You joined Microsoft’s in-house antitrust team back in 2003. The technology industry has changed dramatically in the ten years since then. From your perspective, what have been the most important changes and what lessons should antitrust practitioners draw from those changes?
I only need to look at the desk in my office to answer that question. Ten years ago, I had a company-issued PC on the floor and it was networked to a central server in the basement of my building. I had a telephone running through a central PBX (i.e., Private Branch Exchange) an external back-up hard drive for my documents, a printer, and a cell phone that only made calls and sent texts. Today, I have none of these. My computer is a laptop convertible to a tablet with touch screen and a smart pen. Many of the Office applications that used to run on my computer, or on the server in the basement, are now delivered as a service over the internet from the Azure cloud. I no longer have a phone in my office – all of my telephony is over the web via Lync or Skype from my laptop using a wireless headset. All of my documents are stored in the cloud on SkyDrive, accessible from my choice of devices, and I take them with me to meetings on a Surface tablet running One Note. My Windows Phone has full web and local search, GPS mapping, access to all of my documents, full applications capability (and of course it can still send a text). All of my devices have an operating system (“OS”) with a common Windows user interface. All of my applications and data are accessible from any of these devices, anywhere, at any time. All of which highlights a number of critical changes in the technology industry over the past ten years.
First, competition has moved beyond the x-86 PC to a paradigm of competing ecosystems, starting with Apple’s ecosystem around iOS, the Apple devices that run it, its associated apps, and the App Store. Today, all the categories of hardware, software, and services that I use in the Microsoft ecosystem are available to a high degree in competing alternatives from established non-Windows ecosystems from Apple and Google, and on the horizon from Amazon and Samsung.
In turn, the emergence of these ecosystems is driving the consumerization of IT. By that I mean that today we all want to use our personal devices at work and at play. As a result, we want to be able to use the same interface and access the same functionalities/services/apps on our personal smartphone and our business tablet/laptop. In turn, Ecosystems and consumerization of IT are driving device convergence. For example, I no longer take a laptop to meetings and often leave it at home when I travel. I take a tablet computer instead, or increasingly, just my smartphone (which may be all I need at times given the increasing size and functionality of so-called “phablets”).
Second, the incredible growth in cloud-based technology enabled by broadband speeds has revolutionized how people connect to technology. Enterprise applications, and even my productivity applications such as Office, run in the cloud as a service, not locally using individually licensed code installed on a server down the hall or on my computer. Most data storage and data center functions are—or will soon be—hosted in the cloud.
Third, acquiring and then mining “big data” to fuel machine learning that underpins internet search, online advertising, and virtually all other online services has emerged as a critical input to competition (and significant barrier to entry) in every area of online endeavor. Access to user data at scale is probably the most important, and least understood or appreciated, aspect of competition in tech today. This is because online markets such as search, search advertising, and other online services are multi-sided (e.g., search advertisers pay while end search users play for free). These markets are characterized by strong network and dynamic scale effects, which result in high barriers to entry. For example, in the case of search, more users means more advertisers, which drives higher revenue. In turn, this allows greater revenue sharing with publishing partners at levels a smaller search engine cannot match based on the economics. User scale and associated data also accelerates the pace of innovation and results in a relative scale gap in terms of machine learning among the various search engines. In short, in these markets, data is “king” and your ability to compete is not just a function of how much data you have at a given time, it is a function of how much data you have relative to your competition over time that drives success.
"Access to user data at scale is probably the most important, and least understood or appreciated, aspect of competition in tech today."
So we have learned over the last ten years that competition and markets are not static. PC OSs may have one day been a distinct relevant market, but today the operating system is just one component of a multi-device ecosystem of apps and services founded in access to users and data at scale. Yet, despite this dramatic change and dynamic markets, I am in the camp that believes that antitrust already has the tools to address new-generation anticompetitive conduct. Often, what is lacking is a full appreciation of the extent to which the game has changed, that what once was considered crucial to competition is now often secondary at best. Looking just around the corner, how big data is acquired, controlled, and ultimately used is the next “big thing” in antitrust.
Your responsibilities at Microsoft include both government merger and conduct investigations, with an increasing emphasis in recent years on intellectual property issues. From your viewpoint, what are the three most important issues facing competition authorities around the world?
Certainly, you have alluded to one of the three: striking the right balance between encouraging competition through IP rights—which is good for consumers—and pursuing abusers of the system. Many authorities, and more recently the courts, are rightly focused on this dynamic as it relates to abuse of standard-essential patents (“SEPs”), which are patents that are necessarily infringed by unlicensed implementation of widely adopted formal technical standards such as WiFi. SEPs have market power stemming from the collaborative nature of standards-setting among industry players, and the fact that SEPs generally cannot be worked around. Effective and enforceable commitments to license these SEPs on fair, reasonable and non-discriminatory (“FRAND”) terms and conditions are necessary to keep SEP holders from leveraging their SEPs to “hold up” implementers of the standard for non-FRAND royalties.
