Professional Documents
Culture Documents
Section 2 in A Web 2.0 World - An Expanded Vision of Relevant Product Markets
Section 2 in A Web 2.0 World - An Expanded Vision of Relevant Product Markets
0 WORLD: AN EXPANDED
VISION OF RELEVANT PRODUCT MARKETS
PAMELA JONES HARBOUR
TARA ISA KOSLOV*
Section 2 enforcement by the U.S. antitrust authorities is backat
least, that is what the antitrust enforcers themselves have been saying.
1
An examination of Section 2 enforcement principles could not be more
timely. Our goal in this essay is to share a few ideas for fresh thinking
about one element of Section 2 enforcementdefining relevant prod-
uct marketsthat may inform the agencies reinvigorated analysis of
dominant firm conduct, particularly in important technology sectors of
our economy.
* The authors were until recently, respectively, Commissioner and Attorney Advisor to
Commissioner Harbour, Federal Trade Commission. The views expressed herein are
those of the authors alone and do not necessarily reflect the views of the FTC or any other
individual Commissioner. This essay is based on a paper presented at the symposium,
Issues at the Forefront of Monopolizaton and Abuse of Dominance, University of Haifa,
Israel (May 2426, 2009). We are grateful to the other symposium participants for their
constructive remarks. We acknowledge significant contributions by Jamie Hine, and we
also thank Laurel Price for his helpful comments.
1
See, e.g., Christine A. Varney, Assistant Atty Gen., Antitrust Div., U.S. Dept of Justice,
Vigorous Antitrust Enforcement in this Challenging Era, Remarks Prepared for the U.S.
Chamber of Commerce 5 (May 12, 2009) [hereinafter Vigorous Antitrust Enforcement]
(The Antitrust Division must step forward and take a leading role in the development of
the Governments multi-faceted response to the current market conditions. Vigorous anti-
trust enforcement action under Section 2 of the Sherman Act will be part of the Divisions
critical contribution to this response.), available at http://www.justice.gov/atr/public/
speeches/245777.pdf; Jon Leibowitz, Chairman, Fed. Trade Commn, Remarks at Round-
table Conference with Enforcement Officials, ABA Section of Antitrust Law Spring Meet-
ing (Mar. 27, 2009) (I am very optimistic that under Christine Varney, who is going to be
a terrific Assistant Attorney General for Antitrust, that the Division and the [FTC] will be
much, much more in synch.), ANTITRUST SOURCE, Apr. 2009, http://www.abanet.org/
antitrust/at-source/09/04/Apr09-EnforcerRT4-29f.pdf. Presumably, in suggesting that
the FTC and Antitrust Division have not always been in synch, Chairman Leibowitz was
referring to the Commissions unanimous refusal to endorse a September 2008 DOJ re-
port on Section 2 enforcement policy. See infra note 5 and accompanying text.
769
770 ANTITRUST LAW JOURNAL [Vol. 76
A few months into the Obama Administration, newly appointed Assis-
tant Attorney General Christine Varney, head of the Antitrust Division
of the U.S. Department of Justice, rejected and withdrew
2
the Depart-
ments September 2008 report on Section 2.
3
AAG Varney indicated that
the Department, rather than adopting a monolithic test to govern Sec-
tion 2 matters, would rely upon well-established judicial precedents that
provide guidance on the limits of acceptable conduct by dominant
firms.
4
Varneys comments more closely aligned the DOJ position with
the views of at least three Commissioners of the Federal Trade
Commission.
5
Given these expressions of interest in more vigorous enforcement, it
seems likely that both U.S. antitrust enforcement agencies actively will
be seeking cases to fill their Section 2 investigatory pipelines and, ulti-
mately, to flesh out their Section 2 enforcement agendas. Recent experi-
2
Press Release, U.S. Dept of Justice, Justice Department Withdraws Report on Anti-
trust Monopoly Law (May 11, 2009), available at http://www.justice.gov/opa/pr/2009/
May/09-at-459.html.
3
U.S. DEPT OF JUSTICE, COMPETITION AND MONOPOLY: SINGLE-FIRM CONDUCT UNDER
SECTION 2 OF THE SHERMAN ACT (2008) (now withdrawn; see Press Release, supra note 2),
available at http://www.justice.gov/atr/public/reports/236681.pdf. This report repre-
sented the DOJs effort to synthesize its findings from a wide-ranging series of public
hearings held jointly by the DOJ Antitrust Division and the FTC from June 2006 through
May 2007, relating to the evaluation of unilateral conduct under Section 2. See Fed. Trade
Commn, Public Hearings: Federal Trade Commission and Department of Justice Hearings on Sec-
tion 2 of the Sherman Act: Single-Firm Conduct As Related to Competition, http://www.ftc.gov/
os/sectiontwohearings/index.shtm (listing a collection of materials relating to Section 2
hearings).
4
See, e.g., Varney, Vigorous Antitrust Enforcement, supra note 1, at 9 (citations
omitted):
While the Department is not proposing any one specific test to govern all Sec-
tion 2 matters at this time, I believe the balanced analyses reflected in the lead-
ing cases interpreting the reaches of the Sherman Act provide important
guidance in this regard. In particular, leading Section 2 casesfrom Lorain Jour-
nal v. United States to Aspen Skiing Co. v. Aspen Highlands Skiing Corp. to United
States v. Microsofthighlight a common concern regarding the harmful effects
of a monopolists exclusionary or predatory conduct on competition and, ulti-
mately, consumers. Reinvigorated Section 2 enforcement will thus require the
Division to go back to the basics and evaluate single-firm conduct against these
tried and true standards that set forth clear limitations on how monopoly firms
are permitted to behave. There can be no better charter for our return to funda-
mental principles of antitrust enforcement.
5
The FTC had refused to join the Departments September 2008 Section 2 Report
and had instead issued its own statement. Statement of Commissioners Harbour, Leibo-
witz and Rosch on the Issuance of the Section 2 Report by the Department of Justice
(Sept. 8, 2008), available at http://www.ftc.gov/os/2008/09/080908section2stmt.pdf. See
also Federal Trade Commn, Working Papers: Federal Trade Commission and Depart-
ment of Justice Hearings on Section 2 of the Sherman Act: Single-Firm Conduct As Re-
lated to Competition (developed by FTC staff involved in the joint FTC/DOJ Section 2
hearings), available at http://www.ftc.gov/os/sectiontwohearings/index.shtm.
2010] SECTION 2 IN A WEB 2.0 WORLD 771
ence suggests that the agencies will pay particularly close attention to
firms that appear to be attaining dominance in new-technology markets.
For example, Google can be seen as the new Microsoft, in terms of its
tremendous capacity for innovation and market success, as well as the
intense scrutiny that henceforth will be cast upon the companys every
move.
6
At the end of the Bush Administration, the DOJ Antitrust Divi-
sion closely examined a proposed transaction between Google and Ya-
hoo! relating to search advertising. The parties abandoned the deal in
the face of an imminent enforcement action.
7
The DOJ has been closely
monitoring the class action litigation surrounding the Google Book
Search service,
8
while the FTC investigated the potential competitive im-
plications of interlocking directorates between Google and Apple.
9
6
See Steve Lohr & Miguel Helft, New Mood in Antitrust May Affect Google, N.Y. TIMES,
May 18, 2009, at B1(In this new climate, the stakes appear to be highest for Google, the
rising power of the Internet economy.), available at http://query.nytimes.com/gst/
fullpage.html?res=9B03E3D8163AF93BA25756C0A96F9C8B63.
7
Press Release, U.S. Dept of Justice, Yahoo! Inc. and Google Inc. Abandon Their
Advertising Agreement (Nov. 5, 2008) (explaining that the DOJ was preparing to file an
antitrust lawsuit alleging that the proposed agreement likely would harm competition in
markets for Internet search advertising and Internet search syndication, where the two
firms accounted for 90 percent or more of each relevant market), available at http://www.
usdoj.gov/atr/public/press_releases/2008/239167.pdf; see also Press Release, Yahoo!, Ya-
hoo! Announces Termination of Services Agreement by Google (Nov. 5, 2008), available
at http://yhoo.client.shareholder.com/PRESS/releasedetail.cfm?ReleaseID=345734.
8
See, e.g., Statement of Interest of the United States of America Regarding Amended
Settlement Agreement, The Authors Guild, Inc. v. Google Inc., 05-8136-DC (S.D.N.Y.)
(Feb. 4, 2010) (recommendation that the court should not accept the settlement, but
should continue to encourage the parties to pursue alternatives and, perhaps, provide
additional guidance to the parties), available at http://www.justice.gov/atr/cases/
f255000/255012.pdf; Press Release, U.S. Dept of Justice, Justice Department Submits
Views on Amended Google Book Search Settlement (Feb. 4, 2010), available at http://
www.justice.gov/atr/public/press_releases/2010/255014.pdf; see also Statement of Inter-
est of the United States of America Regarding Proposed Class Settlement, The Authors
Guild, Inc. v. Google Inc., 05-8136-DC (S.D.N.Y.) (Sept. 18, 2009) (comment on earlier
version of settlement), available at http://www.justice.gov/atr/cases/f250100/250180.pdf;
Press Release, U.S. Dept of Justice, Justice Department Submits Views on Proposed
Google Book Search Settlement (Sept. 18, 2009), available at http://www.justice.gov/atr/
public/press_releases/2009/250181.pdf; Miguel Helft, Justice Dept. Opens Antitrust Inquiry
into Google Books Deal, N.Y. TIMES, Apr. 29, 2009, at B5, available at http://www.nytimes.
com/2009/04/29/technology/internet/29google.html.
9
See, e.g., Miguel Helft & Brad Stone, Board Ties at Apple and Google Are Scrutinized, N.Y.
TIMES, May 4, 2009, at B1, available at http://www.nytimes.com/2009/05/05/technology/
companies/05apple.html; Press Release, Fed. Trade Commn, Statement of Bureau of
Competition Director Richard Feinstein Regarding the Announcement that Google CEO
Eric Schmidt Has Resigned from Apples Board (Aug. 3, 2009) (We have been investigat-
ing the Google/Apple interlocking directorates issue for some time and commend them
for recognizing that sharing directors raises competitive issues, as Google and Apple in-
creasingly compete with each other.), available at http://www.ftc.gov/opa/2009/08/
googlestmt.shtm; Press Release, Fed. Trade Commn, Statement of Chairman Jon Leibo-
witz Regarding the Announcement that Arthur D. Levinson Has Resigned from Googles
Board (Oct. 12, 2009) (Google, Apple, and Mr. Levinson [Apple Board Member] should
772 ANTITRUST LAW JOURNAL [Vol. 76
More recently, the Commission initiated an investigation of Googles
proposed acquisition of AdMob, a mobile advertising firm.
