Memorandum in Support re [229] MOTION to Compel Production of Successor-Custodian Documents and Relevant Source Code filed by Commonwealth of Virginia, State of Arizona, State of California, State of Colorado, State of Connecticut, State of Illinois, State of Michigan, State of Minnesota, State of Nebraska, State of New Hampshire, State of New Jersey, State of New York, State of North Carolina, State of Rhode Island, State of Tennessee, State of Washington, State of West Virginia, United States of America. (Teitelbaum, Aaron)
Page 1 PageID#
IN THE UNITED STATES DISTRICT COURT
FOR THE EASTERN DISTRICT OF VIRGINIA
Alexandria Division
UNITED STATES, et al.,
v.
Plaintiffs,
GOOGLE LLC,
Defendant.
)
)
)
)
)
)
)
)
No. 1:23-cv-00108-LMB-JFA
MEMORANDUM OF LAW IN SUPPORT OF
PLAINTIFFS’ MOTION TO COMPEL PRODUCTION OF
SUCCESSOR-CUSTODIAN DOCUMENTS AND RELEVANT SOURCE CODE
Pursuant to Rule 37(a) of the Federal Rules of Civil Procedure and Local Civil Rule
37(A), Plaintiffs, the United States and Co-Plaintiff States, respectfully submit this memorandum
of law in support of Plaintiffs’ motion to compel discovery from Defendant Google in the abovecaptioned action. Plaintiffs respectfully request that the Court compel Defendant Google to: (1)
search and produce relevant materials from the files of successor custodians in key positions at
Google between October 2020 and January 2023; and (2) produce source code and related
explanatory documents concerning the algorithms and other computer code utilized by Google’s
ad tech tools that are at the heart of this action. The parties discussed these topics at length at
several meet and confers, during which Plaintiffs described the relevance and importance of the
information sought and proposed reasonable compromises for Google’s consideration. Google
rejected Plaintiffs’ proposals, failed to timely offer additional information to facilitate
discussions, and made insufficient counterproposals, rendering all three issues ripe for court
intervention.Page 2 PageID#
BACKGROUND
A. Google’s Proposed Document Custodian List is Out-of-Date and Should Be
Revised to Include Any New Custodians Who Have Assumed Key Positions at
Google Since October 2020.
On April 11, in response to Plaintiffs’ March 27, 2023 initial requests for production of
documents (“RFPs”), Google provided a list of proposed custodians whose documents would be
searched for relevant material. Google’s list of proposed custodians was based largely on
custodians that had been agreed-upon during the United States’ pre-complaint investigation of
Google’s conduct in approximately October 2020. Plaintiffs promptly requested that Google
supplement its proposed custodian list to add a limited set of targeted custodians, including
successors of the employees identified by Google as relevant custodians in its April 11 letter (the
“Successor Custodians”), to ensure that Google’s search for relevant documents included
documents from the files of Google employees in key positions across Google’s ad tech
business, including any employees who may have come into those key positions after October
2020.
From the start of this case, Plaintiffs have made clear that while Google produced a
substantial number of documents during the United States’ pre-complaint investigation of
Google’s conduct, Plaintiffs would need to supplement this information with more recent
relevant documents. 1 Plaintiffs made a reasonable proposal to add additional custodians who had
assumed the same key positions in Google’s digital advertising business, the very same positions
at Google that had warranted inclusion in the custodian list back in October 2020.
At the first Scheduling Hearing, Plaintiffs explained that a refreshed production would require
the addition of “new custodians who joined Google in certain key positions that are relevant to
the facts of the lawsuit.” Mar. 24, 2023, Sch. Hearing at 6:1-4. Page 3 PageID#
Well over a month passed after Plaintiffs requested the addition of these Successor
Custodians, and despite frequent meetings to discuss the issue, 2 Google has yet to confirm that it
will add Successor Custodians to its Custodian List. On May 5, Google represented that
compiling a list of Successor Custodians—a list of its own employees serving in key positions
across the ad tech stack—was too burdensome. To address this surprising claim of burden, on
May 10, Plaintiffs provided Google a list of 63 individuals who Plaintiffs believed, based solely
on public records searches, no longer held the titles they had held on or before October 2020, or
where it was unclear who held those positions. All Google had to do was consult with its own
employees to determine who the relevant Successor Custodians were, and agree to add those
individuals to the list.
On May 26, at 2:30 p.m. ET, Google sent additional information about 45 of the Google employees identified in Plaintiffs’ May 10 email. Google confirmed that many of the
individuals Plaintiffs inquired about have already been added to the list of agreed-upon
custodians in this litigation. Yet, with respect to at least ten of these individuals, Google
identified successors without any indication of whether or not Google would agree to add the
narrower set of Successor Custodians to its custodian list. Plaintiffs still have no information
about the remaining 18 Google employees identified in our May 10 email.
Given that the July 7, 2023 deadline for substantial completion of document production is
quickly approaching, Plaintiffs are concerned that Google has not yet confirmed whether it will
agree to include these Successor Custodians in its search for relevant documents. For that reason,
Plaintiffs regrettably feel compelled to seek the Court’s assistance in resolving the issue. To that
end, we respectfully request that the Court compel Google to complete its search for and produce
Exhibit 1, attached to this motion, sets out Plaintiffs’ efforts to reach resolution on these topics. Page 4 PageID#
relevant materials from the custodial files of successor employees described in Plaintiffs’ May
10 correspondence, or otherwise confirm that the employees are already included on Google’s
custodian list.
