Campbell et al v. Facebook Inc.

Filing 178

Joint Notice of Refiling of Class Certification Briefs and Evidentiary Objections (Dkt. 138, 149, 167, 170, and 172) by Facebook Inc. (Attachments: # 1 Exhibit 1 - Replacement for Dkt. 138 (Redacted), # 2 Exhibit 2 - Replacement for Dkt. 149 (Redacted), # 3 Exhibit 3 - Replacement for Dkt. 167 (Redacted), # 4 Exhibit 4 - Replacement for Dkt. 170, # 5 Exhibit 5 - Replacement for Dkt. 172 (Redacted))(Chorba, Christopher) (Filed on 3/21/2016) Modified on 3/22/2016 (vlkS, COURT STAFF).

1
2
3
4
5
6
7

Michael W. Sobol (State Bar No. 194857)
[e-mail]
David T. Rudolph (State Bar No. 233457)
[e-mail]
Melissa Gardner (State Bar No. 289096)
[e-mail]
LIEFF CABRASER HEIMANN & BERNSTEIN, LLP
275 Battery Street, 29th Floor
San Francisco, CA 94111-3339
Telephone: [phone]
Facsimile: 415.956.1008

12

Hank Bates (State Bar No. 167688)
[e-mail]
Allen Carney
[e-mail]
David Slade
[e-mail]
CARNEY BATES & PULLIAM, PLLC
11311 Arcade Drive
Little Rock, AR 72212
Telephone: [phone]
Facsimile: 501.312.8505

13

Attorneys for Plaintiffs and the Proposed Class

8
9
10
11

14

UNITED STATES DISTRICT COURT

15

NORTHERN DISTRICT OF CALIFORNIA

16

OAKLAND DIVISION

17
18
19

MATTHEW CAMPBELL and MICHAEL
HURLEY, on behalf of themselves and all
others similarly situated,
Plaintiff,

20
21

FACEBOOK, INC.,

PLAINTIFFS’ RESPONSE TO
DEFENDANT’S “OBJECTION TO AND
REQUEST TO STRIKE NEW EVIDENCE
AND MISSTATEMENTS OF FACT”

v.

22

Case No. C 13-05996 PJH (MEJ)

23

Defendant.

Date:
Time:
Judge:
Place:

March 16, 2016
9:00 a.m.
Hon. Phyllis J. Hamilton
Courtroom 3, 3rd Floor

24
25
26
27
28
PLAINTIFFS’ RESPONSE TO DEFENDANT’S
“OBJECTION TO AND REQUEST TO STRIKE NEW
EVIDENCE” CASE NO. 13-CV-05996-PJH (MEJ)

1

I.

INTRODUCTION

2

Facebook mischaracterizes Plaintiffs’ reply to Facebook’s opposition to class certification

3

as presenting “new evidence,” in order to file an additional eleven pages of unauthorized surreply

4

argument in an apparent end-run around this Court’s order limiting Facebook’s briefing on this

5

motion. See Dkt. 146 (denying Facebook’s request for an additional 15 pages, but allowing an

6

additional 5 pages, of briefing). Plaintiffs’ reply offers no new evidence, but rather replies to

7

arguments raised by Facebook, drawing on the existing evidentiary record to counter Facebook’s

8

arguments. Facebook’s Objection to this supposed “new evidence” includes argument by counsel

9

and two declarants which—while ultimately nowhere refuting that Facebook collects and exploits

10

users’ private information or that it maintains the “Titan Information” identifying all those

11

users—improperly go to the merits of these issues. Accordingly, Plaintiffs respectfully submit

12

that the Court should overrule Facebook’s objection and strike the surreply argument and

13

declarations, or, in the alternative, grant leave to Plaintiffs to file additional materials to address

14

Facebook’s mischaracterizations contained therein.

15

II.

ARGUMENT

16

A.

Facebook’s Objection is Procedurally Improper.

