Dunstan et al v. comScore, Inc.

Filing 88

MEMORANDUM Opinion and Order. Signed by the Honorable Young B. Kim on 3/2/2012. (aac, )

UNITED STATES DISTRICT COURT
NORTHERN DISTRICT OF ILLINOIS
EASTERN DIVISION
MIKE HARRIS and JEFF DUNSTAN,
individually and on behalf of a class of
similarly situated individuals,
Plaintiffs,
v.
COMSCORE, Inc.,
a Delaware Corporation,
Defendant.

)
)
)
)
)
)
)
)
)
)
)

Case No. 11 CV 5807

Magistrate Judge Young B. Kim

March 2, 2012

MEMORANDUM OPINION and ORDER
In this putative class action, Plaintiffs Mike Harris and Jeff Dunstan allege that
comScore, Inc. (“comScore”) improperly collected and disseminated personal information
belonging to them and a class of similarly situated individuals. They claim that comScore
violated the Stored Communications Act, 18 U.S.C. § 2701(a)(1) and (2), the Electronic
Communications Privacy Act, 18 U.S.C. § 2511(1)(a) and (d), the Computer Fraud and
Abuse Act, 18 U.S.C. § 1030(a)(2)(C), the Illinois Consumer Fraud and Deceptive Practices
Act, 815 Ill. Comp. Stat. Ann. 505/1 (2007), and was unjustly enriched by its actions. This
matter is before the court on the motion of comScore to bifurcate discovery into two
phases—the first relating to class certification, and the second relating to the merits—and to
stay discovery on the merits until further order of the court. For the following reasons, the
motion is granted:

Facts and Procedural History
Plaintiffs allege that comScore, an Internet market research company, induces
individuals to download and install its software by “bundling” it with free items such as
screensavers, games, and third-party computer applications like CD burning software or
greeting card templates. (R. 1, Compl. ¶¶ 13, 33.) After a user installs the comScore
software, the software surreptitiously collects information about the user’s online activity,
scans some of the user’s computer files, and transmits the information to comScore’s servers.
(Id. at ¶¶ 5, 9.) comScore then aggregates the data it mines from the individual users for the
purpose of developing market research reports, which it sells to its clients. (Id. at ¶ 25.)
Plaintiffs allege that they, as well as a class of similarly situated individuals, were
misled by comScore’s Terms of Service (“ToS”). (Id. at ¶ 37.) They allege that in some
cases, comScore’s ToS display screens do not refer to comScore’s full license agreement (id.
at ¶ 38), and in other cases, do not adequately alert consumers to a link containing
comScore’s license agreement (id. at ¶ 40). They further allege that comScore’s ToS and
Privacy Policy are incomplete because they fail to disclose pertinent information such as the
types of modifications that the comScore software will make to the user’s computer settings,
the breadth of personal data that the software will collect from the user’s computer, and the
possibility that the installation of the software will result in comScore scanning files on
computers found on the user’s local networks. (Id. at ¶¶ 7, 10, 11, 37, 51.) Moreover,

2

Plaintiffs allege that the comScore software, once installed, is difficult to remove. (Id. at
¶¶ 57, 58.)
Approximately two million internet users have installed comScore’s software. (Id.
at ¶¶ 31, 67, 70.) Harris alleges that when he downloaded a free screensaver, he unwittingly
installed comScore’s software, but did not agree to comScore’s ToS or understand that the
screensaver was bundled with comScore’s software. (Id. at ¶¶ 67-69.) Harris alleges that
he was able to uninstall the comScore software from his Macintosh computer but only after
“conducting hours of diligent research.” (Id. at ¶ 68.) Dunstan alleges that when he
downloaded a free greeting card template it was secretly bundled with comScore’s software.
(Id. at ¶ 70.) Dunstan alleges that his computer, which ran the Windows operating system,
was “debilitated” by the comScore software. (Id. at ¶¶ 70, 71.) Dunstan alleges that he had
to purchase and use a $40 anti-virus software to remove the comScore software and restore
his computer’s functionality. (Id. at ¶ 73.)
Pursuant to Federal Rules of Civil Procedure 23(b)(2) and (b)(3), Plaintiffs brought
this action on behalf of themselves and all other persons similarly situated. They seek to
represent a class consisting of “all individuals and entities in the United States that have had
comScore’s surveillance software (“Surveillance Software”) installed on their computer(s)
and a Subclass of all individuals and entities in the United States that have incurred costs in
removing the Surveillance Software.” (R. 2, Pls.’ Mot. to Certify Class at 2.)