B. Plaintiffs are Entitled to the Relevant Source Code and Associated
Explanatory Documents That Are at the Heart of This Case.
Plaintiffs’ first set of requests for production, served April 27, 2023, included requests through 43 (the “Source Code RFPs”), which sought the production of source code from the
products and ad tech features described in Plaintiffs’ complaint, 3 as well as data and documents
explaining that source code. Specifically, Request for Production No. 39 (“RFP 39”) requests:
Source code repositories and related repositories, including training
data, launch or release dates, pseudocode, and the implementation
status or roll-out associated with that source code, relating to the
following features and any successors or updates to those features:
(1) Project Bernanke (including Global Bernanke and Project Bell);
(2) Project Poirot (also known as optimized fixed CPM bidding and
Automated Bidding); (3) AdX’s (sellside) dynamic revenue share;
(4) AdX’s predicted highest other bidder (pHOB) algorithm; (5)
Smart Bidding, including any inputs to and outputs from the feature
or program, including but not limited to minimum bid to win; (6)
Dynamic Allocation; (7) Enhanced Dynamic Allocation; (8) Project
Bell; (9) Reserve Price Optimization (RPO); (10) any curation
function for the Related Products that determine which exchanges,
DSPs, and Ad Networks to call; (11) any bid, price floor, or auction
optimization algorithm, product, or feature; and (12) any automated
means or manner of bidding, whether on individual bids or on a
campaign-level basis.
Requests for Production Nos. 40 and 41 also requested documents and data related to
RFP 39. The two subsequent requests also ask for source code:
REQUEST NO. 42. Source code repositories and related
repositories, including training data, launch or release dates,
In the parties’ Modified Protective Order, “source code” is described as “computer code and
associated comments and revision histories, formulas, engineering specifications, or schematics
that define or otherwise describe in detail the algorithms or structure of software or hardware
designs . . . .” ECF No. 203, at 7-8. Page 5 PageID#
pseudocode, and the implementation status or roll-out associated
with that source code, generated since January 1, 2010, related to:
(1) how bids are selected and submitted in any product that
generates, alters, or submits a bid to purchase an impression in any
Relevant Product or in AdMeld; (2) the way or method by which
Relevant Products and AdMeld determine or determined the
winners of any real-time auction, waterfall bidding, and any sales of
impressions.
REQUEST NO. 43. All documents created since January 1, 2005,
related to plain language explanations, code repositories,
pseudocode, design documents, A/B or comparison testing, and any
other manuals or resources explaining the features, functions, and
operation of the source code used to: (1) determine the selection of
bids into or sent from any Relevant Product or in AdMeld; and (2)
determine the winners of any real-time auction, waterfall bidding,
and any sales of impressions.
The parties have had numerous meetings to discuss the production of source code and
related materials, and Plaintiffs have explained repeatedly why access to source code and related
materials is crucial to allow Plaintiffs a fair opportunity to prosecute their case and understand
Google’s potential defenses.
To address Google’s stated concerns about the burden and/or risks associated with
producing source code information, Plaintiffs alternatively offered to accept “pseudocode” –
descriptions of the steps in an algorithm or other software program in plain and natural language
– and related documentation instead, provided that Plaintiffs’ experts be given an opportunity to
review such pseudocode and determine whether it would be a sufficient substitute for the source
code itself. Yet, after waiting weeks to get back to Plaintiffs, on the parties’ May 24 meet and
confer, counsel for Google stated that it could not agree to the proposal because Google did not
have a “complete” set of such pseudocode. Instead, Google told Plaintiffs that it was preparing a
counterproposal to provide on May 30, which is when Google expected to offer MDL plaintiffs
its proposed response to similar RFPs served in that litigation. Page 6 PageID#
On May 26, at 2:30 p.m. ET, Google finally sent a concrete proposal, indicating that it
was prepared to search for and make available, if located, certain but not all of the source code
requested by Plaintiffs. Google’s proposal is silent with respect to the production of available
pseudocode. Plaintiffs do not know whether Google will produce relevant manuals, design
documents, or any other resources necessary for Plaintiffs’ experts to understand the relevant
Source Code. Plaintiffs do not know when to expect Google to have located the relevant Source
Code—their proposal suggests it may not be able to be located and, in fact, may not exist.
Without further progress on this issue, and in light of the limited time available to review
and analyze such intricate and complex source code that goes to the heart of the complaint’s
allegations, Plaintiffs respectfully request Court intervention to resolve this issue once and for
all. Of course, Plaintiffs will continue to meet and confer with Google while this motion is
pending and if further progress can be made on this issue in the interim, Plaintiffs will happily
table this aspect of the motion.