17

Facebook violates Civil L.R. 7-3(d)(1)’s unambiguous requirement that any Objection to

18

Reply Evidence “may not include further argument on the motion.” Facebook’s Objection,

19

including the declarations, constitutes “argument on the motion.” Indeed, the Objection itself

20

acknowledges that Facebook engages in argument concerning the underlying issues, and actually

21

threatens the possibility that it may file further argument in the future.1 Here, Facebook does

22

more than “straddle the line between objections and argument,” but rather engages in unabashed

23

argument. Ross v. Ecolab Inc., No. 13-5097, 2015 WL 5681323, at *15 (N.D. Cal. Sept. 28,

24

2015) (Hamilton, J.) (finding that the specific evidentiary objections in that case did not constitute

25

surreply argument).2 The Declaration of Dale Harrison argues at length regarding the

26
27
28

1

See Dkt. 169-5 (FB Objection), at 4 n. 3 (“Facebook does not respond at this time to all the new
facts and argument raised by Dr. Golbeck. . . . Facebook may seek leave of Court to submit
additional evidence to rebut her contentions”) (emphasis added).
2
Plaintiffs’ counsel acknowledge that Plaintiffs’ reply brief erroneously (and unintentionally)
contains a misplaced quotation mark when quoting a statement from a declaration of Dale
Footnote continued on next page
1

PLAINTIFFS’ RESPONSE TO DEFENDANT’S
“OBJECTION TO AND REQUEST TO STRIKE NEW
EVIDENCE” CASE NO. 13-CV-05996-PJH (MEJ)

1

infeasibility of identifying class members, albeit based on a blatant mischaracterization of Dr.

2

Golbeck’s method. See, e.g., Dkt. 170-2, at 1, ¶3 (taking issue with the “assumptions” used by

3

Dr. Golbeck). The Declaration of Alex Himel misinterprets Dr. Golbeck’s method as requiring a

4

three step query, making it unworkable. See Dkt. 170-1, at 2, ¶6. While Facebook’s arguments

5

lack merit, they nonetheless constitute arguments, and therefore its Objection is improper.

6

Laced with argument, Facebook’s Objection is nothing less than improper surreply.

7

Carr v. Allied Waste Sys. of Alameda Cnty., No. 10-715, 2011 WL 4047495, at *7 (N.D. Cal.

8

Sept. 8, 2011) (Hamilton, J.), aff’d on other grounds, 516 F. App’x 677 (9th Cir. 2013) (denying

9

request to file “Additional Information to Address Misleading Facts Made by Defendant” as

10

improper attempt to submit a surreply). As unauthorized surreply, Facebook’s Objection also

11

violates the spirit of the Court’s order denying its request for additional 15 pages of opposition

12

briefing. After consideration, the Court found that Facebook “has not adequately supported its

13

request for an additional 15 pages.” Dkt. 146 (Order), at 1. The Court did allow an additional 5

14

pages of briefing, which Facebook used in its opposition. Particularly in the face of this order,

15

Facebook should have sought leave prior to filing new briefing. Roe v. Doe, No. 9-682, 2009 WL

16

1883752, at *5 (N.D. Cal. June 30, 2009) (Hamilton, J.); Civ. L.R. 7–3(d) (“Surreplies are not

17

authorized without first obtaining leave of court”; once a reply brief is filed, no additional

18

memoranda may be filed without court approval).

19

Facebook’s Objection is procedurally improper and therefore the Court should strike it.

20
21
22
23
24
25
26
27
28

Footnote continued from previous page
Harrison offered earlier in the proceedings in connection with a discovery dispute. At page 10
line 20, the reply brief should have read “would ‘require consulting with engineers in every group
. . . ’” rather than read “‘would require consulting with engineers in every group. . . ’”. Plaintiffs’
counsel apologize for this inadvertent error. Although Mr. Harrison’s declaration literally states
that identifying every use of message content “may” require consultation with all these engineers,
Plaintiffs nonetheless disagree with Facebook’s assertion that Mr. Harrison’s testimony was
substantively misrepresented. The clear implication of his testimony is that identifying all the
uses Facebook makes of Private Messages is an enormous if not impossible task. After all, Mr.
Harrison never explains how it “may not” be the case that one would need to consult with
engineers in every group. Apparently, Mr. Harrison meant to leave the impression that it “would”
be the case, when offering testimony to obstruct discovery, but now in the context of class
certification invokes his lawyers’ use of the word “may” to back off the intended meaning.
2

PLAINTIFFS’ RESPONSE TO DEFENDANT’S
“OBJECTION TO AND REQUEST TO STRIKE NEW
EVIDENCE” CASE NO. 13-CV-05996-PJH (MEJ)

1

B.

