The following is a guest post from David Balto, former Federal Trade Commission Policy Director. Mr. Balto represented SunGard Data Systems in the US v. SunGard case described in this post.
Antitrust merger enforcement is a unique area of the law. It requires an assessment of whether a merger carries the potential of significantly harming competition. Courts are not very good at predicting the future and justifiably are very reluctant to prevent or unwind an acquisition without strong evidence of likely anticompetitive effects. Appropriately the antitrust enforcers rarely turn to the courts to try to stop business conduct that is typically procompetitive.
This cautionary approach is particularly necessary in software and other high tech markets. Antitrust analysis works best in traditional products, such as industrial products, that have existed for years in which the characteristics of products and the dimensions of competition are well defined. But in software the products are rapidly evolving, demand is ever changing, and the nature of competition can change overnight. Today's so called dominant firm may find itself an afterthought as the market turns to a whole different set of solutions. Not surprisingly, in the past decade the two litigated challenges to high tech mergers, Oracle's acquisition of PeopleSoft and SunGard's acquisition of Comdisco, resulted in stunning defeats for the Department of Justice's Antitrust Division.
That is why many observers were puzzled when the Antitrust Division sued to unwind the merger
between Bazaarvoice, a social software and data analytics company, and PowerReviews, a small provider of online reviews that had less than $12 million
in total revenues at the time of the transaction. (No one can seem to recall anytime the Division has sued to block a merger of a firm with an amount as paltry as $12 million in revenue). The merger involves the exciting software for providing ratings for products on the Internet, a product that did not exist a few years ago. Although the Division seems to highlight some documents that seem to suggest potential anticompetitive effects, the wooden analysis of the complaint reflects a simple structural view that overlooks the many dimensions of competition and the dynamic nature of the market. Rather than fully probing the likely competitive effects and dynamism of the online retail industry, the Division describes markets, consumer choices, and entry conditions that do not reflect reality. As a result the complaint is plagued by internal inconsistencies and fails to recognize the true price constraints that mitigate the potential for any harm the DOJ predicts as a result of this transaction.
The DOJ Fails to Articulate a Proper Relevant Market
Antitrust analysis may sound daunting, but it is very straightforward. The lodestar in any antitrust case is to define the relevant market – that is to determine what are the products that effectively compete with one another. In a merger challenge, if the government does not properly define the relevant market then the case is over. Defining the market can be very challenging, especially in dynamic markets such as software. Not surprisingly, the government's defeats in challenges to software mergers have typically been because they did not define the relevant market properly.
The DOJ defines the relevant market as "product ratings and review platforms," or "PRR platforms," and explains that these platforms "collect and display consumer-generated product ratings and reviews online." It is axiomatic that defining a relevant market establishes the boundary between products that do compete and those that do not, and determines the firms or products that constrain the relevant firm’s exercise of market power. As the Ninth Circuit has opined
, "A relevant market is identified by considering commodities reasonably interchangeable by consumers for the same purposes. Put another way, the relevant market includes all sellers or producers who have actual or potential ability to deprive each other of significant levels of business." As the literature makes clear
, if a putative relevant market is too narrow, and does not account for competitive forces that serve as a real price constraint on the parties, then the analysis risks condemning perfectly legitimate and competitive behavior by imputing market power where it does not exist.
On this count in Bazaarvoice, the DOJ does not get to first base. The DOJ's alleged PRR platform market is too narrow and falls prey to the mischaracterization of market power risk embodied in the literature. PRR platforms are one of many social-technological tools that retailers and manufacturers use to communicate with end-user customers. Bazaarvoice and PowerReviews compete against numerous firms that strive to empower the consumer's voice through social media to "collect, organize, and display consumer-generated product ratings and reviews online." Manufacturers may use popular network-driven social media tools such as Facebook, Twitter, Google+ and Yelp or more nuanced social media tools such as YouTube, Pinterest, and LinkedIn to give the consumer a voice in online product reviews. Alternatively, manufacturers and retailers can include social media tools that are similar, but not identical, to consumer reviews such as question-and-answer, and community forums.