3

Harris1 propounded his first set of interrogatories and document requests to comScore
in December 2011. (R. 67-1, 67-2.) comScore characterizes these requests as “sweeping and
intrusive . . . encompassing virtually all aspects of comScore’s business.” (R. 67, Def.’s
Mem. in Support of Bifurcation at 1.) comScore seeks to delay responding to these requests
except as they relate to class certification issues. (Id. at 11.) To that end, comScore has
agreed to respond to the interrogatories and requests for production that it believes relate to
class certification issues. It has already produced the source code for the software in dispute,
including the source code for its “RK Verify” software, which comScore describes as the
software “that confirms that consumers have viewed and agreed to comScore’s Terms of
Service before installing the software.” (Id. at 2.)
Analysis
comScore argues that it would be inefficient and wasteful to conduct merits discovery
before the assigned District Judge rules on the pending class certification motion. comScore
predicts that if class certification is denied, Plaintiffs will settle or voluntarily withdraw their
complaint because the statutory damages at issue—maximum $1,000 for Harris and $1,040
for Dunstan—are meager. In that event, any merits discovery already performed would be
a wasted, expensive effort. If, on the other hand, the court certifies a class or classes, the
certification order will clarify the issues to be litigated on the merits and will thereby narrow

1

The record does not explain why Dunstan’s name does not appear on the written requests
for discovery but this ruling applies to all parties in this case.
4

the scope of merits discovery. comScore suggests that the court might certify a class limited
to Macintosh users or limited to Windows users because the comScore software for the two
operating systems differs for at least two reasons: (1) comScore sold the data it collected
from Windows users, but not the data collected from Macintosh users (see R. 59, Ans. at
¶ 38); and (2) comScore’s software for Windows did not authorize data collection from local
networks, though its software for Macintosh did allow limited connection to computers
networked with its Macintosh users (see id. ¶ 10). comScore suggests that if the court
certifies a class limited to either group of users, discovery on the merits relating to the other
group would be rendered irrelevant and wasteful.
Plaintiffs disagree. They argue that bifurcated discovery will delay the litigation.
They are particularly sensitive to any delay induced by comScore because of comScore’s
repeated praise of the “rocket docket” of Virginia, its preferred venue. Secondly, they argue
that bifurcation will require increased judicial supervision because the parties will disagree
about the permissible scope of class certification discovery. Third, they argue that comScore
has not met its burden of establishing “good cause” for a protective order delaying merits
discovery.
“The Federal Rules of Civil Procedure give magistrate judges broad discretion in
resolving discovery disputes.” Heyman v. Beatrice Co., Inc., No. 89 C 7381, 1992 WL
245682, at *2 (N.D. Ill. Sept. 23, 1992). That discretion extends to decisions to bifurcate
discovery. See Ocean Atlantic Woodland Corp. v. DRH Cambridge Homes, Inc., No. 02 C