10
Whether
Google deserves all of this attention or not, Googles behavior is likely to
serve as a focal point for discussions regarding the limits of acceptable
conduct for dominant, or potentially dominant, Internet-based firms.
Debateboth domestic and internationalwill continue regarding
the appropriate analytical framework for dominant firm conduct.
11
This
essay will focus on one particular element of that framework. Among the
first principles of antitrust is that any analysis of potentially anticompeti-
tive conduct requires the definition of a relevant product market. In any
industry, product market definition is capable of determining the out-
come of an antitrust inquiry or enforcement action.
12
In technology
markets that evolve rapidly, the challenge of market definition in-
creases; this is especially true of Internet markets characterized by large
numbers of innovators, rapid application development capability, and
relatively low entry barriers. As technologies converge, todays comple-
ments might become tomorrows substitutes. Is it possible for anyone
consumers, business actors, legal counselors, or regulatorsto under-
stand fully how different products may relate to each other a month or a
year down the road?
To ensure that consumers are adequately protected, the antitrust en-
forcement agencies should actively embrace the possibility of defining
be commended for recognizing that overlapping board members between competing
companies raise serious antitrust issues and for their willingness to resolve our concerns
without the need for litigation.), available at http://www.ftc.gov/opa/2009/10/google.
shtm.
10
Google Public Policy Blog, An Update on Our AdMob Acquisition (Dec. 23, 2009), (ac-
knowledging FTC inquiry and receipt of FTC second request, which typically signifies a
substantial merger investigation), http://googlepublicpolicy.blogspot.com/2009/12/up-
date-on-our-admob-acquisition.html. Several months earlier, Googles CEO had an-
nounced that [a]cquisitions are turned on again after the worst of the economic
downturn, and that henceforth Google anticipated making one acquisition a month in
lieu of new hiring. Google CEO Sees One Small Acquisition a Month, REUTERS, Sept. 23, 2009,
available at http://www.reuters.com/article/idUSTRE58M4LV20090923.
11
The International Competition Network (ICN) is a key forum where this discussion
has been ongoing for several years. The ICN Unilateral Conduct Working Group, formed
in May 2006, has been engaged in a detailed comparative examination of the challenges
involved in addressing anticompetitive unilateral conduct of dominant firms and firms
with market power, and has established a goal of promot[ing] greater convergence and
sound enforcement of laws governing unilateral conduct. Intl Competition Network,
Unilateral Conduct Working Group, http://www.internationalcompetitionnetwork.org/
index.php/en/working-groups/unilateral-conduct.
12
See, e.g., United States v. Oracle Corp., 331 F. Supp. 2d 1098, 115861 (N.D. Cal.
2004) (plaintiffs did not meet burden of proof on product market definition; court could
not exclude related software solutions outside the narrowly drawn enterprise software
application markets argued by plaintiffs).
2010] SECTION 2 IN A WEB 2.0 WORLD 773
relevant product markets based on current and future interrelationships
and convergence among various technologies. Under this approach, the
market definition exercise would strive, as usual, to identifyfrom a
functional perspectivewhere in the marketplace consumers currently
benefit from competition. Market definition would, however, also ac-
count for areas where new and meaningful competition is likely to
emerge, based on technological developments as well as emerging con-
sumer preferences.
We propose two types of product market definitions that would em-
body these principles. First, we suggest the definition of markets for
data, separate and apart from markets for the services fueled by these
data.
13
Data market definition would reflect the distinction between data
collection at one point in time and expanded data usage at some later
date. Data market definition also would properly recognize the in-
creased significance and value of the massive and growing data troves
that constantly are generated by Internet activities. Additionally, and im-
portantly, this approach to market definition would be consistent with
marketplace reality: Internet-based firms often derive great value from
user data, far beyond the initial purposes for which the data initially
might have been shared or collected, and this value often has important
competitive consequences. In contrast, product market definitions
based only on a snapshot of current data usage may not accurately cap-
ture this aspect of competition, especially in markets that exhibit net-
work effects based on aggregations of data.
Second, the agencies should consider framing relevant product mar-
kets around privacy issues. In the United States, up until now, privacy
has been thought of primarily as a consumer protection issue. But pri-
vacy is an increasingly important dimension of competition as well,
which is exactly why modern antitrust analysis must take privacy into
account. It makes no sense to maintain an artificial dichotomy between
competition and consumer protection law, especially when their goals
are complementary.
14
The product market definition exercise may be a
13
Such services might include, for example, social networking, mapping, email, photo
sharing, calendaring, document management, and advertising, to name just a few.
14
Competition law strives to provide consumers with an efficient range of choices,
while consumer protection law aims to ensure that consumers can exercise these choices
based on truthful and accurate information. The synergies between competition and con-
sumer protection have been highlighted in, for example, Neil W. Averitt & Robert H.
Lande, Consumer Sovereignty: A Unified Theory of Antitrust and Consumer Protection Law, 65
ANTITRUST L.J. 713, 713 (1997); Spencer Weber Waller, In Search of Economic Justice: Consid-
ering Competition and Consumer Protection Law, 36 LOY. U. CHI. L.J. 631 (2005).
774 ANTITRUST LAW JOURNAL [Vol. 76
useful way to begin the process of making privacy cognizable under the
antitrust laws.
15
We acknowledge, but do not attempt here, the challenge of deriving
these kinds of product market definitions according to traditional mar-
ket definition principles (e.g., defining buyers and sellers, identifying
substitutes, applying the hypothetical monopolist test, etc.). Such efforts
are best undertakenon a case-by-case basisby counselors, enforcers,
and economists who have access to market-specific facts regarding data
acquisition and use. The current joint FTC/DOJ inquiry into proposed
Merger Guidelines reforms
16
may yield useful insights into a methodol-
ogy that will facilitate the product market definition in these complex
technology-driven areasespecially if the agencies endorse an effects-
based analytical approach that supports backing into a product market
definition based on direct effects evidence.
17
15
This essay deals primarily with product market definition in the context of evaluating
unilateral conduct by dominant (or potentially dominant) firms under Section 2. We
note, however, that the same concepts could apply to product market definition in the
merger context (to the extent that courts continue to apply a legal framework that re-
quires plaintiffs to meet their burden of proof by defining relevant product markets).
Merger analysis often is premised on the idea that a merger may create a dominant firm,
which might then be able to engage in anticompetitive conduct that would not have been
possible but for the merger. Given the incipiency approach of the Clayton Act, which
proscribes transactions whose effect may be (not will be) substantially to lessen com-
petition, firms typically are not allowed to consummate a merger or acquisition that is
likely to lead to a dominant market share in a properly defined relevant market. See
Brown Shoe Co. v. United States, 370 U.S. 294 (1962); United States v. Philadelphia Natl
Bank, 374 U.S. 321 (1963); accord FTC v. H.J. Heinz Co., 246 F.3d 708, 713 (D.C. Cir.
2001). An important distinction between merger and monopolization law is that, under
Section 2, dominance (or the likelihood of dominance) is not prohibited outright.
Rather, a firm may seek or attain lawful dominance unilaterally, as long as the firm does
not engage in acts of exclusionary conduct. United States v. Grinnell Corp., 384 U.S. 563,
57071 (1966) (cited as settled law in Verizon Communications Inc. v. Law Offices of Curtis
V. Trinko, LLP, 540 U.S. 398, 407 (2004)).
16
Press Release, Fed. Trade Commn, Federal Trade Commission and Department of
Justice to Hold Workshops Concerning Horizontal Merger Guidelines (Sept. 22, 2009),
available at http://www.ftc.gov/opa/2009/09/mgr.shtm. Press Release, Fed. Trade
Commn, Federal Trade Commission Seeks Views on Proposed Update of the Horizontal
Merger Guidelines (Apr. 20, 2010), available at http://www.ftc.gov/opa/2010/04/hmg.
shtm.
17
See Fed. Trade Commn & U.S. Dept of Justice, Horizontal Merger Guidelines: Ques-
tions for Public Comment (Sept. 22, 2009), available at http://www.ftc.gov/bc/work
shops/hmg/hmg-questions.pdf. Question 2, in particular, asks: Should the Guidelines
be revised to address more fully how the Agencies use evidence about likely competitive
effects that is not based on inferences drawn from increases in market concentration? If
such revisions are undertaken, what types of such direct evidence are pertinent? Fed.
Trade Commn & U.S. Dept of Justice, Horizontal Merger Guidelines: Questions for Pub-
lic Comment (Sept. 22, 2009), available at http://www.ftc.gov/bc/workshops/hmg/hmg-
questions.pdf. The proposed revised Guidelines answer this question in the affirmative,
leading off with a detailed discussion of the types and sources of competitive effects evi-
dence. Fed. Trade Commn & U.S. Dept of Justice, Horizontal Merger Guuidelines, For
2010] SECTION 2 IN A WEB 2.0 WORLD 775
We also acknowledge that these data and privacy concepts arguably
might be addressed as part of a typical entry analysis, rather than at the
market definition stage.
18
In some cases, this choice may not be out-
come-determinative. We believe, however, that in certain cases there
would be value in considering these issues as part of market definition,
for several reasons. Market definition plays an important role in organiz-
ing and shaping investigations and enforcement recommendations. If
putative market definitions do not yield a credible risk of market power,
an investigation may be framed differently, competitive harm may seem
unlikely, and it is possible that any entry analysis might not be as fully
developed.
19
By considering alternate market definitions, however, addi-
tional theories of competitive harm may emerge and be fully tested. In
addition, many judges treat market definition as a threshold issue, espe-
cially when, for efficiencys sake, they seek to decide cases on the nar-
rowest possible grounds. If a judge deems that a plaintiff has not met its
market definition burden, the analysis may be cut off there, and ques-
tions of entry may never be reached. Therefore, market definition is,
and will remain, an important organizing principle and strategic tool for
enforcers and other plaintiffs.
20
I. A PRIMER ON INTERNET TECHNOLOGIES
The following discussion attempts to distill the core market elements
that might affect and inform antitrust analysis, particularly with respect
to product market definition.