These alternative platforms constrain Bazaarvoice, PowerReviews, and other companies that provide third-party review and aggregating services. When a retailer or manufacturer considers purchasing services from these companies, they do not look only at these two options, but also at the available outlets for consumer review generation provided by the ever-increasing array of social media platforms.
In defining markets, the courts rely on a wide variety of evidence including econometric price studies, other pricing evidence, win/loss data, and testimony of customers. None of this is present in the DOJ complaint. Instead the DOJ relies largely on the defendant's documents, but this is a thin reed indeed. Many of these documents are outdated and ignore the realities inherent in this fast-moving industry. It is debatable whether these documents reflected the true nature of competition at the time they were created. It is certain, however, that these documents no longer reflect the current state of competition. For instance, the DOJ twice references an April 2011 email in which a Bazaarvoice executive characterized the nature of the industry and opined that alternatives to Bazaarvoice are "scarce" and "low-quality." In the intensely rapid changing world of the Internet, documents from 2011 are about as relevant as a floppy disk. Instead of 1) demonstrating that this was in-fact true in 2011, and 2) reestablishing that this description remains accurate, the DOJ's complaint merely assumes that both are the case. However, this description does not align with today's online retail industry or its intersection with social media. These are industries highlighted by dynamism, and it would be incorrect to believe that the relationships between PRR platforms and other social media outlets for consumer reviews have remained stagnant.
Even If One Accepts the DOJ's Relevant Product Market, the DOJ Fails to Recognize the Dynamic Nature of Competition
As explained above, a product market of PRR platforms is not a proper relevant market for antitrust purposes. Even assuming arguendo
that PRR platforms constitute a proper market, however, the DOJ's complaint fails to discuss adequately the nature of competition within these parameters.
Two fatal flaws plague the DOJ's analysis. First, the DOJ fails to offer any explanation for portraying the PRR platform market as consisting of just two meaningful competitors and numerous fringe competitors who offer no real constraint. The DOJ attempts to justify this portrayal by analyzing the nature of competition between Bazaarvoice and PowerReviews within the "Internet Retailer 500," but it is unclear why the government focuses so closely on this tiny market segment and it is even less clear whether an impact solely on that segment would violate the law. The DOJ even concedes that the PRR platform industry "can range from simple software solutions a company has developed with internal resources to sophisticated commercial platforms offering a combination of software, moderation services, and data analytics tools." However, despite this wide range of styles and services, the complaint analyzes only a small segment of the market and suggests that these large sophisticated companies have only two effective alternatives.
The Division has made this mistake in the past, and it did not end well. In Oracle
the DOJ "failed to prove that there are a significant number of customers (the 'node') who regard Oracle and PeopleSoft as their first and second choices." Instead, the DOJ tried to make the unilateral effects argument with the unpersuasive facts that it had. Judge Walker admonished the DOJ, stating the "Plaintiffs' attempt to show localized competition based upon customer and expert testimony was flawed and unreliable. Moreover, plaintiffs' evidence was devoid of any thorough econometric analysis such as diversion ratios showing recapture effects." (A "diversion ratio" shows how much of one competitor's business will shift to another competitor if there is a price increase.)
The complaint against Bazaarvoice is equally flawed. The diversion ratios will simply not tell a story wherein a sizeable portion of all participants in the DOJ's (already flawed) market perceive only Bazaarvoice and PowerReviews as next-best options. If the evidence of diversion ratios were available, the DOJ would have presented it already. In fact, this is a consummated merger – the real-life data should show this effect if it is true. Instead, the data likely tells a story of widespread, dissimilar, and largely unpredictable cross-elasticity of demand. It is probably the case that no "node" in the PRR platform industry exists because the dynamic nature and subjective needs of clients dictate that there is no significant captive set of consumers choosing only between Bazaarvoice and PowerReviews.