5

2523, 2004 WL 609326, at *2 (N.D. Ill. March 23, 2004) (“[w]hether to bifurcate discovery
is a matter committed to the discretion of the trial court”). Though the Federal Rules of Civil
Procedure do not explicitly provide for bifurcated discovery, the 2003 Advisory Committee
Notes to Rule 23 recognize that bifurcation is often appropriate: “it is appropriate to conduct
controlled discovery . . . limited to those aspects relevant to making the certification decision
on an informed basis.” Fed. R. Civ. P. 23 Advisory Committee’s Notes. In deciding motions
to bifurcate merits discovery from class certification discovery, courts consider the following
factors: (1) expediency, meaning whether bifurcated discovery will aid the court in making
a timely determination on the class certification motion, see Plummer v. Chicago
Journeyman Plumbers’ Local Union No. 130, U.A., 77 F.R.D. 399, 402 (N.D. Ill. 1977); (2)
economy, meaning “the potential impact a grant or denial of certification would have upon
the pending litigation,” see Gonzalez v. Pepsico, Inc., No. 06-2163-KHV, 2007 WL 1100204,
at *3 (D. Kan. April 11, 2007), and whether the definition of the class would “help determine
the limits of discovery on the merits,” see American Nurses’ Assoc. v. State of Illinois, 1986
WL 10382, at *3 (N.D. Ill. Sept. 12, 1986); and (3) severability, meaning whether class
certification and merits issues are closely enmeshed, see Gray v. First Winthrop Corp., 133
F.R.D. 39, 41 (N.D. Cal. 1990).
An evaluation of these factors leads to the conclusion that bifurcation of discovery is
the more sensible approach to discovery in this particular case. Proceeding with merits
discovery, which may well involve the review of millions of documents not directly relevant

6

to the issues of class certification, may delay the parties’ submission of supplemental briefing
on the class certification issue. Any delay would frustrate the court’s effort to certify the
action as a class action “[a]t an early practicable time,” as is mandated by Rule 23(c)(1)(A).
For this reason, the Manual for Complex Litigation counsels that “[d]iscovery relevant only
to the merits delays the certification decision and may ultimately be unnecessary.” Manual
for Complex Litigation (Fourth) § 21.14 (2011). Regarding this expediency concern,
Plaintiffs do not address whether bifurcation will result in a speedier determination of the
class certification issue, but rather counter that bifurcation will significantly delay the
resolution on the merits. This argument fails because bifurcated discovery will not take
significantly more time than would non-bifurcated discovery. The parties’ report of their
Rule 26(f) planning meeting specifies that non-bifurcated discovery should take 10 months.
(R. 60, Report of the Pty.’s Planning Meeting at 3.) This is approximately the same total
length of time as comScore envisions for bifurcated discovery, excluding the pause while the
assigned District Judge considers the fully-briefed class certification motion—a required
pause before reaching the merits of the case. Moreover, if a class is certified, the definition
of “that class should help determine the limits of discovery on the merits,” American Nurses’
Assoc., 1986 WL 10382 at *3, which will save time as the parties move forward with the
litigation. Overall, the potential for a modest delay in reaching resolution on the merits does
not outweigh the court’s obligation under Rule 23(c) to resolve the class certification issue
“at an early practicable time.”

7

Secondly, considerations of economy weigh in favor of bifurcating discovery in this
case. As comScore points out, the limited statutory damages available to Plaintiffs are likely
an insufficient motivation to litigate in the absence of class certification. Notably, Plaintiffs
do not argue that they would pursue the merits as individual plaintiffs. Instead, they argue
that economy is not an appropriate consideration before the court. (R. 70, Pls.’ Resp. at 7.)
Two treatises disagree with Plaintiffs on this point. According to the Manual for Complex
Litigation, “in cases that are unlikely to continue if not certified, discovery into aspects of the
merits unrelated to certification . . . can create extraordinary and unnecessary expense and
burden.” Manual for Complex Litigation at § 21.14. Similarly, McLauchlin on Class
Actions summarizes that “[c]ourts are more likely to decline requests to stay pure merits
discovery when the nature of the putative representative’s claims suggests that it would
continue to prosecute individual claims if certification is denied,” which suggests that courts
are indeed more likely to grant requests to stay merits discovery when the nature of the
putative representative’s claims indicates that the claimant would not pursue their claims if
certification is denied. See McLauchlin on Class Actions (Eighth) § 3:10 (2011). In this
case, bifurcation of discovery will be economical not only if certification is denied, but also
if it is approved. As discussed above, if comScore’s averments about the differences
between its Macintosh and Windows software are borne out by discovery, the court might
certify a limited class. Waiting until the court has made that determination may avoid
needless discovery into issues that are ultimately not relevant to the litigation.