Public Comment (Apr. 20, 2010), at 36, available at http://www.ftc.gov/os/2010/04/10
0420hmg.pdf. The section on market definition, which now comes later in the document,
emphasizes that [t]he Agencies analysis need not start with market definition and that
[e]vidence of competitive effects can inform market definition, just as market definition
can be informative regarding competitive effects. Id. at 7.
18
For example, the need to amass huge troves of data, or one firms huge lead in
assembling such a data trove, might be characterized as an entry barrier. So might the
ability to offer consumers their desired level of privacy protectionespecially where con-
sumers demand high levels of privacy protection and will only turn to firms with a proven
reputation for respecting consumer privacy.
19
For example, when Google acquired DoubleClick, the Commission adopted product
market definitions that led the transaction to be characterized primarily as a vertical
merger, and therefore likely to be procompetitive. In contrast, Commissioner Harbours
dissent proposed product market definitions that would have enabled the transaction to
be characterized more in terms of a horizontal overlap, which would have led to a differ-
ent set of presumptions regarding competitive effects. See infra notes 4648 and accompa-
nying text.
20
Malcolm Coate & Jeffrey Fischer, Why Markets Matter for Evidence-Based Merger Analysis,
OXERA AGENDA (Jan. 2010), available at http://www.oxera.com/cmsDocuments/Agenda_
January%2010/Market%20definition.pdf.
776 ANTITRUST LAW JOURNAL [Vol. 76
A. WHAT IS WEB 2.0?
The term Web 2.0 is most commonly used to describe the second
generation of the Internet. This generation of Internet technology is
characterized by greater interaction and connectedness among Web
users, along with the use of rich, interactive media.
21
Web 2.0 relies primarily on the sharing of information, which means
that user contributions are highly valuable. The new Web [is] a tool for
bringing together the small contributions of millions of people and
making them matter.
22
Successful Web 2.0 companies have embraced
the power of the web to harness collective intelligence.
23
To give a sim-
ple example, in Web 1.0, one might have accessed a static, online ver-
sion of Encyclopedia Britannica, where the content had not changed since
it was created and posted online. In Web 2.0, in contrast, one might
consult or contribute to Wikipedia, an online reference that constantly
changes and expands based on user input.
24
Web 2.0 takes advantage of multimedia capabilities that are enabled
by the widespread availability of high-bandwidth Internet access. In Web
1.0, a user might have accessed a transcript of the text of Martin Luther
Kings I Have a Dream speech. In Web 2.0, a user can search YouTube
and find hundreds of video clips of that speech. A user might add her
own comments, either within YouTube or, perhaps, on a separate blog
page that provides links to the video. A user who actually attended the
1963 speech might add her own perspective and narrative, along with
scanned snapshots from her own photo albums, which would expand
the historical record available to others. Or a user could mash up ei-
ther the audio or video of Dr. Kings speech with other user- or profes-
sionally-created material, creating a whole new form of expression.
Success in the Web 2.0 world relies on the creation and exploitation
of network effects, which means that a given Web site becomes more
21
For a detailed explanation of the Web 2.0 concept (by one of the people who
originated the term), see Tim OReilly, What Is Web 2.0: Design Patterns and Business Models
for the Next Generation of Software (Sept. 30, 2005), http://www.oreillynet.com/pub/a/
oreilly/tim/news/2005/09/30/what-is-web-20.html.
22
Lev Grossman, Times Person of the Year: You, TIME, Dec. 13, 2006, available at http://
www.time.com/time/magazine/article/0,9171,1569514,00.html. In 2006, Time named
You as its 2006 Person of the Year in recognition of the millions of contributors whose
user-generated content drove the success of Web sites, such as Wikipedia, YouTube, and
MySpace. Silicon Valley consultants call it Web 2.0, as if it were a new version of some old
software. But its really a revolution. Id.
23
OReilly, supra note 21.
24
Wikipedia: About, http://en.wikipedia.org/wiki/Wikipedia:About.
2010] SECTION 2 IN A WEB 2.0 WORLD 777
useful and increases in value with larger numbers of users.
25
One cur-
rent example is the tremendous growth in the number of Facebook
users, particularly among older age groups who were not the sites origi-
nal target audience of college and high school students.
26
As more peo-
ple in this expanded demographic have joined, Facebook has become a
popular means of communication among friends of all ages, extended
family members, and even business associates.
27
With a broad range of
people spending large amounts of time on Facebook, many companies
and organizations are rushing to establish a Facebook presence as a
means to market their goods and services and, perhaps, better under-
stand their target audience. Application developers, too, are vying for
the attention of Facebook users and seeking ways to monetize these new
connections.
28
Another prime example of the role of network effects in establishing
a strong Web 2.0 presence is the Google search engine algorithm.
Googles initial success in the search market derived from a novel and
unique search methodology that excelled at generating highly relevant
search results.
29
Googles popularity has exploded, however, because the
accuracy and relevance of Google search results actually improves as
more and more searches are conducted. This improved performance, in
turn, has attracted even more searchers to Google, which further im-
proves the search results, and so on, in a continually self-reinforcing
25
Seminal literature regarding network effects includes Michael Katz & Carl Shapiro,
Systems Competition and Network Effects, 8 J. ECON. PERSP., Spring 1994, at 93; Michael L.
Katz & Carl Shapiro, Network Externalities, Competition, and Compatibility, 75 AM. ECON. REV.
424 (1985). See also OReilly, supra note 21 (Network effects from user contributions are
the key to market dominance in the Web 2.0 era.).
26
As of April 2010, Facebook had over 400 million active users, more than half of
whom were logging in daily. Facebook Press Room, Facebook Factsheet, http://www.face
book.com/press/info.php?factsheet; Facebook Press Room, Statistics, http://www.face
book.com/press/info.php?statistics; see also Facebook Press Room, Company Timeline,
http://www.facebook.com/press/info.php?timeline (Facebook users doubled from 100
million to 200 million between August 2008 and April 2009, and had tripled to over 300
million by September 2009).
27
This evolution of Facebook appears to be consistent with the original vision of
Facebooks founder, Mark Zuckerberg, who hoped that the Facebook social utility ulti-
mately would become a truly global digital phone book. . . . . [N]ow his vision goes well
beyond the site as a digital phone book. It becomes the equivalent of the phone itself: It is
the main tool people use to communicate for work and pleasure. . . . Facebook will be
where people live their digital lives . . . . See Jessi Hempel, How Facebook is taking over our
lives, CNNMONEY.COM, Mar. 11, 2009, http://money.cnn.com/2009/02/16/technology/
hempel_facebook.fortune/index.htm.
28
The rise in popularity of Twitter and LinkedIn offers similar examples, albeit with a
more targeted focus (Twitter for short bits of time-sensitive information, LinkedIn for
professional connections and information exchanges).
29
See Google, Corporate Information: Technology Overview, http://www.google.com/
corporate/tech.html.
778 ANTITRUST LAW JOURNAL [Vol. 76
loop. As a result, Google has now entered the lexicon as a verb synony-
mous with the very act of Internet searching.
30
Network effects lead to the collection and sharing of ever-expanding
amounts of information, which often is quite valuable to users. But as
has been demonstrated in other technology areas, network effects also
may make it easier for a firm to achieve and maintain a position of mar-
ket dominance. The Microsoft case provides a notable example of the
relationship between network effects and market dominance. In its orig-
inal Findings of Fact, the district court found that the large installed
base of Windows users creates incentives to develop compatible software
applications, which in turn reinforces demand for Windows, in a posi-
tive feedback loop that creates positive network effects.
31
From the
perspective of would-be competitors, however, the court explained that
these network effects made it difficult for a rival operating system to
attract sufficient software development, a situation characterized by the
court as the applications barrier to entry.
32
B. CLOUD COMPUTING
In order to properly analyze evolving Internet markets, it is also neces-
sary to understand the concept of cloud computing. In very simple
terms, cloud computing refers to complex applications and associated
data that reside on remote servers (sometimes halfway across the world)
and are accessed through the Internet, rather than residing on individ-
ual PCs or individual corporate networks. No matter where you are, if
you have a laptop and an Internet connection, you can take advantage
30
For example, the word google now shows up as a verb in a popular dictionary. See
MERRIAM-WEBSTER DICTIONARY (Online Edition 2010), http://www.merriam-webster.
com/dictionary/google.
31
United States v. Microsoft Corp., 84 F. Supp. 2d 9, 20 (D.D.C. 1999) (Findings of
Fact, 39).
32
Id. at 1922 (Findings of Fact, 3644). After the D.C. Circuits main Microsoft
opinion, the remand to the district court for further remedial proceedings, and the even-
tual settlement, the D.C. Circuit considered a subsequent appeal by the Commonwealth
of Massachusetts, which challenged the remedy as inadequate. In denying this appeal, the
D.C. Circuit was careful to distinguish between the effects of Microsofts exclusionary con-
duct and the positive network effects inherent in the operating system market.
[I]t does not follow that, because a proposed requirement could reduce the
applications barrier to entry, it must be adopted. Recall the applications barrier
to entry arose only in part because of Microsofts unlawful practices; it was also
the product of positive network effects. 84 F. Supp. 2d at 20. If the court is not
to risk harming consumers, then the remedy must address the applications bar-
rier to entry in a manner traceable to our [earlier] decision.
Massachusetts v. Microsoft Corp., 373 F.3d 1199, 1226 (D.C. Cir. 2004). In other words,
while the court could impose or condone a remedy that responded directly to Microsofts
exclusionary conduct, remedial action would not be justified solely to counter legitimate
market power acquired as a result of positive network effects.
2010] SECTION 2 IN A WEB 2.0 WORLD 779
of applications that rely on an Internet interface to harness the comput-
ing power and storage capabilities of massive remote servers.
33
Many cloud applications are targeted directly to consumers. Familiar
examples include Web-based e-mail, online calendars, document man-
agement sites, and photo-sharing sites, many of which are made availa-
ble to consumers fee-free. Cloud computing also has many high-level
commercial applications. Increasingly, even large firms with sophisti-
cated computing needs do not have to bring massive computing power
in-house. Instead, a firm might equip itself with basic computers and In-
ternet access, then effectively contract out its entire IT infrastruc-
ture.
34
Cloud computing can lead to significant efficiencies. For example,
highly mobile consumers, who may choose to utilize free cloud-based
applications, enjoy the flexibility of data that are no longer tethered to a
specific location, which facilitates access from anywhere at any time. On
the commercial end, firms might choose to start out by transferring ba-
sic IT needs to the cloud, then scale up their reliance on cloud-based
computing over time to incorporate the companys most mission-critical
software applications.
