-supported Facebook over paying $1/month. Another study found that consumers almost always choose free and ad-supported apps over the 99 cent alternative without ads. For now, consumers seem to be willing to provide information about themselves so that they receive the product for free, even recognizing that their information will be sold to advertisers in exchange for a zero price. Some consumers even may prefer receiving targeted advertisements that are more likely to be relevant to their needs.
Zero-price strategies also may enable new entrants to break into markets, increasing competition and consumer choice. Firms in zero-price markets often compete on quality and innovation, which can benefit consumers.
Competition on non-price factors often leads to product differentiation (or hybrid strategies) to appeal to different consumer preferences. For example, consumers may choose between radio stations based on the amount of advertising or better content quality, or choose online search services based on more accurate results or greater privacy protections. Consumers could benefit from a free market that gives them the ability to choose between these different business models. Of course, these choices can lead to new so-called revealed preferences, should consumers begin to value their privacy, for example, differently. One goal of antitrust enforcement should be to ensure the markets are free for competition by such new business models.
In addition to their many benefits, zero-price strategies also pose challenges for antitrust enforcement.
In the absence of price competition, market definition can be difficult. The traditional analytical test applied by enforcers to define relevant markets, which looks at small but significant and non-transitory increases in price (or “SSNIP”), does not translate directly to a zero-price market. We cannot look at the effects of a five percent increase in price because five percent of zero is still zero. Choosing variables for measuring market shares also can be more complicated where shares of revenue is not an option.
In light of these challenges, and the increased prominence of zero-price strategies in the digital economy, many currently are debating how antitrust enforcement should treat such products and services, especially when offered by a large digital platform.
On one end of the spectrum, some argue that zero-price products and services should be exempt from antitrust scrutiny. They argue that consumers and competition cannot be harmed if users are getting a product for free. An argument Microsoft used, without success, in defending the DOJ’s antitrust action in 1998. Some argue that the benefits of free products and the complexity of the antitrust analysis should lead us to forgo antitrust enforcement in this area entirely.
On the other end of the spectrum, commentators and some foreign antitrust enforcers call for more aggressive enforcement efforts against digital platforms that provide free services. These commentators and enforcers consider the collection of consumer data in exchange for goods to be potentially anticompetitive. Some even propose throwing out the long-standing consumer welfare standard and crafting new rules to address these businesses.
The long history of zero-price strategies teaches us, however, that both of these extreme views are misplaced.
First, we should not exempt zero-price models from antitrust scrutiny and give a free pass to free services. U.S. antitrust laws apply in full to zero-priced products and services. Traditional conduct that is unlawful under the antitrust laws is still unlawful in the zero-price models more prevalent in today’s digital economy.
For example, agreements among competitors to allocate markets or customers are per se unlawful, even in the context of zero-price goods. Naked horizontal agreements to fix levels of non-price competition, such as quality, advertising, privacy protection, similarly raise antitrust concerns. Mergers-to-monopoly and certain tying arrangements could also violate the antitrust laws in cases involving zero-price strategies.
In fact, the Antitrust Division has reviewed conduct and mergers in zero-price products before, such as broadcast television, and we can continue to do so. The argument for antitrust immunity ignores that these firms still compete on non-price dimensions and that zero-price products are often intertwined with other positive-price products and services. As such, firms with zero-price strategies can still engage in anticompetitive conduct, and antitrust enforcers must stand ready to protect competition and consumers.
Indeed, public antitrust enforcement may need to play an even greater role in zero-price settings where the absence of price to the end consumers could make private damages recovery difficult, or where effects on business partners, such as advertisers who must rely on an ongoing relationship, may impede the incentives for private antitrust actions.
Second, our long history with zero-price strategies also tells us that we do not need a wholesale revision of the antitrust laws to address competitive concerns in these contexts. As I’ve said before, our antitrust laws and principles are flexible enough to adapt to the challenges of the digital economy. For this reason, we should not rush to throw out the consumer welfare standard in the context of the zero-price economy. The consumer welfare standard is not limited to looking at price effects. It also takes into account effects on quantity, quality, consumer choice, and innovation. With zero-price goods, it simply becomes more important to focus on these other non-price factors, or to focus on the actual consumers who really pay for the service.
As in all enforcement matters, we should engage in careful case-by-case analyses in which we look closely at the evidence. When appropriate, we can tailor our approach to consider unique characteristics of a zero-price market. The fact that market definition and other issues can be more challenging in the absence of price competition does not mean we should give up on our rigorous, evidence-based approach.
As I mentioned, one of the challenges for antitrust enforcement in the zero-price economy is that it often can be harder to define the relevant antitrust market. Traditional economic tools that rely on measurements of price are not available for the zero-price context. What, then, should we do in defining markets and assessing market power in the zero-price economy?
As a starting point, we must not fall into the trap of believing that, because there is no price, there is no market or market power. At the same time, we also must avoid the temptation of defining markets with subjective definitions that shortcut rigorous analysis. For example, we should not simply apply labels to a market without looking at the real-world competitive dynamics. Nor should we simplify the analysis by defining a market by the identity of the platform providing the product.
Industries with zero-price strategies and multi-sided platforms are often sufficiently complex that a mechanical market definition is likely to obscure competitive realities. We cannot define a relevant market or market power in a vacuum. We need to look at facts of the specific case and the effects on competition.
Second, for zero-price products and services, a proper antitrust analysis often requires looking more broadly than the free product itself. Remember, for-profit firms that use zero price strategies must be making money from somewhere. Otherwise, they would eventually go out of business. Remember that CBS or ABC that provide free over the air broadcast content do not do so without making a profit. The existence of a free product usually indicates that there is a related positive-priced product and that the economics of those two goods are related. A proper antitrust analysis, in most cases, should consider the free product together with its companion money-making product.