8

The final factor—the severability of class certification issues from merits issues—also
points towards bifurcation of discovery.

Though the “boundary between a class

determination and the merits may not always be easily discernible,” Eggleston v. Chicago
Journeymen Plumbers’ Local Union No. 130, U.A., 657 F.2d 890, 895 (7th Cir. 1981), it is
possible to draw general lines in this case. Discovery relevant to class certification will focus
on the prerequisites for class actions provided by Rule 23(a): (1) numerosity of party
plaintiffs; (2) commonality of questions of law and fact; (3) typicality of the representative
parties’ claims and defenses to those of the class; and (4) adequacy of representation. Fed.
R. Civ. P. 23(a). Regarding numerosity, the inquiry is whether “a class approach would be
useful to avoid the practical problems of trying to join many named plaintiffs or otherwise
clog the docket with numerous individual suits,” Eggleston, 657 F.2d at 895, so certification
discovery will seek to determine the approximate number of potential members of the class.
Regarding commonality, the inquiry is whether “there appear to be common questions of law
or fact.” Id. Here, the commonality prong will likely turn on whether the putative class
members’ consent, and the scope of consent, is subject to evaluation on a class-wide basis,
and whether comScore’s adherence to the scope of consent is subject to analysis on a classwide basis. The third prong, typicality, “requires a showing, not unrelated to commonality,
that others suffer from similar alleged grievances” of Plaintiffs. Id. at 896.

Here, the

typicality prong will turn on whether Harris’ and Dunstan’s alleged experiences downloading
and removing the comScore software are similar to those of the putative class. Lastly, the

9

adequacy prong asks whether “the named parties are qualified and capable of fully pursuing
the common goals of the class without collusion or conflicts of interests.” Id.
In this case, the following issues are not relevant to the class certification analysis
because they do not touch on any of the four prongs of Rule 23(a): (1) the development of
the comScore software; (2) internal comScore emails; (3) comScore’s relationships with its
bundling partners (excluding comScore’s agreements with bundling partners to obtain
consent from prospective comScore users); and (4) comScore’s relationships with its clients.
To the extent that class certification matters are tightly intertwined with merits discovery,
comScore must produce the information and/or documents requested so that Plaintiffs may
develop their arguments relevant to class certification. While Plaintiffs are likely correct that
bifurcation will result in more active judicial supervision, that alone is not sufficient to deny
comScore’s motion.
Turning now to the specific interrogatories and requests for production currently at
issue, this court concludes that comScore must answer the following requests as they pertain
to the certification issues: (1) Interrogatories 1-5, 11 (with limitation), 15-17, and 22; and (2)
Requests for Production (“RFP”) 1.2 Interrogatories 1-5 and RFP 1 seek general information
about comScore’s process of answering interrogatories and RFPs. comScore must respond

2

comScore has agreed to respond to Interrogatories 6-8, and 12-14, and Requests for
Production 2, 13, 15, 18, 21-26, 34, 35, and 40-42. (R. 67, Def.’s Mem. in Support of
Bifurcation at 10 n.5 and 11 n.6.) The parties have also agreed to a process for addressing
Interrogatory 23 and RFP 43, which relate to expert reports and discovery. (Id. at 15 and R.
70, Pls.’ Resp. at 13 n.5.)
10