As with network effects, however, cloud computing is not without
risks. Someone elsenot the userbecomes the caretaker of the un-
derlying data that have been fed into a given application. This raises
questions regarding who may have access to, and control of, valuable
data, and what might be done with those data.
35
In addition to potential
privacy and data security implications,
36
one particular area of concern
33
Its always a balancing act, but todays combination of high-speed networks, sophisti-
cated PC graphics processors, and fast, inexpensive servers and disk storage has tilted
engineers toward housing more computing in data centers. Aaron Rocadela, Computing
Heads for the Clouds, BUS. WK., Nov. 16, 2007, available at http://www.businessweek.com/
technology/content/nov2007/tc20071116_379585.htm.
34
Salesforce.com is one of the best known and most successful examples of the cloud
computing business model. The company originally offered cloud-based software solu-
tions for sales and customer service; the company has since expanded its business model
to include a wider range of software applications. See Salesforce.com, What Is Cloud Com-
puting?, http://www.salesforce.com/cloudcomputing/.
35
See, e.g., Bobbie Johnson, Cloud Computing Is a Trap, Warns GNU Founder Richard Stall-
man, GUARDIAN, Sept. 29, 2008, http://www.guardian.co.uk/technology/2008/sep/29/
cloud.computing.richard.stallman (leading advocate of free software warned that cloud
computing applications pose the same risk as other proprietary software, leading to a loss
of control over ones data).
36
For example, consumers may trade some of their privacy for free cloud applica-
tions, funded by revenues from targeted advertising derived from applications data. It
may not always be clear, however, whether consumers fully understand and have con-
sented to this trade-off. With respect to data security, responsibility for maintaining ade-
780 ANTITRUST LAW JOURNAL [Vol. 76
relates to data portability. If cloud services providers adopt proprietary
standards and formats rather than open ones, data portability may be
limited; users may run the risk of lock-in once they commit to a certain
platform, and entry by new services providers may become more
difficult.
37
C. THE COMMERCIAL MODEL AND ITS IMPLICATIONS
As a final background point, it is important to recognize one key as-
pect of todays Internet environment. In many Web 2.0 markets, the
revenue stream is not matched directly to specific Internet-based ser-
vices. Rather, to a large extent, revenueor monetization, to use the
current term of artderives from the accumulation of data, which can
then be put to myriad commercial uses. This model explains how many
Web sites are able to finance the provision of free content and ser-
vices. The sites are subsidized, in effect, by trading on the value of accu-
mulated data. In many instances, the data come from individual
consumers, who may or may not realize that they are paying for free
information or services by disclosing their personal information.
A primary example of this revenue model is behavioral advertisinga
topic to which the FTC has been paying a great deal of attention.
38
Be-
havioral advertising relies on the sophisticated analysis of data about
usersincluding their actions, choices, and revealed preferencesto
make predictions about which advertisements might appeal to which
consumers. Those targeted ads are then served up while a user is brows-
ing a Web site. The ads typically show up along the top or side of the
main content on a page, with different ads being displayed to different
quate safeguards will need to be allocated among the various guardians who have access
to cloud applications data at processing, transmittal, and storage points.
37
See, e.g., Battle of the Clouds, ECONOMIST, Oct. 17, 2009, at 16 (Buyers of cloud services
must take account of the dangers of lock-in.). But see Clash of the Clouds, ECONOMIST, Oct.
17, 2009, at 80, 8182 (the economics of cloud computing may favor open standards, thus
mitigating potential lock-in concerns, although certain closed applications with desira-
ble features still may be popular).
38
See, e.g., Press Release, Fed. Trade Commn, FTC Staff Proposes Online Behavioral
Advertising Privacy Principles (Dec. 20, 2007), available at http://www.ftc.gov/opa/2007/
12/principles.shtm (includes link to text of staff statement); Press Release, Fed. Trade
Commn, FTC Staff Revises Online Behavioral Advertising Principles (Feb. 12, 2009),
available at http://www.ftc.gov/opa/2009/02/behavad.shtm; FED. TRADE COMMN, STAFF
REPORT, SELF-REGULATORY PRINCIPLES FOR ONLINE BEHAVIORAL ADVERTISING (2009) [here-
inafter FTC STAFF BEHAVIORAL ADVERTISING REPORT], available at http://www.ftc.gov/os/
2009/02/P085400behavadreport.pdf; see also Commissioner Pamela Jones Harbour, Con-
curring Statement Regarding Staff Report, Self-Regulatory Principles for Online Behav-
ioral Advertising (Feb. 12, 2009) [hereinafter Harbour Behavioral Advertising
Concurrence], available at http://www.ftc.gov/os/2009/02/P085400behavadharbour.
pdf.
2010] SECTION 2 IN A WEB 2.0 WORLD 781
users who browse the same Web site. This placement method is very
different from the traditional, static advertising model, where a given ad
would be assigned to a specific space in a printed newspaper or a spe-
cific time slot in a given TV show.
In online advertising, an advertiser gets the most bang for its buck
when its ads generate more click-throughsmeaning that a users in-
terest is piqued, and the user completes an action, such as clicking on
the ad to pursue more information. The point of traditional, non-behav-
ioral display advertising is a numbers game: it seeks to place a given ad
in front of as many eyeballs as possible, figuring that if a certain percent-
age of viewers will respond to the ad, more eyeballs will equal more
click-throughs. Behavioral advertising takes a more targeted approach.
It attempts to place highly relevant ads in front of the right sets of eye-
balls, to maximize the likelihood of a click-through from each viewer. If
the ads are likely to be more effective at attracting customers, an adver-
tiser will pay more to place the ads, which will generate a larger revenue
stream for the Web site.
Sometimes ads are served up because they relate directly to the con-
tent on the pages the user has been viewing. One recent article
39
cited
this example: a user is Web surfing to research the purchase of a new
car, and ads for the users chosen brand of car soon start popping up.
That type of advertising does not require much data analysis. In more
sophisticated applications of behavioral advertising, however, ads may
appear that are not entirely related to the content of the page being
viewed. Rather, ads may be selected and served up based not only on the
current page, but also on other aggregated data about the userinclud-
ing, for example, searches the user recently has conducted, other sites
the user has visited, the users geographic location, and other factors.
Here is one stylized (but entirely plausible) example.
40
Imagine an
avid baseball fan who frequently visits fantasy baseball and other sports
Web sites. This user later visits an airline comparison shopping Web site,
and peruses flight options to Baltimore. Next, the user visits the New
York Times Travel Section online, and reads an article about what to do
during a weekend visit to Baltimore. An advertiser selling Baltimore Ori-
oles tickets would love to target this user for an offer for baseball tickets.
The ticket seller is likely to pay an advertising network a premium to
deliver its ad to a user matching this profile, based on the users collec-
39
Diana Dietrich, Google/Doubleclick: Is Privacy an Antitrust Concern (Not to Mention What is
the FTC Doing About Behavioral Advertising)?, SECURE TIMES, Spring/Summer 2008, at 2.
40
This example is based on similar examples contained in the FTC STAFF BEHAVIORAL
ADVERTISING REPORT, supra note 38, at 34.
782 ANTITRUST LAW JOURNAL [Vol. 76
tive interests and the high likelihood that he will be interested in the
content of the ad.
This type of behavioral targeting may be highly efficient, in terms of
finding the right eyeballs for an ad, and matching Internet users with
content and opportunities relevant to their interests. But behavioral ad-
vertising also raises extensive questions about the boundaries of privacy
and, perhaps more importantly, consumers expectations and assump-
tions regarding their privacy.
41
It is unclear whether consumers fully un-
derstand how the business model works, in terms of the price they pay
for free content and services. It is also difficult to quantify the cumula-
tive consequences of sharing dataconsequences that may be difficult
to reverse if a consumer ultimately decides that the revenue model is
not to her liking,
42
and that may be magnified as data-gathering technol-
ogies infiltrate more of our daily lives.
43
As illustrated by Facebooks re-
41
The FTC Staff Behavioral Advertising Report described the agencys ongoing examina-
tion of behavioral advertising, including its implications for consumer privacy. See, e.g., id.
at 10 (the invisibility of the practice to consumers raises privacy concerns, as does the risk
that data collected for behavioral advertisingincluding sensitive data about children,
health, or financescould be misused). The report also set forth revised principles to
govern self-regulatory efforts in this area. Id. at 45.
The Commission recently initiated a series of roundtables to explore privacy issues
more broadly, as discussed infra note 89. Fed. Trade Commn, Exploring Privacy: A
Roundtable Series, http://www.ftc.gov/bcp/workshops/privacyroundtables/index.shtml
[hereinafter FTC Privacy Roundtables].
42
For a more detailed discussion of these concerns and risks, see Harbour Behavioral
Advertising Concurrence, supra note 38.
There may be a tipping pointa point where consumers become sufficiently
concerned about the collection and use of their personal information that they
want to exercise greater control over it, but where any such attempt to exercise
control becomes futile because so much of their digital life already has been
exposed.
Id. at 4.
43
These concerns may be exacerbated as consumers rapidly adopt cloud computing,
mobile devices, and other data-focused technologies. In the cloud and mobile environ-
ments, not only are increasing volumes of data being shared and collected, but a broader
range of data may be available to fuel targeting efforts. Mobile technologies may enable
consumers to be reached by targeted ads throughout the day, even at times when consum-
ers are not sitting down and using traditional computers. In addition, new technologies,
such as smart grid electricity meters that capture detailed information about individual
energy consumption use, may further expand the data pool and thereby raise their own
unique privacy issues. See, e.g., INFO. & PRIVACY COMMR, ONTARIO, & FUTURE OF PRIVACY
FORUM, SMARTPRIVACY FOR THE SMART GRID: EMBEDDING PRIVACY INTO THE DESIGN OF
ELECTRICITY CONSERVATION 3 (Nov. 2009) [hereinafter SMARTPRIVACY FOR THE SMART
GRID] (while smart grid planning efforts are laudable, [w]e must take great care not to
sacrifice consumer privacy amidst an atmosphere of unbridled enthusiasm for electricity
reform.), available at http://www.ipc.on.ca/images/Resources/pbd-smartpriv-smartgrid.
pdf.