2
3

Facebook Could And Should Have Raised Its Surreply Points in Its
Opposition.

Facebook’s Objection contains numerous substantive assertions that respond to material

4

raised in Plaintiffs’ opening brief, not the rebuttal brief. For example, in Alex Himel’s

5

declaration he asserts the “impossib[ility]” of querying EntShares, but this is purely in response to

6

Dr. Golbeck’s query identified in her opening report and is not related to her rebuttal analysis.

7

Dkt. 170-1 (Feb. 26, 2016 Himel Decl.), at 1, ¶5. Mr. Himel also argues that EntShares are used

8

to render URL attachments, and also states (without evidence or elaboration) that “EntShares play

9

a critical role in a number of ways—including user experience and abuse-prevention functions, as

10

well as the functionality of users’ privacy settings.” Id., at 3, ¶10. This testimony is directly

11

related to Facebook’s expert Dr. Goldberg’s opposition opinions (and indeed that was the very

12

testimony to which Dr. Golbeck was responding (see Dkt. 167-1, Ex. 1 (Golbeck Rebuttal

13

Report), at 10, ¶24), and therefore could and should have been included with Facebook’s

14

opposition. Similarly, Dale Harrison’s declaration repeats arguments Facebook made through its

15

opposition regarding creation of EntShare objects and associations, and therefore does not reply

16

to Dr. Golbeck’s rebuttal opinions. Dkt. 170-2 (Feb. 26, 2016 Harrison Decl.), at 4, ¶7, bullet

17

point 4. Accordingly, Facebook’s arguments contained in its Objection should have been

18

presented in its Opposition, and are not the proper subject of surreply.

19

C.

20

Facebook cannot assert that there is anything improper with filing Dr. Golbeck’s Rebuttal

Dr. Golbeck’s Rebuttal Report Contains No “New Evidence.”

21

Report on reply, so therefore asserts that the rebuttal report contains “new evidence.” See Civ.

22

L.R. 7–3(c) (“Any reply to an opposition may include affidavits or declarations . . . .”). To the

23

contrary, Dr. Golbeck’s rebuttal simply addresses arguments raised by Facebook in its opposition,

24

relying on the record of the case. Therefore, Dr. Golbeck’s rebuttal presents nothing new, and

25

does not present any occasion for Facebook to file a surreply.

26
27
28

1.

The “Titan Information.”

Contrary to Facebook’s assertion, Dr. Golbeck’s Rebuttal Report does not present a new
methodology to identify class members. Instead, in direct reply to Facebook’s challenges, the
3

PLAINTIFFS’ RESPONSE TO DEFENDANT’S
“OBJECTION TO AND REQUEST TO STRIKE NEW
EVIDENCE” CASE NO. 13-CV-05996-PJH (MEJ)

1

Rebuttal Report provides clarification and context to the “database query” that Dr. Golbeck

2

specifically described in her initial expert report. Dkt. 138-4, Ex. 2 (Golbeck Report), at 28, ¶103

3

(“To retrieve a list of class members…[a] database query could be used to select the Facebook

4

IDs of everyone whose actions had created an EntShare from a private message.”); Id., at ¶104

5

(“The exact code will vary based on the type of database.”). Rather than identifying “an entirely

6

new data source,” as Facebook wrongly asserts, Dr. Golbeck references the “Titan Information”

7

that Facebook retrieved for the Representative Plaintiffs as the same type of information needed

8

to identify the absent class members. See Dkt. 126-1 (Oct. 6, 2015 Harrison Decl.), at 2, ¶5

9

(stating that Harrison was “able to extract” the “Titan Information” for the Named Plaintiffs).

10

Facebook’s feigned surprise can hardly be credited—Titan Information confirms that the

11

relevant Private Message data resides on Facebook’s system. First, as early as June 1, 2015,

12

Facebook declarant Alex Himel explained that “‘Titan’ was the internal name for the Facebook

13

Messages product.” FB App.3 1607 (June 1, 2015 Himel Decl.), at 1613. Similarly, Facebook

14

produced what it, itself, calls “Titan Info” for the Named Plaintiffs’ intercepted Private Messages.