to these requests to the extent they relate to the Interrogatories and RFPs that comScore has
agreed to answer and is ordered to answer.
Interrogatory 11 seeks discovery about the types of data that third-party entities
purchased, specified by the purchasers. Plaintiffs argue that this discovery is necessary to
establish “whether the class members suffered damages and, if so, whether such damages
predominate . . . or are incidental.” (R. 70, Pls.’ Resp. at 10.) This court agrees that whether
the members of the proposed classes suffered similar damages is relevant to the issue of
commonality, and further agrees that whether comScore sold similar types of panelist data
is relevant to the issue of whether comScore’s alleged breach of the putative class members’
scope of consent is subject to evaluation on a class-wide basis. But, the identities of
comScore’s clients are not relevant to the class certification analysis. Defendant is ordered
to answer this interrogatory without having to identify the third-party purchasers.
Interrogatory 15 seeks information about comScore’s source code control and source
code library retention policies and practices. comScore argues that the source code it has
agreed to produce to Plaintiffs will allow them to “test their core allegations on the merits
while also addressing the central issues on class certification.” (R. 67, Def.’s Mem. in
Support of Bifurcation at 2.) But the information sought by Interrogatory 15 may be
necessary to assess the integrity of comScore’s source code.

11

Interrogatories 16 and 173 ask comScore to identify and describe each type of
information that the comScore software “monitors, collects, retains, and/or transmits” from
the Windows and Macintosh panelists.

Whether comScore collected the same, or

substantially the same, types of content from the panelists is relevant to the commonality
prong of the class certification analysis. Interrogatory 22 seeks factual information relating
to comScore’s position that class certification is inappropriate in this case. This interrogatory
clearly seeks information directly related to the class certification issues.
The remaining interrogatories (9, 10 and 18-21) and RFPs (3-12, 14, 16, 17, 19, 20,
29-33, 36-39 and 44-45) are outside the scope of class certification discovery and are stayed
pending resolution of the class certification issues. Interrogatories 9 and 18 and RFPs 4, 912, 31-33, and 37, seek discovery about comScore’s “bundling partners” and its “Trees for
the Future” program. comScore’s bundling partners are the entities that offer comScore’s
software in conjunction with their own free software applications. (R. 1, Compl. at ¶ 13.)
comScore’s Trees for the Future program induces Internet users to become panelists in
exchange for having trees planted. (R. 70, Pls.’ Resp. at 12.) comScore’s relationships with
third-parties are not relevant to the issues of numerosity, commonality, typicality, and

3

According to comScore, its commitment to produce the relevant source code “addresses
Plaintiffs’ Requests for Production Nos. 2, 27-28, and Interrogatories 16 and 17.” (R. 67,
Def.’s Mem. in Support of Bifurcation at 10 n.5.) This court disagrees in regards to
Interrogatories 16 and 17. While the source code would likely enable Plaintiffs to access the
information sought by Interrogatories 16 and 17, Plaintiffs suggest that having to “go fish”
through the code is burdensome. (R. 70, Pls.’ Resp. at 14.) Because Interrogatories 16 and
17 clearly relate to class certification issues, Plaintiffs are entitled to that discovery now.
12

adequacy of representation. To the extent that Plaintiffs believe that responses to these
Interrogatories and RFPs would address issues of consent for the panelists who obtained the
comScore software through a bundling partner or the Trees for the Future program, they are
unnecessary because comScore has already agreed to produce that information in response
to Interrogatories 13 and 14 and RFPs 15 and 24-26. (R. 67, Def.’s Mem. in Support of
Bifurcation at 11 n.6.) Plaintiffs will also be able to explore comScore’s methods of
obtaining consent from prospective panelists by examining the source code and RK Verify
software.
Interrogatory 10 seeks discovery about the purchasers of comScore’s market research
data. Plaintiffs argue that this discovery is necessary to establish “whether the class members
suffered damages and, if so, whether such damages predominate . . . or are incidental.”
(R. 70, Pls.’ Resp. at 10.) As stated earlier this court agrees that whether the members of the
proposed classes suffered similar damages is relevant to the issue of commonality but the
identities of comScore’s clients are not relevant to the certification analysis.
Interrogatory 19 and RFP 8 seek information related to comScore’s investigation and
termination of its Macintosh panel. Plaintiffs argue that they need this discovery to establish
whether injunctive relief is appropriate. But these requests are overly broad for that purpose.
For example, Interrogatory 19 requests that comScore describe its “investigation of the MAC
PANEL, INCLUDING the reasons for its ultimate termination, and IDENTIFY ALL
DOCUMENTS RELATING TO such investigation.”