2010] SECTION 2 IN A WEB 2.0 WORLD 783
cent experience when it modified its privacy settings,
44
as well as
Googles recent conversion of email accounts into social networking ac-
counts,
45
a delicate balance exists between, on the one hand, the bene-
fits of sharing information, and on the other hand, the desire to
maintain control over data dissemination and use.
II. DATA MARKETS IN CYBERSPACE
With this background in mind, we will explore situations where it
might be appropriate and useful to define an input market for data it-
selfnot just a market for the services fueled by the data.
A. GOOGLE
Commissioner Harbour, an author of this essay, first articulated a po-
sition regarding data markets in her dissenting statement
46
when the
Commission approved the merger of Google and DoubleClick in De-
44
A strong public reaction erupted when, in late 2009 and early 2010, Facebook modi-
fied its privacy policy and the privacy settings available to users. Press Release, Facebook,
Facebook Asks More than 350 Million Users Around the World to Personalize Their Pri-
vacy (Dec. 9, 2009), available at http://www.facebook.com/press/releases.php?p=133917.
A backlash quickly ensued. See, e.g., Jaikumar Vijayan, Facebook Hit With Class Action over
Privacy Changes, COMPUTERWORLD.COM, Feb. 16, 2010, http://www.computerworld.com/s/
article/9157758/Facebook_hit_with_class_action_over_privacy_changes?source=rss_
news; Brian Womack, Facebooks New Information-Sharing Options Attract Criticism, BLOOM-
BERG.COM, Dec. 12, 2009, http://www.bloomberg.com/apps/news?pid=20601103&sid=A_
311Tx4BFwM; Kevin Bankston, Facebooks New Privacy Changes: The Good, The Bad, and The
Ugly, ELECTRONIC FRONTIER FOUNDATION DEEP LINKS BLOG, Dec. 9, 2009, http://www.eff.
org/deeplinks/2009/12/facebooks-new-privacy-changes-good-bad-and-ugly. In the
United States, the Electronic Privacy Information Center (EPIC) filed a complaint and
supplement with the Commission, alleging that Facebooks revised privacy settings vio-
lated user expectations, diminished user privacy, and contradicted Facebooks own repre-
sentations. Electronic Privacy Information Center, In re Facebook, http://epic.org/
privacy/inrefacebook/ (listing materials relating to the EPIC complaint). In Canada, the
Office of the Privacy Commissioner, prompted by a complaint from the Canadian In-
ternet Policy and Public Interest Clinic at the University of Ottawa, relaunched a previ-
ously closed investigation of Facebook; that investigation had culminated in the release of
a July 2009 report examining privacy issues raised by the Web site. Press Release, Office of
the Privacy Commissioner of Canada, Privacy Commissioner Launches New Facebook
Probe (Jan. 27, 2010), available at http://www.priv.gc.ca/media/nr-c/2010/nr-c_100127_
e.cfm; Press Release, Office of the Privacy Commissioner of Canada, Facebook Agrees to
Address Privacy Commissioners Concerns (Aug. 27, 2009), available at http://www.priv.
gc.ca/media/nr-c/2009/nr-c_090827_e.cfm.
45
Miguel Helft, Anger Leads to Apology from Google About Buzz, N.Y. TIMES, Feb. 14, 2010,
available at http://www.nytimes.com/2010/02/15/technology/internet/15google.html.
When initially launched, the Google Buzz social network automatically connected people
based on their email and chat contacts. After an angry response by users, Google altered
the service so that connections were suggested instead.
46
Dissenting Statement of Commissioner Pamela Jones Harbour, Google/DoubleClick,
FTC File No. 071-0170, (Dec. 20, 2007) [hereinafter Harbour Google Dissent], available at
http://www.ftc.gov/os/caselist/0710170/071220harbour.pdf.
784 ANTITRUST LAW JOURNAL [Vol. 76
cember 2007.
47
The competition analysis in Google/DoubleClick might
have proceeded differently if the Commission had taken a more dy-
namic and forward-looking approach to market definition, instead of
relying on a snapshot of current, narrowly-drawn product categories.
48
Even before the merger, it might have been argued that Google held a
significant market share in a possible market for data gathered via
search. Assuming such a product market were defensible, the Commis-
sion might have asked whether Googles acquisition of DoubleClick
would have substantially increased the likelihood that Google would ac-
quire or maintain market power in that market (a question that the
Commission majority did not address).
Furthermore, if that product market definition is valid, Googles
share of the search data market has only increased over time. The
Google search engine has become further entrenched as the dominant
search site,
49
and the firm has accumulated even more search data.
Given the role of network effects, one might wonder whether any other
firm will be able to chip away at Googles search supremacy without ac-
cess to a comparable trove of data.
50
47
Press Release, Fed. Trade Commn, Federal Trade Commission Closes Google/
DoubleClick Investigation (Dec. 20, 2007), available at http://www.ftc.gov/opa/2007/12/
googledc.shtm; Statement of Federal Trade Commission, Google/DoubleClick, FTC File
No. 071-0170 (Dec. 20, 2007), available at http://www.ftc.gov/os/caselist/0710170/0712
20statement.pdf; see also Concurring Statement of Commissioner Jon Leibowitz, Google/
DoubleClick, FTC File No. 071-0170 (Dec. 20, 2007), available at http://www2.ftc.gov/
os/caselist/0710170/071220leib.pdf.
48
In particular, Commissioner Harbour disagreed with the Commission majoritys de-
cision to treat search and display advertising as fundamentally different productscom-
plements versus substitutes, leading to a primarily vertical characterization and an
abbreviated entry analysiseven though evidence strongly suggested that these two types
of advertising already were converging, the two firms were fast becoming horizontal com-
petitors, and the acquisition might create market power. See Harbour Google Dissent,
supra note 46, at 57. She also disagreed with the Commission majoritys distinctions be-
tween premium and remnant Internet advertising space and between direct and interme-
diated sales because she believed that the combination of network effects and
technological advances soon would blur these distinctions. Id. at 78.
49
In January 2010, it was estimated that Google had a market share of over 65 percent
of searches conducted; Yahoo! was a distant second with 17 percent, and Microsofts Bing
had about 11 percent. Press Release, comScore, comScore Releases January 2010 U.S.
Search Engine Rankings (Feb. 11, 2010), http://comscore.com/Press_Events/
Press_Releases/2010/2/comScore_Releases_January_2010_U.S._Search_Engine_Rank-
ings. Microsofts May 2009 launch of Bing appears to have had little effect on search
engine market shares. See Press Release, comScore, comScore Releases May 2009 U.S.
Search Engine Rankings (June 17, 2009), http://comscore.com/Press_Events/Press_Re-
leases/2009/6/comScore_Releases_May_2009_U.S._Search_Engine_Rankings (Google at
65 percent, Yahoo at 20 percent, Microsoft at 8 percent).
50
One might argue that the existence of network effects heightens the incentives to
develop completely new technological solutions (as opposed to incremental improve-
ments of existing technologies) because of the winner-take-all nature of the market. If
2010] SECTION 2 IN A WEB 2.0 WORLD 785
Alternatively, one might define a somewhat broader market, such as
data used for behavioral advertising. This market would include not
only search data, but also data gathered from other sources and applica-
tions that offer clues regarding consumer preferences. Given Googles
extensive reach throughout the cloud, and the popularity of various
Google applications, Google likely has amassed the largest collection of
data that can be used for behavioral advertising. As a result, Google-
affiliated advertising spaces are viewed as very attractive and valuable
locations for advertisers to place their ads, such that advertisers may be
reluctant to spend their limited advertising dollars elsewhere. Existing
competitors and potential entrants may find it difficult to generate suffi-
cient advertising revenue to survive in competition with Google. Mean-
while, Google will continue to amass still greater amounts of search and
application data, which will further strengthen network effects and ham-
per entry. Over time, as a result of its data ownership, Google might be
uniquely positioned to control not only the data market, but also the
related market for behavioral advertising itself.
We are not asserting that Google actually has attained a dominant
position in one or more properly defined relevant product markets.
And even if Google has reached such a milestone, we certainly do not
mean to condemn the firm for having achieved success by innovating
and marketing superior products.
51
Nor do we mean to suggest that
Google has engaged, or is now engaging, in any exclusionary conduct
that would violate Section 2. Our main point is that, if antitrust enforc-
ers wish to evaluate the conduct of Google (or other successful Internet
one of these new technologies proves superior, its adoption is likely to effectuate a para-
digm shift and thereby eliminate any competitive problems in the original market. In
theory, if such a paradigm shift could occur rapidly, the risk of competitive harm would
be small indeed. Presumably, antitrust enforcers would encourage such paradigm-shifting
innovation, and would shape their enforcement approach accordingly. One such exam-
ple might be the DOJs recent decision to allow the proposed Internet search and paid
search advertising agreement between Microsoft and Yahoo!, based on a prediction that
the transaction would accelerate competition-enhancing innovation. See infra text accom-
panying notes 5254.
In contrast, enforcers might be expected to intervene when innovation may be sup-
pressed. Harm to competition may be especially likely if, for example, it would take sev-
eral years for a new technology to shake up and reshape the market. During that time,
any interim harm suffered by consumers (due to potential exclusionary conduct by the
incumbent dominant firm) might be more likely to outweigh any efficiencies generated
by network effects.
51
See United States v. Grinnell Corp., 384 U.S. 563, 57071 (1966) (Section 2 targets
the willful acquisition or maintenance of [monopoly] power as distinguished from
growth or development as a consequence of a superior product, business acumen, or
historic accident.); United States v. Aluminum Co. of Am., 148 F.2d 416, 430 (2d Cir.
1945) (antitrust law does not condemn a successful competitor who has attained a mo-
nopoly by virtue of his superior skill, foresight and industry).
786 ANTITRUST LAW JOURNAL [Vol. 76
firms) through a Section 2 (or EU Article 82, now Article 102) lens,
defining data markets may be a useful toolespecially if the enforcers
seek (as they should) to craft an analytical framework that accurately
reflects how these firms, their competitors, and their customers interact
in the real world.
It appears that the Department may have been leaning towards this
approach in its recent analysis of the Microsoft/Yahoo! transaction, as
evidenced by the public statement explaining the DOJs rationale for
closing its investigation of the parties proposed search and advertising
agreement. The statement highlighted the unusual relationship be-
tween scale and competitive performance in the search and paid search
advertising industry, and explained how the quality of Microsofts search
algorithms would be enhanced via access to a larger set of queries as a
result of the transaction.