Second, the DOJ completely ignores the concept of self-help in the social media consumer reviews industry. At its core, the products supplied by Bazaarvoice and PowerReviews are based on simple technology. These companies create software that appears on a retailer's website and enables consumers to provide first-hand product reviews. The companies also provide differing analytic and syndication services, both of which are a function of nothing more than intelligent use of data. There is nothing stopping retailers and/or manufacturers from creating the same service and extracting value from the data. Unsurprisingly, companies often perform some or all of these tasks themselves. Amazon
stands out as a leader in providing consumer review platforms and uses the data to drive marketing and sale decisions. Zappos
, the online shoe and apparel company from Henderson, Nevada provides its own consumer review platform on its website, and uses this information not only to improve sales and marketing, but to provide an added level of consumer care.
Like the question of the consumer "node," the DOJ has also failed to account for internal solutions as a price constraint. Once again the DOJ is forgetting an important lesson from a past defeat. In SunGard
, the DOJ tried to block the merger of two firms that provided computer disaster recovery services, which sounded like tremendously sophisticated and complex services. But the court found that self-help ("internal hotsite solutions") was a perfectly adequate option for many customers. The DOJ had portrayed the notion of internal hot sites as expensive and difficult to create, and suggested that not enough customers would turn to internal solutions to prevent the merging parties from raising prices. Judge Huvelle disagreed, and pointed out that, not only did internal solutions exist in some capacity, but that the incentive to create internal solutions would increase alongside any increase in price. Furthermore, the evidence demonstrated that customers had varying needs, and "any generalizations regarding customer behavior cannot be arrived at with any certainty, since it depends on a host of factors, including the type of equipment a customer must duplicate, the particular circumstances and needs of the customer, and in some cases, the size of the customer's operations."
The same can be said for customers of social media consumer review -- any attempt to predict the future needs and behaviors of customers is nothing more than generalization and speculation based upon incomplete data, an uncertain technological future, and dynamic and varied customer needs.
The DOJ Fails to Account Adequately for Entry and Expansion, Both of Which are Likely
The DOJ asserts that anticompetitive harm resulting from this transaction will not be corrected by additional competitors entering the market or existing participants expanding. The rationale for this assertion lies primarily in the DOJ's contention that Bazaarvoice's syndication network creates an insurmountable entry barrier. This statement ignores the fact that PowerReviews entered the market and competed effectively without offering a syndication product on par with Bazaarvoice's. Furthermore, the DOJ makes no attempt to quantify the number of Bazaarvoice customers that take advantage of the syndication offering. In fact, many manufacturers and retailers choose not to utilize this service, instead preferring to outsource to another vendor or perform the analytics in-house.
Notwithstanding these factual oversights, the assertion that Bazaarvoice's syndication network is a barrier to entry fails. The aggregation of data through the creation of consumer reviews is a profitable endeavor, but it is also an easily repeated endeavor. Bazaarvoice's reviews and sophisticated analysis may make it a better competitor but it does nothing to cement Bazaarvoice as an enduring competitor in the face of an improved service. Allegations of network effects as barriers to entry are made far too lazily, and the DOJ would have the trier of fact believe that a piece of data can only be captured once, or is a zero-sum game. This is just not the case. There is competition for data just as there is competition for any other product. Finally, as the value of data continues to increase, retailers and manufacturers will have less incentive to continue outsourcing this portion of the business to Bazaarvoice.
Unsurprisingly, entry is already occurring in this alleged market. Reevoo
are new entrants looking to disrupt competition, while Amazon
are established market participants looking to grow their profits at the expense of companies like Bazaarvoice. The DOJ's entire theory of harm is premised on a presumption of stagnancy that runs contrary to the nature of the high-tech and electronic commerce industries.
Antitrust enforcement in high tech markets poses special challenges -- to recognize the dynamic fast paced nature of competition, the fluidity of product markets, and the opportunities for new forms of rivalry. Unfortunately, the complaint in the Bazaarvoice case takes a static approach hinged to a few outdated documents. Without more it is unlikely a court will take the draconian step of unwinding this merger.