15

See FB App. 1535 (Facebook’s Second Suppl. Resps. and Objs. to Pltfs.’ Narrowed Second Set of

16

Interrogatories), at 1553 (“Facebook has identified and produced representations of the data

17

stored for messages, referred to as ‘Titan Information’”) (emphasis added), and Ex. A thereto, at

18

1555.4 In other words, obtaining the Private Message data that Facebook indisputably stores

19

would most logically occur via the Titan Information, based upon Facebook’s own

20

representations and characterizations of the data it has produced, thus far. Dr. Golbeck’s Rebuttal

21

Report simply reiterates this point, mentioning Titan Information by name, just as Facebook’s

22

own witnesses already had.

23

Dr. Golbeck’s clarification that her methodology relies on the existence of Titan

24

Information is the necessary by-product of Facebook’s misleading implication that certain data

25
26
27
28

3

Citations to “FB App.” are to the Appendix of Evidence filed by Facebook in connection with
its opposition to Plaintiffs’ Motion for Class Certification.
4
Dr. Golbeck lists Facebook’s interrogatory responses which reference the “Titan Info” in her
initial report, as did Plaintiffs in their opening motion. See Dkt. 138-4, Ex. 2 (Golbeck Report), at
Ex. B; Dkt. 138 (Motion for Class Certification), at 13.
4

PLAINTIFFS’ RESPONSE TO DEFENDANT’S
“OBJECTION TO AND REQUEST TO STRIKE NEW
EVIDENCE” CASE NO. 13-CV-05996-PJH (MEJ)

1

fields do not exist in its attempt to confuse the most commonsense and elemental steps in a

2

straightforward, objective process. Dkt. 149 (FB Opp. Br.), at 14 (claiming that Dr. Golbeck’s

3

proposal did not entail searching data sufficient to identify certain class members). The Rebuttal

4

Report merely pointed out what Facebook already knew (but chose to ignore) when it claimed

5

that the methodology would yield incomplete results. Namely, Dr. Golbeck clarified that Titan

6

Information contains the needed data, specifically: “a. message sender; b. message recipient; c. a

7

timestamp . . . and d. whether or not a URL . . . [is] associated with the message.” Dkt. 167-1,

8

Ex. 1 (Golbeck Rebuttal Report), at 2, ¶8. Nowhere does Facebook claim that any other

9

information would be required to properly identify the members of the class. Dr. Golbeck has not

10

changed her methodology—it remains the same sound, workable procedure that was articulated in

11

the initial expert report. She has simply attempted to remedy Facebook’s misdirection.

12

Facebook’s newly-submitted complaints that compiling the class list from Private

13

Message data would be “impossible” are untimely and incorrect. See Dkt. 170-1 (Feb. 26, 2016

14

Himel Decl.), at 1 ¶5. Dr. Golbeck’s opening and rebuttal reports simply state that because

15

Facebook stores the relevant data associated with users’ Private Messages, “[a] database query

16

could be used to select the Facebook user IDs of everyone whose actions had created an EntShare

17

from a private message.” Dkt. 138-4, Ex. 2 (Golbeck Report), at 28, ¶¶103-104. With respect to

18

each interception at issue in this case, Facebook’s source code creates an “EntShare” encoded

19

with specific attributes indicating it originated as a Private Message attachment with URL

20

content.5 These relevant EntShares each have a unique Facebook ID (an “fbid”) which appears in

21

the attachment field for the specific message that created it. The messages which have such

22

EntShare attachments will themselves have a “message ID,” which links to the other relevant

23

information (sender, recipient, and date) needed to identify class members. In what has to be

24

labeled a blatant mischaracterization, Facebook states that Dr. Golbeck’s method requires the

25
26
27
28

5

Improperly arguing the merits in its Objection, Facebook asserts that EntShare objects specific
to URL attachments in private messages cannot be uniquely identified. See, e.g., Dkt. 170-2
(Feb. 26, 2015 Harrison Decl.), at 2, ¶6. Dr. Golbeck disagrees, based on her review of
Facebook’s source code, which shows that EntShare attributes include those which show the
presence of a shared URL attachment. See Dkt. 167-1, Ex. 1 (Golbeck Rebuttal Report), at 2, ¶8;
Dkt. 138-4, Ex. 2 (Golbeck Report), at 27, ¶99.
5

PLAINTIFFS’ RESPONSE TO DEFENDANT’S
“OBJECTION TO AND REQUEST TO STRIKE NEW
EVIDENCE” CASE NO. 13-CV-05996-PJH (MEJ)