ARGUMENT
The Federal Rules of Civil Procedure provide a broad baseline for discovery as a party is
entitled “to discover any material that is relevant to any party’s claim or defense, is
nonprivileged, and is proportional to the needs of the case.” Fed. R. Civ. P. 26(b)(1); BASF Plant
Sci., LP v. Commonwealth Sci. & Indus. Rsch. Org., 2019 WL 8108060, at *2 (E.D. Va. July 3,
2019) (citing Fed. R. Civ. P. 26(b)(1)). This right to relevant discovery includes “[a]ny matter
that bears on, or that reasonably could lead to other matter that could bear on, any issue that is or
may be in the case.” Tube-Mac Indus., Inc. v. Campbell, No. 2:20CV197, 2021 WL 6332354, at
*1 (E.D. Va. June 10, 2021) (quoting Oppenheimer Fund, Inc. v. Sanders, 437 U.S. 340, (1978)). Federal Rule of Civil Procedure 37 and Local Civil Rule 37(A) allow parties to seek a Page 7 PageID#
court order to compel responses to discovery requests. The party refusing discovery “generally
bears the burden to prove why the requested evidence is not discoverable.” BASF Plant Sci.,
2019 WL 8108060, at *2. As discussed below, Plaintiffs have requested custodial documents and
carefully-tailored requests for production that are both highly relevant to their allegations and
Google’s defenses.
A. Google’s Burden Objections to Plaintiffs’ Request for Successor Custodian
Documents Are Unpersuasive.
Plaintiffs have made a simple request: Google should collect documents from custodians
who have occupied roles relevant to this litigation for the entire relevant time period—including
the most recent, post-October 2020 time period. Despite this straightforward request, Google has
claimed that identifying Google employees who have held these roles during recent years is itself
too burdensome of a request. In an attempt to understand the burden, Plaintiffs asked whether
Google would be willing to make a few phone calls to its Human Resources (or equivalent)
departments to track down the relevant information. Plaintiffs were told Google would have to
conduct “dozens and dozens of interviews,” and Google could not just “press a button” for this
information. Plaintiffs then made efforts to lessen Google’s burden by running searches on social
media sites and other publicly available sources to begin to identify individuals who have left the
company since October 2020. Despite Plaintiffs’ best efforts to assist Google in this process,
Plaintiffs are not in a position to determine and verify a list of relevant custodians covering the
post-October 2020 time period—only Google has access to its own employment records and
databases, and it does not seem a burdensome task for a company with multiple law firms
representing it to do what is necessary to ensure that its document custodian list reflects current
employees in relevant positions. Page 8 PageID#
After nearly two months of negotiations and with barely more than a month remaining
before the substantial completion deadline—two and a half hours before Plaintiffs’ deadline to
file this motion—Google provided information about forty-five Google employees, including the
names of eleven successors not yet on Google’s custodian list. Google has not said whether
Google would agree to add any of the eleven newly-identified successors to its custodian list.
And there are still eighteen employees identified by Plaintiffs who hold job titles that suggest
their documents would be highly relevant and responsive to Plaintiffs’ discovery requests but
who nonetheless may or may not already be included on Google’s list of custodians. 4 Google’s
May 26 letter says nothing about these eighteen employees. Plaintiffs respectfully request that
the Court order Google to promptly identify any new custodians necessary to add to its custodian
list in order to cover the post-October 2020 time period and promptly produce responsive
documents from those employees’ custodial files.
B. Plaintiffs’ Experts Need the Opportunity to Examine the Source Code and
Related Documents That Are at the Heart of Plaintiffs’ Allegations.
Courts routinely order the production of source code where it is deemed relevant. See
Trustees of Columbia Univ. v. Symantec Corp., 2014 WL 12979655, at *1 (E.D. Va. Apr. 22,
2014) (ordering patent infringer to produce source code for its accused product, rejecting burden
argument); see also Spendlove v. RapidCourt, LLC, 2019 WL 7143664, at *7-8 (E.D. Va. Dec.
23, 2019) (finding Plaintiffs’ need for source code outweighed harm to Defendants, whose
interests could be protected by an appropriate order). Even though source code may be
competitively sensitive, “[t]here is, however, no source code exception to the production
It is possible that a substantial portion of these identified custodians are already on Google’s
custodian list, which would substantially diminish any purported burden on Google from adding
these names to the custodian list. Page 9 PageID#
requirements of Fed. R. Civ. P. 34.” Apple Inc. v. Samsung Elecs. Co., 2012 WL 1595784, at *(N.D. Cal. 2012) (sanctioning defendant for failing to produce source code pursuant to order).
While Google claimed in its responses to Plaintiffs’ first requests for production that Plaintiffs
must make a showing that it is “necessary” to receive source code, the standard to produce
source code is not one of necessity, but of relevance. Spendlove, 2019 WL 7143664 at *(finding that “courts have required [parties] to produce source code and/or algorithm information
demonstrating how their software works when relevant to the litigation at issue.”).
Source code and the documents that supplement and explain the workings of source code,
are highly relevant to Plaintiffs’ claims and Google’s defenses. How the algorithms and products
work, including how bids are curated and auction winners are chosen, are central disputed issues
in this case. Indeed, throughout the complaint, Plaintiffs allege that Google’s algorithms and
features were developed to give Google an unfair anticompetitive advantage. Many of the
anticompetitive acts alleged are, at a very basic level, Google’s development and implementation
of source code by engineers and project managers. For that reason, accessing and understanding
the source code is imperative. Plaintiffs’ experts must be given a full and fair opportunity to
verify, analyze, and describe to a jury precisely what Google did that allowed it to obtain an
unfair competitive advantage.
