Campbell et al v. Facebook Inc.
Filing
109
MOTION for Extension of Time to File Plaintiffs' Motion for Extension of Class Certification and Summary Judgment Deadlines filed by Matthew Campbell, Michael Hurley. (Attachments: # 1 Proposed Order, # 2 Declaration of David Rudolph, # 3 Exhibit 1, # 4 Exhibit 2, # 5 Exhibit 3, # 6 Exhibit 4, # 7 Exhibit 5, # 8 Exhibit 6, # 9 Exhibit 7, # 10 Exhibit 8, # 11 Exhibit 9, # 12 Exhibit 10, # 13 Exhibit 11, # 14 Exhibit 12, # 15 Exhibit 13, # 16 Exhibit 14, # 17 Exhibit 15, # 18 Exhibit 16, # 19 Exhibit 17, # 20 Exhibit 18, # 21 Exhibit 19, # 22 Exhibit 20, # 23 Exhibit 21)(Sobol, Michael) (Filed on 9/16/2015)
EXHIBIT 12
June 19, 2015
Christopher Chorba
Direct: +1 213.229.7396
Fax: +1 213.229.6396
CChorba@gibsondunn.com
Client: 30993-00028
VIA ELECTRONIC MAIL
Hank Bates, Esq.
Carney Bates & Pulliam, PLLC
11311 Arcade Drive
Little Rock, AR 72212
Re:
Campbell v. Facebook, Inc., N.D. Cal. Case No. 13-cv-05996-PJH
Dear Hank:
I write in response to your second letter dated June 17, which is a self-serving attempt
to posture and misstate the history regarding Plaintiffs’ discovery requests and the parties’
extensive discussions regarding those requests.
Although it is true that Plaintiffs propounded their First Set of Document Requests on
January 26, 2015, you know that we reached out to you and your colleagues proactively (and
immediately) to meet and confer over several concerns that Facebook had regarding the
breadth and proposed time period for these requests, which purported to reach back to the
date that Plaintiffs believed Facebook was launched to the public in 2006. On February 11,
we had a lengthy telephone conference to discuss our concerns, and you invited us to serve
written responses that proposed compromises on these requests. Facebook then served its
written responses on March 9. Within days, we started discussing these responses—first,
during a brief, in-person conversation before the Case Management Conference on
March 12, and then later during a lengthy, approximately four-hour telephonic call on
March 17. As you know, the call did not resolve our differences, because we were unable to
receive answers to several very straightforward questions—including, for example, whether
the class definition alleged in the Complaint accurately reflected the purported class that
Plaintiffs seek to represent in this lawsuit.
Nevertheless, our conversations continued over the next month, including in
Magistrate Judge James’ courtroom after the discovery conference on April 13. (You will
undoubtedly recall these discussions, particularly the concerns that we reiterated regarding
the breadth of Plaintiffs’ requests, which you attempted to dismiss with a cavalier response
that we should know that this is “how the game is played.”) Following that conference, and
at Plaintiffs’ insistence to prioritize immediate production of “source code,” Facebook was
required to divert its ongoing efforts and prioritize “source code”-related materials by June 1.
Of course, Facebook’s ongoing efforts to collect and produce responsive documents related
to Plaintiffs’ other requests were not put on hold, as we wrote to you on May 13 with a list of
proposed custodians and search terms, and we also suggested another compromise regarding
June 19, 2015
Page 2
the time period for Plaintiffs’ requests. You responded two weeks later, on May 27, and
noted that Plaintiffs had no further suggestions to the proposed search terms.
Last week, on June 12, we identified additional custodians and accepted Plaintiffs’
proposed compromise on the relevant time period for the document requests (April 1, 2010
to December 30, 2013). We also cited our concerns with Plaintiffs’ apparent position that
any individual identified on any email would need to be added to the list of custodians: “If
we were to include as a custodian any person copied on a potentially relevant e-mail (no
matter their actual involvement in the issue), the number of custodians would increase
exponentially. This approach is inconsistent with the proportionality requirement in
Rule 26(b)(2)(C) and (g)(1)(B), the Stipulated Order re Discovery of Electronically Stored
Information in this case (Dkt. 74), as well as the District Court’s ESI Guideline 1.03.” We
also offered to discuss this issue with you.
