Before the Court is Google’s Motion for Review of the Special Master’s Order Granting in Part Plaintiffs’ Motion to Compel Written Discovery Regarding Google Chats (the “Order”). (Dkt. #585). The Order addressed the Plaintiff States’ motion that Defendant Google LLC be compelled to respond to certain requests for so-called “discovery on discovery.” Discovery on discovery has been described as follows: “In certain circumstances where a party makes some showing that a producing party’s production has been incomplete, a court may order discovery designed to test the sufficiency of that party’s discovery efforts in order to capture additional relevant material.” Freedman v. Weatherford Int’l Ltd., No. 12 CIV. 2121 LAK JCF, 2014 WL 4547039, at *2 (S.D.N.Y. Sept. 12, 2014).

Here, the States asked that Google be required to: (1) provide them with the start and end dates for any legal holds sent to Google’s custodians since January 1, 2015 (Interrogatory No. 31); (2) produce any litigation holds or reminders that were issued to any custodians, or the like, since January 1, 2018 (Requests for Production Nos. 126 and 127); (3) produce the “system-wide backend log” previously produced in another federal case involving Google (Request for Production No. 133); and (4) produce the transcript and accompanying documents from the deposition of Google’s corporate representative from a case currently pending in Texas state court. See (Dkt. #558). The Special Master granted the motion in part. Specifically, he required Google to: (1) provide a full answer to the States’ Interrogatory No. 31; and (2) produce responsive documents to Requests for Production Nos. 126, 127, and 133, with certain modifications. The Order otherwise denied the States’ requests. See (Dkt. #558).

Google appeals two components of the Order: (1) the Special Master’s finding that the States have made a “preliminary showing of spoliation”; and (2) the requirement that Google produce information about litigation holds issued before the State of Texas announced its investigation of Google’s ad tech business in September 2019. (Dkt. #585 at 13). After reviewing the parties’ briefing and related record materials, the Court will grant Google’s appeal and modify the Order to confirm that no ruling has been made on spoliation at this time, preliminary or otherwise, and to require that Google produce litigation hold information only from the date that Texas announced its investigation of Google in September 2019.

I. STANDARD OF REVIEW

Under Federal Rule of Civil Procedure 53, the Court reviews the Special Master’s findings of fact and legal conclusions de novo. FED. R. CIV. P. 53(f)(3)–(4).

II. THE PRELIMINARY SHOWING OF SPOLIATION FINDING

The first component of the Special Master’s Order appealed by Google is his finding that the Plaintiff States made a preliminary showing of spoliation. Initially, the States requested that the Special Master compel the discovery of “any litigation holds or hold notices,” as well as “any litigation hold reminders . . . issued to any custodian, deponent, or person likely to have discoverable information identified in this matter.” (Dkt. #540 at 15) (citing Dkt. #348-2) (cleaned up). The States argued they required these litigation holds because

Google has testified that, prior to October 2021, its Chat Retention Policy did not discuss employees’ obligations to preserve chats when under legal/litigation holds and that those obligations were instead discussed in the hold notices themselves. And because Chats defaulted to “history-off” for employees under litigation holds until 2023, those notices were apparently the only way an employee would know to turn their history “on” and preserve their Chats.

(Dkt. #540 at 15) (emphasis added). Consequently, the States argued that the holds would “reveal (a) who was reminded to preserve relevant Chats and when, (b) what they were instructed to do, and thus (c) the scope and degree of Google’s noncompliance with its preservation obligations.” (Dkt. #540 at 15). The States further argued that the designated 30(b)(6) deponent had been unable to adequately speak to Google’s Chat preservation policies. (Dkt. #540 at 15).

Following briefing and a hearing, the Special Master ruled that “the States have made a preliminary showing of spoliation, that Google has not carried its burden to prove privilege, and that Google’s litigation holds are therefore discoverable in part.” (Dkt. #558 at 11) (emphasis added). This holding was driven by “ample evidence” supporting “a reasonable deduction that relevant Chats were not preserved or were destroyed and that Google may not have complied with its duty to preserve relevant documents and communications.” (Dkt. #558 at 9). The Special Master observed that “Chat history is only preserved if two steps are taken: (1) a Google employee must toggle his/her Chat history “on”; and (2) that employee must be subject to a litigation hold.” (Dkt. #558 at 8). Thus, “if an employee was not subject to a litigation hold at a particular date and time, [then] there is no possibility that the employee’s Chat history was preserved.” (Dkt. #558 at 8). For these reasons, the Special Master determined that “the start and end dates of Google’s litigation holds [are] a critical fact in this case.” (Dkt. #558 at 8).[1] 

The Court agrees with the substance of the Special Master’s ruling, i.e., that the requested discovery on discovery concerning litigation holds for Google Chat is warranted. Notably, this aspect of the Special Master’s ruling is unchallenged by Google. The only objection to the ruling is that it is framed, in part, as resulting from a “preliminary showing of spoliation.” Such a finding is not necessary to support the requested discovery on discovery, and the Court declines to reach any conclusions on spoliation at this time, preliminary or otherwise