Privacy and data protection is another critical issue for authorities. There is a balance with respect to offering consumers more control over privacy, and recognizing that advertising and advertisers have legitimate and strong roles to play. Privacy is an issue reaching a tipping point, while at the same time privacy policies among the major players are increasingly divergent. At Microsoft, we share consumers’ privacy concerns, and we are making concrete efforts to help give them more control, today and in the future. Last year, we built “Do Not Track” technology as the default option for our browser software, Internet Explorer 10. We decided to do this after consumers overwhelmingly told us that they want more control over how their personal information is used online. More recently, we launched Bing for Schools, which enables students to conduct searches in school without the distractions that result from advertising.
And, as I mentioned earlier, it is hard to overstate the importance of data (much of it gleaned from consumers who may be not entirely aware of the data being collected) in the ecosystem environment. And as you might suspect, I do not think that competition authorities have fully grasped the impact of Google’s search and advertising monopolies on competition in a variety of technology markets. All of these issues are interrelated to some extent.
Speaking of search and advertising, you were actively involved in the U.S. Federal Trade Commission’s antitrust investigation of Google’s search advertising practices, which closed in January of this year after Google made some voluntary commitments concerning its future conduct. Microsoft management described the FTC settlement in a blog post as “weak” and “unusual.” The European Commission appears to be nearing a settlement in its Google search advertising investigation and other jurisdictions continue to investigate the company’s search practices. What are the most important things that you disagree with regarding the FTC’s decision, and why? What may other antitrust agencies take into account in their respective search investigations?
Yes, the search settlement with Google was disappointing on a number of levels, starting with the fact that Google’s undertakings are voluntary and thus, unenforceable against Google absent a consent decree. It is clear that competition authorities—despite a lot of hard work by the talented, dedicated people in those agencies—are still grappling with the extent and impact of Google’s search and search advertising monopolies on competition in a number of markets.
Most recently, for example, we have seen Google’s affirmative steps to block high-quality YouTube apps for smartphone platforms that do not use Google’s dominant search engine as a default, such as Windows Phone. By having two different interfaces—a high-quality one made available for phone platforms that use Google search as their default search engine, such as the iPhone, and a low-quality one for phones that do not—Google degrades the YouTube viewing experience for customers that do not buy phones that default to Google search.
"Looking just around the corner, how big data is acquired, controlled, and ultimately used is the next “big thing” in antitrust."
On the positive side, the European Commission has identified a number of serious concerns in its investigation. First, in its general search results, Google gives preferential treatment to its own vertical search services over competitors’ links. Second, Google copies content from competing vertical search services and uses it in its own offerings, which reduces competitors’ incentives to invest in the creation of original content for the benefit of internet users. Third, agreements with web advertisers result in de facto exclusivity requiring them to obtain all or most of their requirements of search advertisements from Google. Fourth, Google contractual restrictions on software developers prevent them from offering tools that allow the seamless transfer of search advertising campaigns across Google’s ad platform (AdWords) and other platforms for search advertising.
After a recent market test, the Commission found Google’s proposed undertakings to resolve these concerns to be insufficient. We are waiting to see what happens next. In the meantime, I am hopeful that a number of other agencies around the world who are also investigating will look even more broadly at Google’s conduct—to see the forest as well as the trees, so to speak. Google possesses a very powerful monopoly and has used it to blunt actual and potential competition in a variety of markets, and that is a problem for competition and is inhibiting innovation. I suspect that is what the European Commission heard a lot about this summer when it was market testing Google’s proposed remedies to settle that investigation.
Some critics of the FTC’s decision in the Google search advertising investigation say that there should have been more of a focus on privacy and Google’s control over user data. Do you believe that privacy and “big data” are antitrust issues and, if so, why? What unique challenges do such issues present for antitrust practitioners?
Yes, I think these are antitrust issues. The fact is that access to data affects a wide range of other technologies. The most well-known examples are web search and search advertising. But a host of applications and websites now rely on user data, whether it is actively placed in the cloud by the user (e.g., when creating saved account information), collected in the course of online activity, scanned by sending email to certain email platforms, or generated as the result of the user carrying a phone running a particular OS as they are moving around. Many companies use that data in return for providing services, making the user’s data akin to the “price” being charged for those services. And no one would challenge the role of price in assessing consumer harm.