52
The Department predicted that Microsofts
larger data pool may enable more effective testing and thus more rapid
innovation of potential new search-related products, changes in the
presentation of search results and paid search listings, other changes in
the user interface, and changes in the search or paid search algo-
rithms.
53
The statement suggested that this enhanced performance . . .
should exert correspondingly greater competitive pressure in the mar-
ketplace,
54
presumably creating a stronger number two firm to chal-
lenge the market leader Google.
As a matter of practical significance, early-stage product market defi-
nition affects the output of any government investigation.
55
In conduct
as well as merger investigations, antitrust enforcers typically issue sub-
poenas or other compulsory process to obtain documents, data, and
narrative responses. Tentative product market definitions will be incor-
porated into the subpoenas, and these definitions will affect the struc-
ture and scope of the entire investigation. If data troves are not squarely
included within the scope of compulsory process (separate from the
products and services that rely on the data), it may become difficult for
enforcement staff to pursue data-driven theories as the case evolves be-
cause the staff may be less able to evaluate other potential or intended
52
Press Release, U.S. Dept of Justice, Statement of the Department of Justice Antitrust
Division on Its Decision to Close Its Investigation of the Internet Search and Paid Search
Advertising Agreement Between Microsoft Corporation and Yahoo! Inc. (Feb. 18, 2010),
available at http://www.justice.gov/atr/public/press_releases/2010/255377.pdf.
53
Id.
54
Id.
55
Accord Dietrich, supra note 39, at 45 (discussing short-term practical implications if
government investigations more closely scrutinize data usage, and suggesting how outside
parties and counsel might respond to this new category of interest).
2010] SECTION 2 IN A WEB 2.0 WORLD 787
uses for the data.
56
For this reason, enforcers who are developing investi-
gative plans involving Internet firms should think creatively and con-
sider the inclusion of one or more data markets, such as data gathered
via search, data used for behavioral advertising, or other types of data
markets that might be suggested by case-specific facts.
B. PRECEDENT FOR DATA MARKETS: OTHER CASES
When Commissioner Harbours Google/DoubleClick dissent was issued,
some critics suggested that the data market concept did not reflect
mainstream antitrust analysis. This critique is inconsistent with several
significant merger cases that, for all intents and purposes, have been
about data markets. While there may not yet have been a pure Section 2
case based squarely on data markets, the concept is defensible and is
consistent with past agency practice.
1. CCC/Mitchell
One such example is the CCC/Mitchell merger case,
57
where the Com-
mission obtained a preliminary injunction to block a proposed merger
between competitors in the markets for estimatics (electronic systems
used to estimate the cost of automobile collision repairs) and total loss
valuation systems (software systems used to value passenger vehicles that
have been totaled). Insurance companies are the primary customers of
these systems.
56
Commissioner Harbours Google/DoubleClick dissent raised similar concerns. It sug-
gested that if the Second Requests in the merger investigation had included data markets,
the staff might have been better equipped to probe the parties intentions for their com-
bined data troves. These answers would have been relevant to the antitrust analysis, might
have further informed the consumer protection inquiry regarding privacy implications
and presumably would have had some binding effect on the parties. Instead, the staffs
investigation relied on the parties voluntary and non-binding representations about fu-
ture data usage. Harbour Google Dissent, supra note 46, at 9.
57
Press Release, Fed. Trade Commn, FTC Launches Suit to Block Merger of CCC and
Mitchell (Nov. 25, 2008), available at http://www.ftc.gov/opa/2008/11/cccmitchell.shtm;
FTC v. CCC Holdings Inc., No. 08-2043-RMC (D.D.C. filed Mar. 18, 2009) (Memorandum
Opinion) [hereinafter CCC Opinion] (granting preliminary injunction), available at http:/
/www.ftc.gov/os/caselist/0810155/090309cccmitchellpublicopinion.pdf. Case materials
(federal court proceedings) are available at http://www.ftc.gov/os/caselist/0810155/
index.shtm.
The Commission also issued a Part 3 complaint in a parallel administrative proceeding.
CCC Holdings Inc. & Aurora Equity Partners III L.P., FTC Docket No. 9334 (Nov. 25,
2008) (Complaint). The Part 3 case was dismissed following the district courts granting
of a preliminary injunction and the parties subsequent abandonment of the proposed
merger. CCC Opinion, supra. Case materials (administrative proceedings) available at
http://www.ftc.gov/os/adjpro/d9334/index.shtm.
788 ANTITRUST LAW JOURNAL [Vol. 76
The complaint alleged relevant product markets for systems.
58
How-
ever, the key assets that would have been combined were extensive pro-
prietary databases of parts and labor costs, going back many years, for
virtually all vehicles on the road. As alleged by the Commissions com-
plaint and found by the court, it would be extremely costly and time-
consuming for another firm to create a comparable database.
59
The par-
ties did offer software systems used to analyze and process the data, as
well as other related services. But the software and services were secon-
dary in the Commissions analysis, in large part because they would be
easier for a potential entrant to replicate. The rich trove of historical
data drove the success of each of the merging firms, and the lack of
access to a comparable database was viewed as a significant entry barrier.
2. Thomson/Reuters
Another example is Thomson/Reuters,
60
a DOJ merger settlement from
February 2008. According to the complaint and other public docu-
ments, the parties were two of a very few firms that sold financial data
used by investment managers, investment bankers, traders, and other
institutional customers. These data were used by customers to make in-
vestment decisions and to provide advice to their firms and clients. Spe-
cifically, the DOJ complaint alleged the following three relevant product
markets: (1) fundamentals data; (2) earnings estimates data; and (3)
aftermarket research reports.
61
58
FTC v. CCC Holdings Inc., No. 08-2043-RMC, Complaint 15 (D.D.C. filed Nov. 26,
2008) (Complaint), available at http://www.ftc.gov/os/caselist/0810155/081126
cmpltforpitro.pdf.
59
Among the significant barriers to entry in the relevant markets, as cited in the
FTCs complaint, were the substantial time and expense required to assemble, edit, and
maintain estimatics and TLV [total loss valuation] databases relating to virtually every
vehicle driven in the United States, as required by customers. Id. 25. The district court
agreed:
The difficulty and cost of developing and maintaining an entirely new parts and
labor database that is accepted by the market would be significant barriers to
new entrants. . . . It would take a number of years, untold thousands of man-
hours, and millions of dollars of investment to create and maintain a competi-
tive parts and labor database.
CCC Opinion, supra note 57, at 3839.
60
Press Release, U.S. Dept of Justice, Justice Department Requires Thomson to Sell
Financial Data and Related Assets in Order to Acquire Reuters (Feb. 19, 2008) [hereinaf-
ter Thomson/Reuters News Release], available at http://www.usdoj.gov/atr/public/press
_releases/2008/230250.pdf.
61
United States v. Thomson Corp., No. 08-0262, Complaint at 1431 (D.D.C. filed
Feb. 19, 2008), available at http://www.usdoj.gov/atr/cases/f230200/230281.pdf.
2010] SECTION 2 IN A WEB 2.0 WORLD 789
As part of the settlement, Thomson, the acquiring firm, was required
to divest copies of three financial datasets. The DOJ announcement
62
specifically noted that the remedies imposed by the Department were
consistent with remedies obtained by the European Commission (EC)
in a parallel settlement,
63
which may provide some insight regarding
how the EC might view the concept of a data market
3. Hearst/First DataBank
An older, but still relevant, example dates back to December 2001,
when the Commission accepted a landmark settlement with The Hearst
Trust to resolve charges stemming from an April 2001 complaint.
64
The
complaint had alleged that Hearst illegally acquired a monopoly in the
market for electronic integratable drug information databases, which
are used by pharmacies, hospitals, doctors, payors, and patients to ob-
tain information about drug prices, drug effects, drug interactions, and
drug eligibility under various payment plans.
According to the complaint, after unlawfully acquiring its sole com-
petitor in this market, Hearst drastically raised prices, and virtually all
customers acceded to the price increases. Despite these enormous
price increases, three years later there had been no significant new entry
into the relevant product market.
65
The structural relief portion of the
remedy required Hearst to divest (among other things) a copy of the
entire electronic integratable drug information database that Hearst
had acquired, along with access to information needed for updates.
66
4. Ticketmaster/Live Nation
Another example is the combination of Ticketmaster and Live Na-
tion, which was the subject of a DOJ antitrust investigation and a settle-
ment that includes both structural and behavioral relief. The DOJ
62
Thomson/Reuters News Release, supra note 60, at 2.
63
Press Release, European Commn, Mergers: Commission Clears Acquisition of
Reuters by Thomson Subject to Conditions (Feb. 19, 2008) (IP/08/260). Case materials
available at http://europa.eu/competition/mergers/cases/index/m94.html.
64
Press Release, Fed. Trade Commn, Hearst Corp. to Disgorge $19 Million and Divest
Business to Facts and Comparisons to Settle FTC Complaint (Dec. 14, 2001), available at
http://www.ftc.gov/opa/2001/12/hearst.shtm; see also Press Release, Fed. Trade
Commn, FTC Charges Hearst Trust with Acquiring Monopoly in Vital Drug Information
Market (Apr. 4, 2001), available at http://www.ftc.gov/opa/2001/04/hearst.shtm; FTC v.
The Hearst Trust, No. 01-00734 (D.D.C. filed Apr. 15, 2001) (Complaint) [hereinafter
Hearst Trust Complaint], available at http://www.ftc.gov/os/2001/04/hearstcmp.htm.
65
Hearts Trust Complaint, supra note 64, 3738.
66
Final Order and Stipulated Permanent Injunction, FTC v. The Hearst Trust, No. 01-
00734 (D.D.C. filed Dec. 18, 2001), available at http://www.ftc.gov/os/2001/12/hearst
finalorder.pdf.
790 ANTITRUST LAW JOURNAL [Vol. 76
enforcement action focused on protecting competition in a market for
primary ticketing services to major concert venues, a market in which
Ticketmaster already was a leader and Live Nation was an emerging
threat.
67
While a data market does not appear to have been formally
defined, in crafting the relief the DOJ explicitly acknowledged the im-
portance of data as a valuable competitive asset.
Notably, the Proposed Final Judgment requires the merged firm to
provide current clients with their ticketing data promptly upon request,
in a reasonably usable form, if clients wish to move their business to
another primary ticketing firm.