1

assembly of a separate and freestanding list of all existing “message IDs” whether they were

2

pertinent to the class or not. But the subject of the query, as laid out with example query code in

3

Dr. Golbeck’s opening report, is for the relevant EntShare, not the message ID. Dkt. 138-4, Ex. 2

4

(Golbeck Report), at 28, ¶104. The resulting query would identify the relevant message ID, and

5

with the same query the date parameters for the class period can be applied, which would yield

6

the applicable user IDs of sender and recipient contained in other data fields associated with the

7

message ID. See id., at 28, ¶103; Dkt. 167-1, Ex. 1 (Golbeck Rebuttal Report), at 2, ¶9. Dr.

8

Golbeck’s statement that “[b]y starting with a list of all message IDs, a database query could be

9

written ...” (id.) just means that Titan Information includes the message IDs, where the EntShares

10

themselves do not. Thus, the parade of impossibilities Facebook marches out relies upon a

11

deceptively false premise.

12

Accordingly, it is of no consequence that Titan Information was not referred to by name,

13

as it was described in substance, and Facebook was of course well aware of its existence. When

14

Facebook challenged the ability to use Titan Information, Dr. Golbeck properly replied with

15

further specifics from the previous disclosed record.

16
17

2.

“Scribeh_share_stats.”

Plaintiffs have set forth evidence showing that Facebook scans Private Messages while in

18

transit to create a so-called EntShare with certain attributes of the Private Message, including

19

those representing the URL content of a message, which is redirected through discrete code for

20

later use unrelated to sending the message. As set forth in Dr. Golbeck’s opening report, one

21

such place this Private Message content is sent to is a table called “share_stats” (or sometimes

22

called “share_stats Hive table” within Facebook), a name which should speak for itself. In its

23

opposition, Facebook claimed it no longer puts Private Message data contained in the share_stats

24

table as it once did “to inform the URLs displayed in the Recommendations Feed.” FB App.

25

1693 (Jan. 15, 2016 Fechete Decl.), at 1698, ¶14. So, in her rebuttal report, Dr. Golbeck

26

confirmed that her search of the source code showed existing references to the share_stats table.

27

As further evidence of Facebook’s logging of Private Message data, Dr. Golbeck amplified on the

28

connected mechanism used to gather data, “scribeh_share_stats,” a code function that retrieves
6

PLAINTIFFS’ RESPONSE TO DEFENDANT’S
“OBJECTION TO AND REQUEST TO STRIKE NEW
EVIDENCE” CASE NO. 13-CV-05996-PJH (MEJ)

1

the same Private Message data for logging. Dkt. 167-1, Ex. 1 (Golbeck Rebuttal Report), at 11-

2

12, ¶¶31-32. As its name implies, scribeh_share_stats “writes” pertinent Private Message

3

attributes to logs, acting like a giant vacuum within Facebook’s system, sucking up immense

4

volumes of user data for future use. Id., at 12, ¶34.

5

References linking scribeh_share_stats to share_stats have been a part of this case since

6

long before Dr. Golbeck’s Rebuttal Report. For example, describing a change to Facebook’s

7

logging functionality, one document states: “I wanted to make sure that noone was using the

8

share_stats Hive table (verified with

9

change (data will be logged to scribeh_share_stats) and a backend change (data will be collected

10

using ptail).” See FB000003810. This document shows that data previously being logged in the

11

“share_stats Hive table” was, going forward, going to be logged to “scribeh_share_stats,” which

12

demonstrates that the scribeh_share_stats and share_stats functionality was related (if not

13

identical). Moreover, contrary to Mr. Himel’s assertion, Dr. Golbeck did link scribeh_share_stats

14

to share_stats, when she noted that scribeh_share_stats creates a report with an attribute for

15

“type,” defined as “S_S,” which elsewhere in the code represents share_stats. Dkt. 167-1, Ex. 1

16

(Golbeck Rebuttal Report), at 12, ¶32.

17

and checked the HiPal history). This is both a frontend

Moreover, Mr. Himel’s newly-submitted testimony is even more inappropriate in light of

18

the fact that Facebook has withheld the source code configuration tables related to share_stats and

19

scribeh_share_stats that would support the (otherwise unsupported) assertions he now makes.