Plaintiffs’ experts intend to use Google’s source code (along with the documents and data
explaining this source code) to explain how the central products and features in this case operate.
This information will allow Plaintiffs’ experts to explain the operation of Google’s features and
algorithms to the jury, as well as enable Plaintiffs’ experts to develop accurate models of
Google’s conduct. Page 10 PageID#
During the investigation, the United States attempted to understand key functions of
Google’s products through its investigative depositions of Google employees. But those efforts
were thwarted by witnesses that disclaimed specific knowledge about the products and their
precise features. In making the requested source code available, Google has the benefit of a Protective
Order that it helped draft and prepare and which specifically contemplates the production of
source code. See ECF No. 203. This Protective Order specifically mentions and defines “Source
Code”; lists “Source Code” in its bucket of “Highly Confidential” information; and provides a
detailed ten-page “Source Code Protocol” for the step-by-step production of source code. Id. at
App. C. As another court has held: “There are often protections put in place for the disclosure
and viewing of such sensitive information. With such protections in place, there is no elevated
standard when seeking the discovery of source code; a party need only satisfy the general
provisions of RCFC 26(b)(1) when seeking the ‘discovery of the source code as relevant to its
claims.’” Thales Visionix, Inc. v. United States, 149 Fed. Cl. 38, 48 (Fed. Cl. 2020) (quoting
Baron Servs., Inc. v. Media Weather Innovations LLC, 717 F.3d 907, 913 n.9 (Fed. Cir. 2013))
(emphasis added). Here, Google has all the protections it needs and has advocated for, and its
claims of “substantial risk of serious harm” cannot be permitted to prevent the production of
essential source code where Google has sought, and agreed to, detailed and onerous protections
in a Protective Order.
Google argues in its responses to Plaintiffs’ first requests for production that the materials
received in discovery during the investigation should be adequate for Plaintiffs’ needs. But as the
United States has stated to Google on multiple meet and confers, the United States did not
receive a full and complete production of source code and source code-related documents during
the investigation. Page 11 PageID#
Nonetheless, Plaintiffs attempted to find a reasonable way to respond to Google’s
concerns while maintaining Plaintiffs’ ability to obtain necessary information through discovery.
But in light of Google’s failure to promptly provide any workable counterproposal, Plaintiffs
must request relief from this Court. As with the previous request, Plaintiffs will continue to meet
and confer with Google while this motion is pending in the hopes that an agreement on this
important question can be reached promptly.
CONCLUSION
For the foregoing reasons, Plaintiffs respectfully request that the Court compel Google to
produce documents from the files of Successor Custodians and to make full and complete
responses to Plaintiffs’ RFPs 39-43. Page 12 PageID#
Dated: May 26, Respectfully submitted,
JESSICA D. ABER
United States Attorney
JASON S. MIYARES
Attorney General of Virginia
/s/ Gerard Mene
GERARD MENE
Assistant U.S. Attorney
2100 Jamieson Avenue
Alexandria, VA Telephone: (703) 299-Facsimile: (703) 299-Email: Gerard.Mene@usdoj.gov
/s/ Andrew N. Ferguson
ANDREW N. FERGUSON
Solicitor General
STEVEN G. POPPS
Deputy Attorney General
TYLER T. HENRY
Assistant Attorney General
/s/ Julia Tarver Wood
JULIA TARVER WOOD
/s/ Michael J. Freeman
MICHAEL J. FREEMAN
/s/ Kelly D. Garcia
KELLY D. GARCIA
/s/ Amanda M. Strick
AMANDA M. STRICK
/s/ Aaron M. Teitelbaum
AARON M. TEITELBAUM
United States Department of Justice
Antitrust Division
450 Fifth Street NW, Suite Washington, DC Telephone: (202) 307-Fax: (202) 616-Email: Kelly.Garcia@usdoj.gov
Office of the Attorney General of Virginia
202 North Ninth Street
Richmond, VA Telephone: (804) 692-Facsimile: (804) 786-Email: thenry@oag.state.va.us
Attorneys for the Commonwealth of
Virginia and local counsel for the
States of Arizona, California,
Colorado, Connecticut, Illinois,
Michigan, Minnesota, Nebraska, New
Hampshire, New Jersey, New York,
North Carolina, Rhode Island,
Tennessee, Washington, and West
Virginia
Attorneys for the United States
PDF Page 1
PlainSite Cover Page
PDF Page 2
Case 1:23-cv-00108-LMB-JFA Document 230 Filed 05/26/23 Page 1 of 12 PageID# 1937
IN THE UNITED STATES DISTRICT COURT
FOR THE EASTERN DISTRICT OF VIRGINIA
Alexandria Division
UNITED STATES, et al.,
v.
Plaintiffs,
GOOGLE LLC,
Defendant.