Increasingly, we are seeing traditional foreclosure practices by vertically-integrated companies such as Google, collecting big data and using it in downstream markets in competition with customers of its online advertising services. However, there is a major difference in that the data input in these cases —including data documenting our every interaction with a website such as clicks, time of engagement, even where a person may hover its cursor over a link—has not been produced by the website but belongs to consumers who have unknowingly relinquished control over it. I can easily foresee a time in the not too distant future where, subject to privacy law requirements, all market players whose competitive offerings depend on access to users’ data and who comply with privacy law requirements should have access to such big data on the ground that it is a public good, access to which will preserve competition in online services.
We also make a tremendous effort to be good stewards of users’ information and, not to violate the trust those users place in us, have built very strong privacy protections into our policies and products. I think that is true for most technology companies with some notable exceptions. User data is important for many technology companies as they compete with one another, and the privacy protections on that data can be an important differentiating feature or a service or term of supply that should factor into antitrust analysis of competitive dynamics.
And, of course, if one company gains control over user data at a certain chokepoint—as Google has with its search monopoly—it can wield that data to blunt potential competition in downstream or adjacent markets. In a world of ecosystems, that kind of control can be incredibly powerful.
One area in which Microsoft and the US and EU antitrust authorities do appear to share common ground has been the focus on addressing alleged abuses of standard-essential patents, or SEPs. Why do you believe this is an important competition law issue? And what do you see as the major problems that remain unresolved concerning SEPs?
I am not surprised that the US and EU authorities have found common ground over SEPs. As you know, SEPs are patents that are necessarily infringed by unlicensed implementation of a widely deployed standard adopted by a standard-setting organization, or SSO. The IEEE 802.11x series of Wi-Fi standards is an example; the standard contains a number of patented technologies contributed by different owners. When someone contributes patented technology to a standard, they typically voluntarily agree to license those patents on a fair, reasonable, and non-discriminatory, or FRAND, basis. Such a commitment prevents them from getting their patented technology into a standard that everyone needs for their products, and then turning around and using that patent to exclude rivals from the marketplace or to extract non-FRAND royalties and other licensing terms from implementers of the standard, often called “patent hold up.”
But what has happened is that certain SEP owners have found ways to nonetheless engage in patent hold up. For example, they have brought infringement claims on SEPs before the U.S. International Trade Commission (“ITC”) to seek exclusion orders barring importation into the U.S. of products that implement these standards, such as Xbox, or they have breached their FRAND promise by demanding non-FRAND licensing terms for their SEPs and seeking injunctions on those SEPs in court to unfairly pressure the implementer to accede to their demands. Facing the risk of exclusion from the marketplace, implementers can be forced to take non-FRAND terms, which can have anti-competitive effects in the market and a negative impact on the consumers of interoperable products and services.
Fortunately, we are seeing some pretty strong reactions to these tactics by courts and authorities. These include the FTC’s settlement with Google where Google agreed (at least, in theory) not to seek injunctions or exclusion orders on SEPs it acquired from Motorola, as well as the FTC’s settlement with Bosch that included similar restrictions on the use of FRAND-encumbered SEPs. Most recently, the Obama Administration vetoed an exclusion order against Apple barring importation of certain iPhone models by the ITC, something that has not been done since the 1980s, and instructed the ITC to take the public interest into account when dealing with exclusion requests based on FRAND-encumbered SEPs. In the meantime, the European Commission issued a Statement of Objections against Samsung based on a complaint by Apple that alleged improper use of SEPs.
We also have seen a number of related court decisions in the U.S. In 2012, Judge Posner denied Motorola Mobility’s claim for an injunction against Apple on its standard-essential patent because a FRAND royalty would provide all the relief to which Motorola would be entitled if it proved infringement of its SEPs. And in another case, Realtek, a United States district court in San Diego recently enjoined a SEP holder from pursuing an ITC exclusion order on its SEPs. And finally, Microsoft recently concluded a jury trial in federal district court in Seattle on its claims that Motorola’s conduct in seeking injunctions, even when judged alone apart from its general course of conduct, violated Motorola’s duty of good faith and fair dealing and thus breached its FRAND licensing commitments. On September 4, a unanimous jury rendered its verdict in favor of Microsoft on all claims and awarded damages to Microsoft in the amount of $14.5 million.