13

Similarly, RFP 8 requests all

documents and ESI relating to the investigation and termination of the Macintosh panel.
How comScore investigated its Macintosh panel and its reasons for terminating it are not
relevant to the appropriateness of injunctive relief in this case. If Plaintiffs need discovery
to verify comScore’s termination of the Macintosh panel, they should propound a more
limited request to that end.
Interrogatory 20 and RFP 44 seek discovery about comScore’s public relations
response to the litigation. comScore’s efforts to manage its public relations are not relevant
to the Rule 23 prerequisites for class certification. Interrogatory 21, which seeks discovery
about comScore’s affirmative defenses, goes to the merits and is not relevant to class
certification.
RFPs 3, 5-7, and 14 seek information about comScore’s development of its software.
This request is unduly burdensome and unlikely to yield discovery relevant to the
certification issues.

The members of the putative class were not impacted by the

development of comScore’s software, but by the software itself.
RFPs 16, 17, 19, 20, 28,4 36, and 38 seek communications between comScore and its
employees on a variety of topics. At this stage in the litigation, Plaintiffs need to establish
that comScore’s software impacted the putative class members in a common manner.

4

According to comScore, its commitment to produce the relevant source code “addresses
Plaintiffs’ Requests for Production . . . 27-28.” (R. 67, Def.’s Mem. in Support of
Bifurcation at 10 n.5.) Because RFP 28 also requests communications between comScore
and its employees, the source code is not responsive.
14

comScore’s internal communications are less relevant to that issue than how the software
actually impacted the putative class. Plaintiffs have access to the source code, which, in
conjunction with the discovery responses mandated in this order, should demonstrate how
the software impacted the members of the putative class.
RFPs 27,5 29, and 30 seek documents, ESI, and communications relating to
comScore’s use and sale of information personal to Harris and other panelists. This court
agrees with Plaintiffs that whether comScore sold similar types of personal information of
other panelists is relevant to commonality, and whether comScore sold Harris’s personal
information is relevant to typicality. However, as framed, the scope of these requests are
overly broad as they would include comScore’s contracts and contract negotiations with its
clients, neither of which is relevant to the Rule 23 analysis. Furthermore, comScore’s answer
to Interrogatory 11 should provide Plaintiffs with information on the types of personal data
comScore sold to others.
RFP 45, which seeks information about comScore’s liability insurance, is also not
relevant to the class certification issues. However, based on the parties’ joint statement in
their Report of the Parties’ Planning Meeting, comScore should have produced a copy of the

5

According to comScore, its commitment to produce the relevant source code “addresses
Plaintiffs’ Requests for Production . . . 27-28.” (R. 67, Def.’s Mem. in Support of
Bifurcation at 10 n.5.) In fact, RFP 27 requests documents and ESI relating to agreements
between comScore and the panelists. The source code, alone, is therefore not responsive to
RFP 27.
15

relevant policies on December 7, 2011, pursuant to Federal Rule of Civil Procedure
26(a)(1)(A)(iv).6 (R. 60 at 1.)
Conclusion
For the foregoing reasons, comScore’s motion to bifurcate discovery is granted.
Discovery on the merits is stayed until after the court rules on Plaintiffs’ motion to certify the
action as a class action. comScore is ordered to respond to the written discovery requests as
detailed herein by March 23, 2012.
ENTER:

_________________________________
Young B. Kim
United States Magistrate Judge

6

Because the parties have not provided the court with substantive arguments regarding RFP
39, the court is unable to determine whether this request is relevant to the class certification
or merits issues. The parties should meet and confer in regard to RFP 39.
16