)
)
)
)
)
)
)
)
No. 1:23-cv-00108-LMB-JFA
MEMORANDUM OF LAW IN SUPPORT OF
PLAINTIFFS’ MOTION TO COMPEL PRODUCTION OF
SUCCESSOR-CUSTODIAN DOCUMENTS AND RELEVANT SOURCE CODE
Pursuant to Rule 37(a) of the Federal Rules of Civil Procedure and Local Civil Rule
37(A), Plaintiffs, the United States and Co-Plaintiff States, respectfully submit this memorandum
of law in support of Plaintiffs’ motion to compel discovery from Defendant Google in the abovecaptioned action. Plaintiffs respectfully request that the Court compel Defendant Google to: (1)
search and produce relevant materials from the files of successor custodians in key positions at
Google between October 2020 and January 2023; and (2) produce source code and related
explanatory documents concerning the algorithms and other computer code utilized by Google’s
ad tech tools that are at the heart of this action. The parties discussed these topics at length at
several meet and confers, during which Plaintiffs described the relevance and importance of the
information sought and proposed reasonable compromises for Google’s consideration. Google
rejected Plaintiffs’ proposals, failed to timely offer additional information to facilitate
discussions, and made insufficient counterproposals, rendering all three issues ripe for court
intervention.
PDF Page 3
Case 1:23-cv-00108-LMB-JFA Document 230 Filed 05/26/23 Page 2 of 12 PageID# 1938
BACKGROUND
A. Google’s Proposed Document Custodian List is Out-of-Date and Should Be
Revised to Include Any New Custodians Who Have Assumed Key Positions at
Google Since October 2020.
On April 11, in response to Plaintiffs’ March 27, 2023 initial requests for production of
documents (“RFPs”), Google provided a list of proposed custodians whose documents would be
searched for relevant material. Google’s list of proposed custodians was based largely on
custodians that had been agreed-upon during the United States’ pre-complaint investigation of
Google’s conduct in approximately October 2020. Plaintiffs promptly requested that Google
supplement its proposed custodian list to add a limited set of targeted custodians, including
successors of the employees identified by Google as relevant custodians in its April 11 letter (the
“Successor Custodians”), to ensure that Google’s search for relevant documents included
documents from the files of Google employees in key positions across Google’s ad tech
business, including any employees who may have come into those key positions after October
2020.
From the start of this case, Plaintiffs have made clear that while Google produced a
substantial number of documents during the United States’ pre-complaint investigation of
Google’s conduct, Plaintiffs would need to supplement this information with more recent
relevant documents. 1 Plaintiffs made a reasonable proposal to add additional custodians who had
assumed the same key positions in Google’s digital advertising business, the very same positions
at Google that had warranted inclusion in the custodian list back in October 2020.
1
At the first Scheduling Hearing, Plaintiffs explained that a refreshed production would require
the addition of “new custodians who joined Google in certain key positions that are relevant to
the facts of the lawsuit.” Mar. 24, 2023, Sch. Hearing at 6:1-4.
2
PDF Page 4
Case 1:23-cv-00108-LMB-JFA Document 230 Filed 05/26/23 Page 3 of 12 PageID# 1939
Well over a month passed after Plaintiffs requested the addition of these Successor
Custodians, and despite frequent meetings to discuss the issue, 2 Google has yet to confirm that it
will add Successor Custodians to its Custodian List. On May 5, Google represented that
compiling a list of Successor Custodians—a list of its own employees serving in key positions
across the ad tech stack—was too burdensome. To address this surprising claim of burden, on
May 10, Plaintiffs provided Google a list of 63 individuals who Plaintiffs believed, based solely
on public records searches, no longer held the titles they had held on or before October 2020, or
where it was unclear who held those positions. All Google had to do was consult with its own
employees to determine who the relevant Successor Custodians were, and agree to add those
individuals to the list.
On May 26, at 2:30 p.m. ET, Google sent additional information about 45 of the 63
Google employees identified in Plaintiffs’ May 10 email. Google confirmed that many of the
individuals Plaintiffs inquired about have already been added to the list of agreed-upon
custodians in this litigation. Yet, with respect to at least ten of these individuals, Google
identified successors without any indication of whether or not Google would agree to add the
narrower set of Successor Custodians to its custodian list. Plaintiffs still have no information
about the remaining 18 Google employees identified in our May 10 email.
Given that the July 7, 2023 deadline for substantial completion of document production is
quickly approaching, Plaintiffs are concerned that Google has not yet confirmed whether it will
agree to include these Successor Custodians in its search for relevant documents. For that reason,
Plaintiffs regrettably feel compelled to seek the Court’s assistance in resolving the issue. To that
end, we respectfully request that the Court compel Google to complete its search for and produce
2
Exhibit 1, attached to this motion, sets out Plaintiffs’ efforts to reach resolution on these topics.
3
PDF Page 5
Case 1:23-cv-00108-LMB-JFA Document 230 Filed 05/26/23 Page 4 of 12 PageID# 1940
relevant materials from the custodial files of successor employees described in Plaintiffs’ May
10 correspondence, or otherwise confirm that the employees are already included on Google’s
custodian list.
B. Plaintiffs are Entitled to the Relevant Source Code and Associated
Explanatory Documents That Are at the Heart of This Case.