68
The DOJs Competitive Impact State-
ment explains that this data portability provision reduces venues
switching costs and lowers barriers to other companies competing for
Defendants primary ticketing clients because it ensures that those
venue clients will not be forced to relinquish valuable data if they decide
to switch primary ticketing service providers.
69
This aspect of the relief
addresses the elimination of horizontal competition between estab-
lished Ticketmaster and newcomer Live Nation in the market for pri-
mary ticketing services. With respect to this horizontal overlap as it
existed at the time of the merger, the remedy arguably would have been
the same if the DOJ had chosen to define a data market comprising, for
example, data used for primary ticketing. Defining a ticketing data mar-
ket, however, might have enabled the remedy to reach even further,
covering expanded uses of historical ticketing data that would have gen-
erated further horizontal competition between the parties in the future.
The Proposed Final Judgment also addresses a potential, primarily
vertical, competitive concern relating to the practice of data mining in
ticket sales.
70
As explained in one press report that arose when the trans-
action was first announced, concert promoters mine data from past
performances, and use this information when bidding against rivals for
future performances. For instance, promoters look at detailed informa-
67
Press Release, U.S. Dept of Justice, Justice Department Requires Ticketmaster En-
tertainment Inc. to Make Significant Changes to Its Merger with Live Nation Inc. (Jan. 25,
2010), available at http://www.justice.gov/atr/public/press_releases/2010/254540.pdf;
Competitive Impact Statement, United States v. Ticketmaster Entertainment, Inc. & Live
Nation, Inc., No. 10-00139 (D.D.C. filed Jan. 25, 2010) [hereinafter Ticketmaster Competi-
tive Impact Statement], available at http://www.justice.gov/atr/cases/f254500/254544.
pdf.
68
[Proposed] Final Judgment IX.C, United States. v. Ticketmaster Entertainment,
Inc. & Live Nation, Inc., No. 10-00139 (D.D.C. filed Jan. 25, 2010), available at http://
www.justice.gov/atr/cases/f254500/254558.pdf.
69
Ticketmaster Competitive Impact Statement, supra note 67, at 17.
70
Adam Satariano, Live Nation Rivals See Hard Rain from Dylan Data, BLOOMBERG.COM,
Mar. 4, 2009, http://www.bloomberg.com/apps/news?pid=20601103&sid=AZ.ptvJ54zIU&
refer=US.
2010] SECTION 2 IN A WEB 2.0 WORLD 791
tion regarding how fast tickets sell and at what prices, the most popular
time of year for concerts, the ages and addresses of ticket buyers, and
perhaps even their contact information. Historically, these types of data
have been provided to promoters by Ticketmaster, the largest ticketing
system. After combining with Live Nation, the biggest concert promoter,
Ticketmaster might not have the same incentives to share its vast collec-
tion of ticketing data with rival concert promoters. Moreover, if Tick-
etmasters incentives were to change as a result of the merger, it is
unclear whether any other market participant would be able to amass or
offer a dataset as extensive as Ticketmasters.
71
The Proposed Final Judg-
ment addresses this concern by restricting the combined firms ability to
use its valuable ticketing data in its promotion and artist management
lines of business, unless those data are made available on the same terms
to other promoters or artist managers not affiliated with Ticketmaster.
As explained in the Competitive Impact Statement, this provision will
prevent the combined Ticketmaster/Live Nation from abusing their
position in the primary ticketing market to impede competition among
promoters and artist managers.
72
With respect to the data mining concerns, the DOJ investigation ap-
pears to have yielded sufficient evidence to justify relief based on vertical
foreclosure concernsnamely, Ticketmasters changed incentives as a
primary ticketer upon vertically integrating with a concert promoter. If
the evidence had not been strong enough to support a vertical theory of
harm, a data market definition might have provided a horizontal hook
to address the data mining concerns. The DOJ might have defined a
market comprising the frequently-mined historical ticketing data them-
71
The Ticketmaster/Live Nation Merger: What Does It Mean for Consumers and the Future of the
Concert Business?: Hearing Before the S. Comm. on the Judiciary, Subcomm. on Antitrust, Competi-
tion Policy and Consumer Rights, 111th Cong. (Feb. 24, 2009) (statement of Seth Hurwitz,
Co-Owner, I.M.P. Productions and 9:30 Club, Washington, DC), available at http://judici-
ary.senate.gov/hearings/testimony.cfm?id=3674&wit_id=7625:
[T]here are situations where I may be forced to use TicketMaster, either by a
venue contract, or perhaps where TicketMaster would be the unquestionably
better provider of ticket service. And here is the big problem with that: If this
merger is allowed to happen, my biggest competitor will have access to all of my
sales records, customer information, on sale dates for tentative shows, my ticket
counts, they can control which shows are promoted and much more. This will
put ALL independent promoters at an irreparable competitive disadvantage.
This would be like Pepsi forcing Coke to use its services as its distributor, and
pretend that the intelligence Pepsi gathers wont harm Coke. It just cant hap-
pen and maintain a fair and level playing field.
72
Ticketmaster Competitive Impact Statement, supra note 67, at 17.
792 ANTITRUST LAW JOURNAL [Vol. 76
selves, and imposed relief provisions targeted directly at preserving cur-
rent and future competition in that market.
73
III. MARKETS RELATED TO PRIVACY PROTECTION
In addition to data markets, there may be other types of relevant
product markets that make particular sense in cyberspace. In her Google/
DoubleClick dissent, Commissioner Harbour raised the idea of competi-
tion on privacy dimensions, and suggested that privacy should be cogni-
zable under the antitrust laws.
74
Defining relevant product markets that
recognize privacy as a dimension of competition might provide the miss-
ing link for such cognizability.
A. PRIVACY PROTECTION AS A FORM OF NON-PRICE COMPETITION
It is widely accepted that firms compete not only on price, but also on
various non-price dimensions that are important to consumers. In a
thought-provoking 2001 essay, then-FTC Commissioner Thomas Leary
discussed the challenges of defining relevant product markets when
dealing with differentiated product markets that are influenced by indi-
vidual consumer tastes.
75
Commissioner Leary questioned whether tradi-
tional models of market definition would continue to make sense as the
assumption of homogeneous commodity products becomes progres-
sively less realistic in the real world of non-homogeneous products, ser-
vices, and experiences.
76
He warned, however, that it would be equally
unappealing to ignore potentially growing areas of antitrust concern
simply because the traditional approaches are inadequate.
77
A similar message may apply to the antitrust analysis of privacy protec-
tions.
78
It would be entirely inappropriate to ignore consumers con-
73
Lacking the detailed confidential information available to the DOJ, and with great
respect for the comprehensive settlement the DOJ obtained in this matter, we do not
claim that our data market approach would have yielded better relief. We raise the Tick-
etmaster/Live Nation example primarily to demonstrate that enforcers should think cre-
atively about market definition when facing a situation where competitive harm seems
likely, but traditional market definitions do not appear to provide a path toward needed
relief.
74
Harbour Google Dissent, supra note 46, at 911.
75
See Thomas B. Leary, The Significance of Variety in Antitrust Analysis, 68 ANTITRUST L.J.
1007 (2001).
76
Id. at 1008.
77
Id. at 100809.
78
See Peter P. Swire, Testimony Submitted to the Federal Trade Commission Behavioral
Advertising Town Hall (Oct. 18, 2007), available at http://www.ftc.gov/os/comments/
behavioraladvertising/071018peterswire.pdf. Professor Swire points out that as more data
are collected, individuals with high privacy preferences may perceive a significant reduc-
tion in the quality of the search product. Echoing themes set forth by Commissioner
2010] SECTION 2 IN A WEB 2.0 WORLD 793
cerns about privacy-based competition, simply because product market
definition might prove difficult. As demonstrated by recent studies, on-
line privacy is an important issue for many consumers, especially with
regard to targeted behavioral advertising.
79
Moreover, consumer aware-
ness of privacy issues continues to grow, driven in large part by enforc-
ers increased scrutiny and consumer education efforts, which have led
firms to improve transparency regarding their privacy policies. Appar-
ently, the online firms are listeningmany of the biggest Internet
names publicize their privacy policies as a way to attract and retain users.
Even more importantly, these firms react directly to each others privacy
policy changes. At one point in 2008, Google, Yahoo!, and Microsoft
each shortened the amount of time they would retain personal data
gathered from users Web surfing.
80
Interestingly, Microsoft announced
that it would anonymize its data after six monthscompared to the
firms then-existing eighteen-month policybut only if its rivals would
Leary, Professor Swire contends that this sort of quality reduction is a logical component
of antitrust analysis. Id. at 5.
Professor Robert Lande has expressed similar thoughts.
Antitrust is actually about consumer choice, and price is only one type of choice.
The ultimate purpose of the antitrust laws is to help ensure that the free market
will bring to consumers everything they want from competition. This starts with
competitive prices, of course, but consumers also want an optimal level of vari-
ety, innovation, quality, and other forms of nonprice competition. Including pri-
vacy protection.
Robert H. Lande, The Microsoft-Yahoo Merger: Yes, Privacy Is an Antitrust Concern,
FTC:WATCH, Feb. 25, 2008, at 1; see also Averitt & Lande, supra note 14, at 713 (Con-
sumer sovereignty exists when two fundamental conditions are present. There must be a
range of consumer options made possible through competition, and consumers must be
able to choose effectively among these options.).
79
See Joseph Turow et al., Americans Reject Tailored Advertising and Three Activities
that Enable It (Sept. 29, 2009) (majority of consumers would prefer not to be subjected to
behavioral targeting, including 66 percent overall and 55 percent of 18-to-24 year olds;
significantly higher percentages of people, between 7386 percent, object to specific types
of tailored advertising), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=
1478214; see also Stephanie Clifford, Two-Thirds of Americans Object to Online Tracking, N.Y.
TIMES (Sept. 29, 2009) (discussing Turow study), available at http://www.nytimes.com/
2009/09/30/business/media/30adco.html; Press Release, TRUSTe Behavioral Targeting:
Not that Bad?! TRUSTe Survey Shows Decline in Concern for Behavioral Targeting; Consumers
Want Relevant Ads Online, But Still Worry About Their Online Privacy (Mar. 4, 2009) (while
consumer discomfort with behavioral advertising declined from 57 percent to 41 per-
cent between 2008 and 2009, 72 percent of those surveyed said that online advertising was
intrusive and annoying when they were faced with irrelevant ads); Stephanie Clifford,
Many See Privacy on Web as Big Issue, Survey Says, N.Y. TIMES, Mar. 16, 2009, at B5, (report-
ing results of TRUSTe survey; noting conclusion that the vast majority of Americans view
online privacy as a really or somewhat important issue), available at http://www.ny-
times.com/2009/03/16/technology/internet/16privacy.html.