20

Notably, Mr. Himel does not say that Facebook no longer logs Private Message content via

21

“scribeh_share_stats.” In October 2015, Plaintiffs requested, and Facebook refused to produce,

22

the databases containing the configuration tables. See Nov. 20, 2015 email from Facebook

23

counsel Joshua Jessen to Plaintiffs’ counsel David Rudolph (arguing that the requested databases

24

are irrelevant and stating “the production of any of this newly-requested data would be

25

inappropriate”). Plaintiffs are seeking to compel production of this data, and Facebook should

26

not, under the guise of “surreply” material, be allowed to introduce into evidence testimony

27

related to source code functionality that Facebook has refused to produce.

28
7

PLAINTIFFS’ RESPONSE TO DEFENDANT’S
“OBJECTION TO AND REQUEST TO STRIKE NEW
EVIDENCE” CASE NO. 13-CV-05996-PJH (MEJ)

1

D.

2
3

Nothing In Facebook’s Surreply Should Dissuade The Court From Certifying
the Class.

Facebook’s two primary arguments in the Objection—concerning ascertainability of the

4

class and Facebook’s improper use of Private Message content—offer no basis to deny class

5

certification. On ascertainability, Facebook does not dispute that Titan Information includes all

6

the necessary information to identify all class members. Instead, it essentially bickers that the

7

precise code for the query has not necessarily yet been presented in final form. The bickering

8

misses the main point, that is, that Plaintiffs’ expert has identified the existence of the data and

9

general methodologies for extracting it that can be successful, even if it takes, quite

10

unsurprisingly, working within the systems (a luxury she has not yet been afforded) to finalize the

11

details. On improper use, again, Facebook nowhere denies that its source code shows it is

12

systematically logging Private Message content and that it has historically made use of that

13

content for commercial purposes. It merely bickers that so far some of what Plaintiffs identify in

14

the source code as pointing to improper use is no longer operative. Moreover, the logging in and

15

of itself constitutes an interception, and is particularly concerning given the immense extent of

16

Facebook’s logging and hence storage of the associated data.

17

III.

18

CONCLUSION
For the foregoing reasons, Plaintiffs respectfully request that the Court overrule and deny

19

Facebook’s Objection to and Request to Strike New Evidence, and to strike the declarations of

20

Alex Himel and Dale Harrison filed therewith. Alternatively, if the Court declines to strike

21

Facebook’s two declarations, Plaintiffs respectfully request leave to file a response thereto,

22

including a declaration from their expert.

23
24
25
26
27
28
8

PLAINTIFFS’ RESPONSE TO DEFENDANT’S
“OBJECTION TO AND REQUEST TO STRIKE NEW
EVIDENCE” CASE NO. 13-CV-05996-PJH (MEJ)

1

Dated: March 9, 2016

By:

/s/ Michael W. Sobol
Michael W. Sobol

2
3
4
5
6
7

Michael W. Sobol (State Bar No. 194857)
[e-mail]
David T. Rudolph (State Bar No. 233457)
[e-mail]
Melissa Gardner (State Bar No. 289096)
[e-mail]
LIEFF CABRASER HEIMANN & BERNSTEIN, LLP
275 Battery Street, 29th Floor
San Francisco, CA 94111-3339
Telephone: [phone]
Facsimile: 415.956.1008

8
9
10
11
12

Rachel Geman
[e-mail]
Nicholas Diamand
[e-mail]
LIEFF CABRASER HEIMANN & BERNSTEIN, LLP
250 Hudson Street, 8th Floor
New York, NY 10013-1413
Telephone: [phone]
Facsimile: 212.355.9592

13
14
15
16
17
18

Hank Bates (State Bar No. 167688)
[e-mail]
Allen Carney
[e-mail]
David Slade
[e-mail]
CARNEY BATES & PULLIAM, PLLC
11311 Arcade Drive
Little Rock, AR 72212
Telephone: [phone]
Facsimile: 501.312.8505

19
Attorneys for Plaintiffs and the Proposed Class
20
21
22
23
24
25
26
27
28
9

PLAINTIFFS’ RESPONSE TO DEFENDANT’S
“OBJECTION TO AND REQUEST TO STRIKE NEW
EVIDENCE” CASE NO. 13-CV-05996-PJH (MEJ)