Plaintiffs’ first set of requests for production, served April 27, 2023, included requests 39
through 43 (the “Source Code RFPs”), which sought the production of source code from the
products and ad tech features described in Plaintiffs’ complaint, 3 as well as data and documents
explaining that source code. Specifically, Request for Production No. 39 (“RFP 39”) requests:
Source code repositories and related repositories, including training
data, launch or release dates, pseudocode, and the implementation
status or roll-out associated with that source code, relating to the
following features and any successors or updates to those features:
(1) Project Bernanke (including Global Bernanke and Project Bell);
(2) Project Poirot (also known as optimized fixed CPM bidding and
Automated Bidding); (3) AdX’s (sellside) dynamic revenue share;
(4) AdX’s predicted highest other bidder (pHOB) algorithm; (5)
Smart Bidding, including any inputs to and outputs from the feature
or program, including but not limited to minimum bid to win; (6)
Dynamic Allocation; (7) Enhanced Dynamic Allocation; (8) Project
Bell; (9) Reserve Price Optimization (RPO); (10) any curation
function for the Related Products that determine which exchanges,
DSPs, and Ad Networks to call; (11) any bid, price floor, or auction
optimization algorithm, product, or feature; and (12) any automated
means or manner of bidding, whether on individual bids or on a
campaign-level basis.
Requests for Production Nos. 40 and 41 also requested documents and data related to
RFP 39. The two subsequent requests also ask for source code:
REQUEST NO. 42. Source code repositories and related
repositories, including training data, launch or release dates,
3
In the parties’ Modified Protective Order, “source code” is described as “computer code and
associated comments and revision histories, formulas, engineering specifications, or schematics
that define or otherwise describe in detail the algorithms or structure of software or hardware
designs . . . .” ECF No. 203, at 7-8.
4
PDF Page 6
Case 1:23-cv-00108-LMB-JFA Document 230 Filed 05/26/23 Page 5 of 12 PageID# 1941
pseudocode, and the implementation status or roll-out associated
with that source code, generated since January 1, 2010, related to:
(1) how bids are selected and submitted in any product that
generates, alters, or submits a bid to purchase an impression in any
Relevant Product or in AdMeld; (2) the way or method by which
Relevant Products and AdMeld determine or determined the
winners of any real-time auction, waterfall bidding, and any sales of
impressions.
REQUEST NO. 43. All documents created since January 1, 2005,
related to plain language explanations, code repositories,
pseudocode, design documents, A/B or comparison testing, and any
other manuals or resources explaining the features, functions, and
operation of the source code used to: (1) determine the selection of
bids into or sent from any Relevant Product or in AdMeld; and (2)
determine the winners of any real-time auction, waterfall bidding,
and any sales of impressions.
The parties have had numerous meetings to discuss the production of source code and
related materials, and Plaintiffs have explained repeatedly why access to source code and related
materials is crucial to allow Plaintiffs a fair opportunity to prosecute their case and understand
Google’s potential defenses.
To address Google’s stated concerns about the burden and/or risks associated with
producing source code information, Plaintiffs alternatively offered to accept “pseudocode” –
descriptions of the steps in an algorithm or other software program in plain and natural language
– and related documentation instead, provided that Plaintiffs’ experts be given an opportunity to
review such pseudocode and determine whether it would be a sufficient substitute for the source
code itself. Yet, after waiting weeks to get back to Plaintiffs, on the parties’ May 24 meet and
confer, counsel for Google stated that it could not agree to the proposal because Google did not
have a “complete” set of such pseudocode. Instead, Google told Plaintiffs that it was preparing a
counterproposal to provide on May 30, which is when Google expected to offer MDL plaintiffs
its proposed response to similar RFPs served in that litigation.
5
PDF Page 7
Case 1:23-cv-00108-LMB-JFA Document 230 Filed 05/26/23 Page 6 of 12 PageID# 1942
On May 26, at 2:30 p.m. ET, Google finally sent a concrete proposal, indicating that it
was prepared to search for and make available, if located, certain but not all of the source code
requested by Plaintiffs. Google’s proposal is silent with respect to the production of available
pseudocode. Plaintiffs do not know whether Google will produce relevant manuals, design
documents, or any other resources necessary for Plaintiffs’ experts to understand the relevant
Source Code. Plaintiffs do not know when to expect Google to have located the relevant Source
Code—their proposal suggests it may not be able to be located and, in fact, may not exist.
Without further progress on this issue, and in light of the limited time available to review
and analyze such intricate and complex source code that goes to the heart of the complaint’s
allegations, Plaintiffs respectfully request Court intervention to resolve this issue once and for
all. Of course, Plaintiffs will continue to meet and confer with Google while this motion is
pending and if further progress can be made on this issue in the interim, Plaintiffs will happily
table this aspect of the motion.