As the foregoing summary reflects, it is only in the last few weeks that we have been
able to reach agreement on the custodians, search terms, and date range for Plaintiffs’
document requests. Accordingly, your complaints about the volume of the production to
date, and your assertion that there has been “five months” of delay, plainly misstates the
record. Over the last several weeks, we have worked diligently and at great expense to our
client to continue to collect responsive materials from the custodians identified in our
May 13 and June 12 correspondence. Those efforts continue, and we anticipate having
another tranche of documents ready shortly.
As we continue our efforts, it has become apparent that the broad search terms that
we initially proposed in our letter of May 13, when applied to the agreed custodians, are
overbroad. Specifically, these terms have resulted in a potential review population of
approximately 600,000 unique documents. Two particular terms had exceptionally high hit
counts, together adding over 330,000 unique documents to the review set.1 After a review of
several thousand documents, we believe that this search term set is extremely overbroad, and
we are currently utilizing a predictive coding tool to further cull this set. As you may know
“predictive coding” applies advanced machine learning techniques to the text of documents
to automatically classify unreviewed documents as responsive or nonresponsive. Although
relatively new, this procedure is an accepted method of narrowing the review population
pursuant to the Federal Rules. See, e.g., Moore v. Publicis Groupe SA, No. 11 Civ.
1279(ALC)(AJP), 2012 WL 1446534, at *1, *3 (S.D.N.Y. Apr. 26, 2012) (overruling
plaintiff’s objections “that the predictive coding method contemplated in the ESI protocol
1
The following two terms, ((message* or messenger or titan) and (spam* or filter or “junk” or
“unsolicited”)” and ((message* or messenger or titan) w/25 (process*)), added 146,901 and 183,497 unique
documents, respectively, to the review set.
June 19, 2015
Page 3
lacks generally accepted reliability standards [and] that the use of such method violates Fed.
R. Civ. P. 26,” and instead affirming the magistrate judge’s decision, saying that “under the
circumstances of this particular case, the use of the predictive coding software as specified in
the ESI protocol is more appropriate than keyword searching”); Rio Tinto PLC v. Vale S.A.,
No. 14-3042, 2015 WL 872294, at *1 (S.D.N.Y. Mar. 2, 2015) (“In the three years since Da
Silva Moore, the case law has developed to the point that it is now black letter law that where
the producing party wants to utilize TAR [technology assisted review] for document review,
courts will permit it.”).
The model is trained from a subset of documents that we manually reviewed and can
be iteratively strengthened to improve accuracy.2 Our goal is to continue iterating the model
until we achieve a recall rate that returns a statistically significant and industry-accepted
percentage of relevant documents, when applied to a subset that was manually reviewed for
relevance. We are utilizing these methods in order to identify the most relevant documents
from an enormous set, in order to permit them to be reviewed and produced as fast as
possible. We would welcome the opportunity to confer with you regarding these methods
and your thoughts on a fair, reasonable, and proportionate review process. See N.D. Cal. ESI
Guideline 2.02 (recommending conferring regarding “[o]pportunities to reduce costs and
increase efficiency and speed, such as by conferring about the methods and technology used
for searching ESI to help identify the relevant information and sampling methods to validate
the search for relevant information”).
To the extent Plaintiffs have any objections to this approach, we should discuss them
as soon as possible. We would remind you that Judge Hamilton noted at the Case
Management Conference that she “agreed” with Facebook’s position that the parties should
focus their discovery efforts on the open factual issues identified in her Motion to Dismiss
ruling, as well as issues related to class certification. When we discuss a resolution of any
open disputes, it will be most constructive for you to identify the priority items that Plaintiffs
require for these issues, and consider “tabling” any unrelated requests. In addition, the chart
of requests attached to your letter includes several items to which Facebook objected and/or
to which no responsive documents exist.
2
Certain documents are excluded from the predictive coding process because they do not have quality text
for the model to analyze, including multimedia files, picture files, system files and documents with very
little text.
June 19, 2015
Page 4
We look forward to a constructive dialogue on how to resolve these issues. Rather
than continuing to send self-serving (and misleading) letters, you should let us know when
you are available for a call next week, or confirm that you are available for an in-person
meeting on June 24.
Sincerely,
Christopher Chorba
cc:
All Counsel of Record
Disclaimer: Justia Dockets & Filings provides public litigation records from the federal appellate and district courts. These filings and docket sheets should not be considered findings of fact or liability, nor do they necessarily reflect the view of Justia.
Why Is My Information Online?