But there are unresolved issues. One is what exactly constitutes a FRAND offer and how does a court figure that out—I would point readers to Judge Robart’s very detailed April 25 opinion in the Microsoft v. Motorola case, which was the first time a U.S. court stepped in to determine the FRAND value for a SEP, and likely will be a roadmap for other courts . Specifically, Motorola’s demand for 2.25% of the end product selling price for its SEPs was not FRAND, and that the FRAND royalty was rather a set per-unit amount—$0.005 for each H.264-compliant Windows or Xbox product and $0.0347 for each 802.11-compliant Xbox product. For example, the 2.25% royalty that Motorola demanded would have been $11.25 for a $500 Dell laptop, but the actual FRAND royalty was just $0.005.
Another issue that the agencies in particular are grappling with is so-called patent “hold out,” which some claim can occur when a potential licensee refuses to accept a FRAND offer for a license to SEPs. “Hold out” (a refusal to take a FRAND license for SEPs) should be less of a concern, however, when one considers that the SEP holder can take any dispute between the parties to court and seek monetary relief that, if awarded, can be enforced by the court. Most importantly, patent hold out does not raise an antitrust concern, while patent hold up using SEPs most certainly does.
“[M]uch of the current litigation in the so called “smartphone patent wars” could have been be avoided if companies were willing to recognize the value of others’ creations in a way that is fair. Intellectual property rights are a two-way street."
Clearly, patents are becoming an ever-increasing focus of your practice. In addition to disputes involving SEPs, you recently have been involved in high profile transactions involving patents sold by Novell, Nortel, and Kodak. In light of what we have been discussing, why are patents of such interest to Microsoft and other technology companies right now?
Well, as I mentioned earlier, I think intellectual property has been developing as a competitive issue in recent years. From Microsoft’s perspective, the most basic consideration is the amount of investment we put into R&D compared to companies in other industries and indeed, compared to most of our competitors. Microsoft spent $10.4 billion on R&D in fiscal year 2013. Whether we incorporate all of that research into our products or not, we have a right to recoup that investment if someone else wants to use the technology that we have developed through that R&D. On the other hand, much of the current litigation in the so called “smartphone patent wars” could have been be avoided if companies were willing to recognize the value of others’ creations in a way that is fair. Intellectual property rights are a two-way street. This is why we have paid others more than $4 billion over the last decade to license others’ IP and implement it in our own products. As a general matter, we just place a high value on IP, for use in our products or our partners’ products, or in its own right as the end result of somebody’s investment of resources and creativity.
I think you can relate some of the more recent IP developments to some of the same trends we discussed earlier. For example, with the evolution of ecosystems, we have to worry about more than just software IP, but also hardware. That has been amplified by the variety of patents that read on cloud-related technology, and especially the number of standards at play. We want to be able to develop products in these areas ourselves and we want our partners to be able to do the same. When we negotiate licenses or when we negotiate broader cross-licenses or other agreements, patents are valuable consideration. It is pretty clear that our competitors feel the same way.
Finally, you do have infringement concerns that drive interest in patents. I think the open source movement has contributed a great deal to the software industry and many open source developers try very hard to avoid infringing others’ patents. But there are some operating systems out there, such as Android, which look and function very, very similarly to proprietary OSs that took many years of investment and development. I think a number of proprietary software developers have been concerned for some time that Android is free riding on their investments, a practice which distorts competition between mobile platforms.
Another major issue, which has drawn the attention of President Obama, is the patent litigation initiated by “Patent Assertion Entities” (“PAEs”), which in the President’s words “don’t actually produce anything themselves” and instead develop a business model “to essentially leverage and hijack somebody else’s idea and see if they can extort some money out of them.” Based on public reports, by the time this interview is published the Federal Trade Commission likely will be conducting a market study of PAEs (under its Section 6(b) of the FTC Act authority) and their impact on competition. What do you hope the FTC focuses on in conducting this market study, and why?
I trust that the FTC focuses on the truly bad actors and not ignore the pro-competitive aspects of secondary IP markets generally. Like many large technology companies, we have been on the receiving end of infringement lawsuits that are, as the president suggested, little more than “extortion.” I would similarly categorize the nuisance-value suits, where PAEs seek minimal amounts but from many, many defendants, hoping the value of settling minor suits is such that they can collect a payday in the aggregate. The president and the FTC are absolutely right to look at these kinds of abuses.
However, they should seriously consider the extent to which it makes sense to focus enforcement efforts and scarce resources on all entities that are getting rich by exploiting imperfections in the patent system and the courts, i.e., economically efficient actors, rather than focus their attention on improving the system such that it is less conducive to enabling such actors to thrive.
“If you accept that patents have an inherent value as the embodiment of someone’s investment of time, effort, and thought, then it makes sense to have a [secondary] market for those—to compensate the inventors and also to encourage other inventors to invest their time and effort in new technology."