ARGUMENT
The Federal Rules of Civil Procedure provide a broad baseline for discovery as a party is
entitled “to discover any material that is relevant to any party’s claim or defense, is
nonprivileged, and is proportional to the needs of the case.” Fed. R. Civ. P. 26(b)(1); BASF Plant
Sci., LP v. Commonwealth Sci. & Indus. Rsch. Org., 2019 WL 8108060, at *2 (E.D. Va. July 3,
2019) (citing Fed. R. Civ. P. 26(b)(1)). This right to relevant discovery includes “[a]ny matter
that bears on, or that reasonably could lead to other matter that could bear on, any issue that is or
may be in the case.” Tube-Mac Indus., Inc. v. Campbell, No. 2:20CV197, 2021 WL 6332354, at
*1 (E.D. Va. June 10, 2021) (quoting Oppenheimer Fund, Inc. v. Sanders, 437 U.S. 340, 351
(1978)). Federal Rule of Civil Procedure 37 and Local Civil Rule 37(A) allow parties to seek a
6
PDF Page 8
Case 1:23-cv-00108-LMB-JFA Document 230 Filed 05/26/23 Page 7 of 12 PageID# 1943
court order to compel responses to discovery requests. The party refusing discovery “generally
bears the burden to prove why the requested evidence is not discoverable.” BASF Plant Sci.,
2019 WL 8108060, at *2. As discussed below, Plaintiffs have requested custodial documents and
carefully-tailored requests for production that are both highly relevant to their allegations and
Google’s defenses.
A. Google’s Burden Objections to Plaintiffs’ Request for Successor Custodian
Documents Are Unpersuasive.
Plaintiffs have made a simple request: Google should collect documents from custodians
who have occupied roles relevant to this litigation for the entire relevant time period—including
the most recent, post-October 2020 time period. Despite this straightforward request, Google has
claimed that identifying Google employees who have held these roles during recent years is itself
too burdensome of a request. In an attempt to understand the burden, Plaintiffs asked whether
Google would be willing to make a few phone calls to its Human Resources (or equivalent)
departments to track down the relevant information. Plaintiffs were told Google would have to
conduct “dozens and dozens of interviews,” and Google could not just “press a button” for this
information. Plaintiffs then made efforts to lessen Google’s burden by running searches on social
media sites and other publicly available sources to begin to identify individuals who have left the
company since October 2020. Despite Plaintiffs’ best efforts to assist Google in this process,
Plaintiffs are not in a position to determine and verify a list of relevant custodians covering the
post-October 2020 time period—only Google has access to its own employment records and
databases, and it does not seem a burdensome task for a company with multiple law firms
representing it to do what is necessary to ensure that its document custodian list reflects current
employees in relevant positions.
7
PDF Page 9
Case 1:23-cv-00108-LMB-JFA Document 230 Filed 05/26/23 Page 8 of 12 PageID# 1944
After nearly two months of negotiations and with barely more than a month remaining
before the substantial completion deadline—two and a half hours before Plaintiffs’ deadline to
file this motion—Google provided information about forty-five Google employees, including the
names of eleven successors not yet on Google’s custodian list. Google has not said whether
Google would agree to add any of the eleven newly-identified successors to its custodian list.
And there are still eighteen employees identified by Plaintiffs who hold job titles that suggest
their documents would be highly relevant and responsive to Plaintiffs’ discovery requests but
who nonetheless may or may not already be included on Google’s list of custodians. 4 Google’s
May 26 letter says nothing about these eighteen employees. Plaintiffs respectfully request that
the Court order Google to promptly identify any new custodians necessary to add to its custodian
list in order to cover the post-October 2020 time period and promptly produce responsive
documents from those employees’ custodial files.
B. Plaintiffs’ Experts Need the Opportunity to Examine the Source Code and
Related Documents That Are at the Heart of Plaintiffs’ Allegations.
Courts routinely order the production of source code where it is deemed relevant. See
Trustees of Columbia Univ. v. Symantec Corp., 2014 WL 12979655, at *1 (E.D. Va. Apr. 22,
2014) (ordering patent infringer to produce source code for its accused product, rejecting burden
argument); see also Spendlove v. RapidCourt, LLC, 2019 WL 7143664, at *7-8 (E.D. Va. Dec.
23, 2019) (finding Plaintiffs’ need for source code outweighed harm to Defendants, whose
interests could be protected by an appropriate order). Even though source code may be
competitively sensitive, “[t]here is, however, no source code exception to the production
4
It is possible that a substantial portion of these identified custodians are already on Google’s
custodian list, which would substantially diminish any purported burden on Google from adding
these names to the custodian list.
8
PDF Page 10
Case 1:23-cv-00108-LMB-JFA Document 230 Filed 05/26/23 Page 9 of 12 PageID# 1945
requirements of Fed. R. Civ. P. 34.” Apple Inc. v. Samsung Elecs. Co., 2012 WL 1595784, at *1
(N.D. Cal. 2012) (sanctioning defendant for failing to produce source code pursuant to order).
While Google claimed in its responses to Plaintiffs’ first requests for production that Plaintiffs
must make a showing that it is “necessary” to receive source code, the standard to produce
source code is not one of necessity, but of relevance. Spendlove, 2019 WL 7143664 at *7
(finding that “courts have required [parties] to produce source code and/or algorithm information
demonstrating how their software works when relevant to the litigation at issue.”).
Source code and the documents that supplement and explain the workings of source code,
are highly relevant to Plaintiffs’ claims and Google’s defenses. How the algorithms and products
work, including how bids are curated and auction winners are chosen, are central disputed issues
in this case. Indeed, throughout the complaint, Plaintiffs allege that Google’s algorithms and
features were developed to give Google an unfair anticompetitive advantage. Many of the
anticompetitive acts alleged are, at a very basic level, Google’s development and implementation
of source code by engineers and project managers. For that reason, accessing and understanding
the source code is imperative. Plaintiffs’ experts must be given a full and fair opportunity to
verify, analyze, and describe to a jury precisely what Google did that allowed it to obtain an
unfair competitive advantage.