80
Kim Dixon, Yahoo Cuts Data Retention to Three Months, REUTERS, Dec. 17, 2008, availa-
ble at http://www.reuters.com/article/technologyNews/idUSTRE4BG2VP20081217.
794 ANTITRUST LAW JOURNAL [Vol. 76
follow suit.
81
Yahoo! subsequently announced that it would retain data
for only three months, albeit according to a different anonymization
standard.
82
And one industry commentator noted that first-mover
Google had started this competition, putting other firms in a position
where they needed to respond.
83
To bring this discussion back to potential Section 2 applications, con-
sider whether a dominant or potentially dominant firm would have the
same incentives to adapt its privacy policieseither in response to con-
sumer demand or as a reaction to competition from other firms. If
achieving a dominant market position might change the firms incen-
tives to compete on privacy dimensions, this is a consequence that anti-
trust enforcers might wish to explore further.
84
Defining a privacy-based
relevant product market would be one way to frame the analysis accord-
ing to traditional antitrust principles, while still accounting for the real-
world priorities of todays Internet consumers.
85
B. INNOVATION COMPETITION
As suggested by the different anonymization standards applied by the
big Internet firms, another form of privacy-related competition may in-
volve the underlying technologies used to protect privacy.
The use of consumer data for behavioral advertising and similar appli-
cations can be highly efficient. But these efficiencies must be reconciled
with the demand for privacy protection. Currently, there is innovation
competition to introduce technological solutions that strike an accept-
able balance between these two objectives, and the pace of such innova-
81
Id.; see also Thomas Claburn, Microsoft Reduces Search Data Retention to Six Months, INFO.
WK., Dec. 8, 2008, available at http://www.informationweek.com/news/internet/search/
showArticle.jhtml?articleID=212300250.
82
Dixon, supra note 80. Yahoo! said it would delete the final segment of the Internet
Protocol (IP) address to make it no longer unique or identifiable, while Microsoft de-
cided to delete the entire IP address to cut off any link to identifying information.
83
Id. (quoting Ari Schwartz, Vice President, Center for Democracy and Technology).
84
See Dietrich, supra note 39, text accompanying note 36 (setting forth a hypothetical
scenario of how a merger might affect incentives to compete based on privacy policies).
85
Privacy-based relevant product markets might also provide a way to reconcile the fact
that consumerswhose data are collected, and whose privacy is at stakefrequently are
not the customers of the services that exploit data troves. See Harbour Google Dissent,
supra note 46, at 10 (because individual consumers had no financial or business relation-
ship with Google and DoubleClick, the majoritys product market approach did not ade-
quately reflect the values of the consumers whose data were at stake). But see Dietrich,
supra note 39, at 56 (mainstream antitrust thinking unlikely to make a radical depar-
ture from accepted antitrust analysis to embrace a privacy-interest proxy that circum-
vents the disconnect between the interests of data-supplying users and those of potentially
impacted customers).
2010] SECTION 2 IN A WEB 2.0 WORLD 795
tion is likely to be accelerated in the coming years.
86
While there is not
yet a clear consensus on the optimal level of competition to incentivize
innovation,
87
intuitively it would seem that innovation competition is
more likely to be robust in a competitive market, where multiple firms
are racing to commercialize their innovations, win market share, and
reap profits.
88
Again, going back to the Section 2 context, it makes sense
to ask whether a dominant or potentially dominant firm is likely to have
the same incentives to develop, or otherwise seek out, innovative new
technologies for the efficient protection of privacy. Absent pressure
from competitors who might provide more attractive alternatives to pri-
vacy-prioritizing consumers, a dominant firm might rationally choose to
innovate less vigorously around privacy or, perhaps, to dole out privacy-
protective technologies to the marketplace more slowly.
C. EFFECT OF PRIVACY REGULATION ON COMPETITION
Professor Randal Picker opines that choices on how to regulate online
datastreams will have consequences beyond privacy.
89
As he explains, re-
strictions on the use of data will directly influence how much competi-
tion is able to emerge in related technology markets.
90
In other words,
decisions about privacy regulation affect not only privacy; these deci-
sions have broader competitive significance as well. Competing philoso-
86
See Harbour Behavioral Advertising Concurrence, supra note 38, at 45 (inadequate
privacy protection technologies, such as the opt-out cookie, offer the illusion of consumer
choice but are confusing and unreliable; the industry should focus its efforts on develop-
ing viable and transparent alternatives); see also id. at 8 (increased attention to privacy
will create opportunities for firms to develop attractive new privacy tools and to market
these features to distinguish themselves from competitors).
87
Compare, e.g., Statement of Commissioner Pamela Jones Harbour, Genzyme Corp.,
FTC File No. 021-0026, (Jan. 13, 2004) (laying out theoretical basis for favoring a general
presumption of anticompetitive effects . . . in the extreme case of a merger to monopoly
that eliminates all competition and diversity in [an] innovation market), with Statement
of Chairman Timothy J. Muris, Genzyme Corp., FTC File No. 021-0026, (Jan. 13, 2004)
(explaining lack of theoretical or empirical link between increased competition and re-
duced innovation) (citing COUNCIL OF ECONOMIC ADVISERS, ECONOMIC REPORT OF THE
PRESIDENT 176 (1999) (To the extent there is consensus, it is that neither the presence of
many competitors nor pure monopoly correlates systematically with optimal levels of
innovation.)).
88
Cf. SMARTPRIVACY FOR THE SMART GRID, supra note 43, at 7 (beyond sizeable public
investment in smart grid technology, significant private sector investment has been de-
voted to developing new products and services in a market projected to reach $100 billion
by 2030; venture capital valued at over $900 million was invested between 2000 and 2008).
89
Randal C. Picker, Competition and Privacy in Web 2.0 and the Cloud, 103 NW. U. L. REV.
COLLOQUY 1 (2008), http://www.law.northwestern.edu/lawreview/colloquy/2008/25/
LRColl2008n25Picker.pdf.
90
Id. at 3 (How we choose to regulate these datastreams is the central regulatory issue
of the emerging computer infrastructure. Our choices here obviously have privacy conse-
quences but also for how much competition will emerge. These are tightly linked.).
796 ANTITRUST LAW JOURNAL [Vol. 76
phies of privacy regulation, and resulting changes in the legal regime,
should play a role in establishing product market definitions based on
privacy protections.
Around the world, privacy laws are evolving in response to new learn-
ing about the psychology of consumer behavior, greater understanding
of consumer preferences, and deeper insights regarding the value of
consumer data, all buoyed by a steady undercurrent of technological
innovation and changing business practices.
91
Ideally, among various
factors, legislators will consider whether a given set of privacy regula-
tions would be likely to facilitate the creation or maintenance of a mo-
nopolyeither by entrenching a dominant firm, or by establishing
mechanisms that would enable a dominant firm to engage in exclusion-
ary conduct.
One example might be the issue of data portability. Imagine that a
given legal regime were to encourage greater consumer control over
data (e.g., through open standards), such that a market emerged to ac-
commodate the porting of data relatively easily among applications. In
that entry-friendly environment, if consumers were unhappy with the
level of privacy protection offered by a popular application or service,
consumers would be better able to vote with their feet (or, more accu-
rately, their data) and switch to competing providers, without losing the
accrued value of their personal datasets. Under those circumstances,
dominance would be hard to achieve and even harder to illegally main-
tain. But in a regime without easy data portability, it would be more
difficult for dissatisfied consumers to shop around for a competitor of-
fering a different balance between data exploitation and privacy
protection.
91
In the United States, for example, the FTC recently launched a roundtable series to
explore, and obtain public input on, various models for promoting consumer privacy,
consistent with core values of transparency, consumer control, and accountability that
should govern any approach to privacy. FTC Privacy Roundtables, supra note 39. A key
area of inquiry will be whether existing legal requirements and self-regulatory regimes
provide an adequate framework for the protection of consumer privacy interests, both
today and in the future. For example, it is possible that the well-accepted reliance on
notice and choice regimes, combined with harm-based approaches to enforcement, does
not work well in the realm of behavioral advertising. See David C. Vladeck, Director, Bu-
reau of Consumer Protection, Fed. Trade Commn, Promoting Consumer Privacy: Ac-
countability and Transparency in the Modern World, Keynote Address at the New York
University School of Law Information Law Institute Workshop on Federal Privacy Legisla-
tion (Oct. 2, 2009), available at http://www.ftc.gov/speeches/vladeck/091002nyu.pdf;
David C. Vladeck, Director, Bureau of Consumer Protection, Fed. Trade Commn, Pri-
vacy: Where Do We Go from Here?, Address at the International Conference of Data
Protection and Privacy Commissioners (Nov. 6, 2009), available at http://ftc.gov/
speeches/vladeck/091106dataprotection.pdf.
2010] SECTION 2 IN A WEB 2.0 WORLD 797
IV. CONCLUSION
It is our hope that the Commission and other antitrust enforcement
agencies will seriously consider the ramifications of Internet markets
driven by troves of datain the Web 2.0 world, the cloud computing
environment, and whatever comes next. Internet markets continue to
evolve rapidly, as Internet firms gain enhanced abilities to collect and
process vast quantities of data. We encourage the agencies to think cre-
atively about how best to enforce the antitrust laws in this environment.
One principled approach may involve defining data-based relevant
product markets that fully capture current and future marketplace reali-
ties. In the spirit of sound competition policy, the law should be en-
couraged to develop in ways that honor a dynamic approach to market
definition and analysis.
Second, we believe that it is important to recognize and explore the
nexus between competition and privacy, regardless of how privacy ulti-
mately is incorporated into antitrust analysis. Going forward, we hope
that the Commission will indeed leverage its greatest resourcethe ex-
pertise of its talented staffto take a more integrated approach to pri-
vacy and competition and to more aggressively pursue issues at the
intersection of privacy and competition in a variety of industries (includ-
ing, for example, health care and electricity markets). This approach
should include the definition of privacy-based relevant product markets,
where applicable, if such definitions will help to make privacy cogniza-
ble under the antitrust laws. The Commission is perfectly positioned to
capitalize on its dual competition and consumer protection expertise,
and fulfill its critical advocacy mission on behalf of consumers, as it
grapples with these cutting-edge issues.