This is particularly important for secondary IP markets that serve a healthy competitive purpose. Sometimes, inventors do not have the resources themselves to monetize their patents, which may be quite valuable and represent important innovations. If you accept that patents have an inherent value as the embodiment of someone’s investment of time, effort, and thought, then it makes sense to have a market for those—to compensate the inventors and also to encourage other inventors to invest their time and effort in new technology.
Rather than close down the secondary market for patents, the president and the FTC could begin by taking a few concrete steps to ameliorate conditions contributing to the worst abuses. Greater transparency, e.g., a requirement that patent holders disclose all of the patents that they own, and a loser pays regime in patent litigation are just two examples.
Some have suggested that “hybrid PAE activity,” or so-called “patent privateering,” is particularly problematic from an antitrust perspective due to the potential for “raising rivals’ costs.” Do you share that view? Why or why not?
Well, I think it depends on the activity. As with PAE activity, generally, you want to look out for abuse of the system without destroying the system altogether. A company may spend a tremendous amount on a particular technology, then find that it may not be useful to implement it in its products. It may make sense for that company to assign its patents to another entity that can monetize the patents and help recoup the investment in their development. So you want to enable that kind of activity and all the related benefits from it—better valuation of assets and more deal-making, increased innovation, and more vigorous competition.
We also need to be careful in lumping any input cost into the “raising rivals’ costs” bucket. Whether you agree with the theory as a basis for antitrust liability or not, all inputs raise cost. That is why the academic work in this area has always spoken in terms of artificially raising rivals costs. In the patent space, we have to ask, can a patent license artificially raise rivals’ costs, or is the licensee just paying what the market determines is the correct price for a given input that is being used by that company?
I do think it is possible for such activity to cross the line and raise antitrust concerns, though I would analyze it as I would any antitrust issue. In other words, you need to define a relative market, assess if there is market power, define what is the specific exclusionary act or acts, evaluate business justifications, look at standing and antitrust injury, etc. History has shown us that many business practices that were condemned due to lack of understanding, like most of the infamous 9 licensing “no-nos” from the 1970s, are widely recognized as procompetitive today.
Some of the more recent mergers and proposed transactions in which you have played a role include Microsoft’s acquisition of Skype (2011) and Google’s proposed search partnership with Yahoo (2008). What was the most interesting transaction that you were involved in, and why?
That is a hard question, I have been very fortunate over the years to have been involved in a number of really interesting and challenging situations. The search and search advertising markets in general have been a focus of mine given the increase in mergers since 2006, when Google acquired DoubleClick, with a lot of very important issues on the table. If I had to pick one example, I would pick our transaction with Yahoo!. As you know, the Department of Justice (“DOJ”) moved to block the failed Google/Yahoo! search deal in 2008, and our search deal with Yahoo! was approved shortly thereafter. Not only did these deals involve markets that were relatively new to the antitrust authorities at that time, but having two 3-2 deals in the same markets before the agencies in close proximity to one another, and then successfully making the case that our deal with Yahoo! was procompetitive and should be cleared was particularly satisfying. We also had to overcome a natural skepticism of many in the government and in the media to view Microsoft as an underdog in any space.
Prior to joining Microsoft, you spent 10 years working on antitrust matters as part of the in-house legal team at American Airlines. How have your experiences in the technology industry differed from your experiences in the airline industry? In which ways are they similar?
Of course, the airline industry is a more mature industry with a history of heavy regulation, while many technology businesses are relatively young and grew up with much less regulation. But in many ways, the airline industry is similar to some technology businesses. In scale businesses, larger networks are often able to provide higher quality offerings to consumers. And, over the years, the nature of these scale businesses has given rise to somewhat similar scrutiny from antitrust authorities.
Speaking of airlines and of larger networks, what is your view of the consolidation that has occurred in this industry during the last decade? How has that consolidation benefitted or harmed consumers? And do you wish to share any views on the recently proposed merger between your former employer, American Airlines and US Airways?
The turmoil in the airline industry is well-documented. Margins are extremely thin and disruptive bankruptcies have become commonplace. Low-cost carriers have proliferated and fares have stayed relatively low. Indeed, I believe if you look at airfares over the last 10 years, you will see that in real terms that fares have dropped. Against this background and given the relatively few overlaps involved, I think the DOJ probably got it right in clearing airline mergers in recent years, including Delta/Northwest, United Airlines/Continental, and Southwest/Airtran. I also think the American/US Airways deal is a deal that should probably be allowed to proceed. There are similarly few overlaps. American is in bankruptcy and this deal would allow it to emerge with a much-improved network better able to compete with its larger rivals. From the outside, I am not sure DOJ has adequately explained what makes this deal different from ones the same administration recently cleared.