Plaintiffs’ experts intend to use Google’s source code (along with the documents and data
explaining this source code) to explain how the central products and features in this case operate.
This information will allow Plaintiffs’ experts to explain the operation of Google’s features and
algorithms to the jury, as well as enable Plaintiffs’ experts to develop accurate models of
Google’s conduct.
9
PDF Page 11
Case 1:23-cv-00108-LMB-JFA Document 230 Filed 05/26/23 Page 10 of 12 PageID# 1946
During the investigation, the United States attempted to understand key functions of
Google’s products through its investigative depositions of Google employees. But those efforts
were thwarted by witnesses that disclaimed specific knowledge about the products and their
precise features. 5
In making the requested source code available, Google has the benefit of a Protective
Order that it helped draft and prepare and which specifically contemplates the production of
source code. See ECF No. 203. This Protective Order specifically mentions and defines “Source
Code”; lists “Source Code” in its bucket of “Highly Confidential” information; and provides a
detailed ten-page “Source Code Protocol” for the step-by-step production of source code. Id. at
App. C. As another court has held: “There are often protections put in place for the disclosure
and viewing of such sensitive information. With such protections in place, there is no elevated
standard when seeking the discovery of source code; a party need only satisfy the general
provisions of RCFC 26(b)(1) when seeking the ‘discovery of the source code as relevant to its
claims.’” Thales Visionix, Inc. v. United States, 149 Fed. Cl. 38, 48 (Fed. Cl. 2020) (quoting
Baron Servs., Inc. v. Media Weather Innovations LLC, 717 F.3d 907, 913 n.9 (Fed. Cir. 2013))
(emphasis added). Here, Google has all the protections it needs and has advocated for, and its
claims of “substantial risk of serious harm” cannot be permitted to prevent the production of
essential source code where Google has sought, and agreed to, detailed and onerous protections
in a Protective Order.
5
Google argues in its responses to Plaintiffs’ first requests for production that the materials
received in discovery during the investigation should be adequate for Plaintiffs’ needs. But as the
United States has stated to Google on multiple meet and confers, the United States did not
receive a full and complete production of source code and source code-related documents during
the investigation.
10
PDF Page 12
Case 1:23-cv-00108-LMB-JFA Document 230 Filed 05/26/23 Page 11 of 12 PageID# 1947
Nonetheless, Plaintiffs attempted to find a reasonable way to respond to Google’s
concerns while maintaining Plaintiffs’ ability to obtain necessary information through discovery.
But in light of Google’s failure to promptly provide any workable counterproposal, Plaintiffs
must request relief from this Court. As with the previous request, Plaintiffs will continue to meet
and confer with Google while this motion is pending in the hopes that an agreement on this
important question can be reached promptly.
CONCLUSION
For the foregoing reasons, Plaintiffs respectfully request that the Court compel Google to
produce documents from the files of Successor Custodians and to make full and complete
responses to Plaintiffs’ RFPs 39-43.
11
PDF Page 13
Case 1:23-cv-00108-LMB-JFA Document 230 Filed 05/26/23 Page 12 of 12 PageID# 1948
Dated: May 26, 2023
Respectfully submitted,
JESSICA D. ABER
United States Attorney
JASON S. MIYARES
Attorney General of Virginia
/s/ Gerard Mene
GERARD MENE
Assistant U.S. Attorney
2100 Jamieson Avenue
Alexandria, VA 22314
Telephone: (703) 299-3777
Facsimile: (703) 299-3983
Email: Gerard.Mene@usdoj.gov
/s/ Andrew N. Ferguson
ANDREW N. FERGUSON
Solicitor General
STEVEN G. POPPS
Deputy Attorney General
TYLER T. HENRY
Assistant Attorney General
/s/ Julia Tarver Wood
JULIA TARVER WOOD
/s/ Michael J. Freeman
MICHAEL J. FREEMAN
/s/ Kelly D. Garcia
KELLY D. GARCIA
/s/ Amanda M. Strick
AMANDA M. STRICK
/s/ Aaron M. Teitelbaum
AARON M. TEITELBAUM
United States Department of Justice
Antitrust Division
450 Fifth Street NW, Suite 7100
Washington, DC 20530
Telephone: (202) 307-0077
Fax: (202) 616-8544
Email: Kelly.Garcia@usdoj.gov
Office of the Attorney General of Virginia
202 North Ninth Street
Richmond, VA 23219
Telephone: (804) 692-0485
Facsimile: (804) 786-0122
Email: thenry@oag.state.va.us
Attorneys for the Commonwealth of
Virginia and local counsel for the
States of Arizona, California,
Colorado, Connecticut, Illinois,
Michigan, Minnesota, Nebraska, New
Hampshire, New Jersey, New York,
North Carolina, Rhode Island,
Tennessee, Washington, and West
Virginia
Attorneys for the United States
12
Back to 