The relationship between free and paid products in the zero-price economy also informs whom we treat as the relevant consumer.
There’s a saying that “If you’re not paying for it, you’re the product.” Users of a zero-price product, in some cases, simply may be the “input” to a final product sold to a different set of consumers. For example, broadcast TV providers use audiences as an input to a product ultimately sold to advertisers. This does not mean we ignore alleged anticompetitive effects for TV viewers, but it does mean we may need to consider effects on advertisers and advertising prices in our holistic analysis.
Third, in the context of zero-price goods, it becomes particularly important to look at non-price competition. Price is only one dimension of competition. Firms in zero-price markets presumably continue to compete against each other on quality, choice, and innovation. These competitive factors are part of our traditional consumer welfare standard and deserve our attention and perhaps renewed focus.
Fourth, it’s helpful to take a step back and think about the purpose of market definition. Market definition is not an end to itself. The purpose of defining a relevant market and assessing market power is to identify competitive constraints that limit a firm’s ability to engage in behavior that harms competition and consumers. Where firms are limited by competitive constraints, the free market can self-correct without the need for antitrust enforcement. When there is such competition, consumers can switch away from firms engaging in behavior that raises price, degrades quality, and harms their experience.
A recent Harvard Business Review article observed a number of constraints on a digital platform’s success, even in markets characterized by network effects. The article first points out that the strength of network effects can vary dramatically and can change over time. While network effects often lead to competition “for the market” rather than “in the market,” they do not necessarily create durable market power.
In the video game industry, for example, companies have alternated taking the lead over the years depending on who has the newest technical advantage and hit games. In the 1990s, Windows was a leading operating system for application developers seemingly entrenched by the strength of its network effects, but technological developments, allowed to compete by the Antitrust Division’s enforcement activity, and shifting consumer preferences helped Android and iOS mobile operating systems to provide new choices and new markets for the consumer.
Vulnerability to “multi-homing,” when a user forms ties with multiple platforms at once, is another constraint on platforms. Multi-homing may constrain the ability for a platform to engage in anticompetitive behavior because consumers can easily shift consumption to the other platform. All else equal, multi-homing occurs more often when the cost of adopting an additional platform is low, and is especially common in zero-price markets. Take Uber and Lyft, for example. It is essentially costless for a user to download the Uber and Lyft applications on a phone, and passengers can easily switch between services depending on which one is cheaper, as long as there are enough drivers on the other side of the platform.
Multi-homing occurs in zero-price markets because they are often highly differentiated. Because platforms do not compete on price, they differentiate themselves from each other by choosing particular levels of quality or by offering different features. Consumers may find certain features or content of different platforms attractive, and therefore join multiple platforms. Advertisers also may rely on many differentiated platforms to distribute their ad campaigns.
As these and other examples demonstrate, there are many different factors that play into a platform’s success or failure. We, as antitrust enforcers, need to pay close attention to competitive constraints, to first make sure we don’t enforce away pro-competitive business models, but also to be vigilant that conduct does not unreasonably constrain the ability of new competitors to compete on the merits.
The title of this conference is “Internet Platforms’ Rising Dominance, Evolving Governance,” and we have been asked, among other things, to “evaluate the appropriate strategies for overseeing platforms.” In answering that question from the perspective of antitrust enforcement, we should remember that market power is not unlawful on its own. The question for antitrust enforcement is not how big the platform is, but whether what the platform is doing harms competition.
Some critics say that platforms are too big, have too much data, or too many users. We must be vigilant to study the causes and effects of market power. In doing so, we must remind ourselves, however, that an efficient firm capturing unsatisfied customers from an inefficient rival or incumbent is, as the Supreme Court explained, “precisely the sort of competition that promotes the consumer interests that the Sherman Act aims to foster.” The benefit of our capitalist free market system is that winners and losers are determined by the market, not chosen by the government.
Many of today’s large digital platforms have grown because they provide innovative and disruptive services that consumers seem to like and want to use. In my view, it’s no accident, and a point of pride, that a majority of these leading platforms are American companies. Our pro-market economic and legal structures and our venture capital community foster innovation and entrepreneurship. The fact that successful companies can reap the benefits of their hard work encourages the next generation of innovators and entrepreneurs and inserts the dynamic competition that best benefits consumers. Therefore, as the Supreme Court explained in Trinko, “[t]o safeguard the incentive to innovate, the possession of monopoly power will not be found unlawful unless it is accompanied by an element of anticompetitive conduct.”
Moreover, antitrust law is concerned about consumer welfare, but only where consumer harm results from the reduction of competition, which distorts the free market. We also are wise as enforcers and policymakers to remember that, as the Supreme Court explained, “[t]he purpose of the [Sherman] Act is not to protect businesses from the working of the market; it is to protect the public from the failure of the market.” Concerns over privacy, inadequate notice, unauthorized use of data, and data protection, are legitimate policy issues that need to be discussed, but should not lead to distortions of our antitrust standards to address them. Fortunately, our antitrust laws are flexible enough to adapt to digital platforms, should evidence show that their practices harm competition.
The digital economy and the increasing number of zero-price strategies have brought numerous benefits to consumers. Our role as antitrust enforcers is to protect the free market, in which consumers are, as The Rolling Stones say, “free to choose what [product or service] they please any old time.” We do this ideally by guarding against collusive and exclusionary behavior that distorts the free market.