Spoliation of evidence “is the destruction or the significant and meaningful alteration of evidence.” Guzman v. Jones, 804 F.3d 707, 713 (5th Cir. 2015) (quoting Rimkus Consulting Grp. v. Cammarata, 688 F.Supp.2d 598, 612 (S.D. Tex. 2010)). And relevant here, “[e]lectronically stored information is routinely deleted or altered and affirmative steps are often required to preserve it. Such deletions, alterations, and losses cannot be spoliation unless there is a duty to preserve the information, a culpable breach of that duty, and resulting prejudice.” Rimkus, 688 F.Supp.2d at 612.

The Court need not conclude that there has been “spoliation” to allow discovery on discovery. Discovery on discovery is permitted when deficiencies in a party’s production suggest that additional, relevant materials may exist or did exist. In this regard, and as a general matter, “[i]t is not the court’s role to dictate how a party should search for relevant information absent a showing that the party has abdicated its responsibility,” and “[a] responding party is best situated to preserve, search, and produce its own [electronically stored information],” which “[p]rinciple . . . is grounded in reason, common sense, procedural rules, and common law, and is premised on each party fulfilling its discovery obligations without direction from the court or opposing counsel[, and eschewing ‘discovery on discovery,’] unless a specific deficiency is shown in a party’s production.” Moore v. Westgate Resorts, No. 3:18-CV-410, 2020 WL 113352, at *12 (E.D. Tenn. Jan. 9, 2020) (cleaned up; quoting [and adding omitted language from] The Sedona Principles, Third Edition: Best Practices, Recommendations & Principles for Addressing Electronic Document Production, 19 SEDONA CONF. J. 1 (2018)).

To make this showing of a specific or material deficiency in the other party’s production to require the responding party to engage in additional searches or discovery efforts, or to obtain “discovery on discovery” that is “both relevant and proportional to the needs of the case” under Rule 26(b)(1), Uschold v. Carriage Servs., Inc., No. 17CV04424, 2019 WL 8298261, at *4 (N.D. Cal. Jan. 22, 2019), the requesting party must make a showing, including through “the documents that have been produced,” that allows the Court to make “a reasonable deduction that other documents may exist or did exist and have been destroyed” or must “point to the existence of additional responsive material,” Hubbard v. Potter, 247 F.R.D. 27, 29, 31 (D.D.C. 2008); accord Berkeley*IEOR v. Teradata Operations, Inc., No. 17 C 7472, 2021 WL 4306159, at *1 (N.D. Ill. Sept. 22, 2021) (collecting cases). The States have made this showing as to the specific, requested discovery on discovery at issue regarding Google’s litigation hold, as the Special Master properly concluded. No further finding is needed to support the Special Master’s discovery on discovery ruling, and Google does not challenge the substance of the ruling.

Accordingly, the Court modifies the Order to delete any suggestion that the Court has made a ruling on spoliation, preliminary or otherwise.

III. GOOGLE’S PRE-2019 LITIGATION HOLDS

The Order also provides that Google must provide the start and end dates for legal holds going back to January 1, 2015, over four years before the State of Texas announced its investigation into Google’s ad tech products. The Court will modify the Order to adjust the relevant date to September 9, 2019, the date Texas first announced its antitrust investigation into Google’s ad tech products.

A. The Special Master’s Ruling and the Arguments on Review

This issue arose when the States moved to compel a full answer to their Interrogatory No. 31, which requested Google provide “the start and end dates for any legal holds applicable to any custodian identified in this matter from January 1, 2015 to the present.”[2] (Dkt. #558 at 6). The Special Master ordered Google to answer the interrogatory and provide the requested pre-2019 litigation hold information, concluding that “Google anticipated litigation concerning the antitrust and regulatory issues regarding ad tech, including DoubleClick, far before 2019” and observing that “Google’s own privilege logs assert attorney client and work product privilege over DoubleClick and ad tech-related communications as far back as 2006.” (Dkt. #558 at 6) (cleaned up).

The Order demonstrates concern about Google’s reliance on privilege as a “sword and shield” as shown by Google’s “continued reliance on privilege objections grounded on anticipation of litigation as a shield to refuse the production of documents well before 2019, while simultaneously insisting that its preservation obligations in this case did not arise until 2019 when it first anticipated litigation upon notice from the State of Texas.” (Dkt. #558 at 7). The States agree with this argument on review, stating that the Special Master “was right to recognize this earlier obligation” since “Google foresaw this litigation as early as 2006.” (Dkt. #600 at 14).

Google argues that the Special Master’s holding lacks a legal basis because “Google could not have reasonably anticipated this litigation until Texas began its investigation in 2019” and because “the Special Master’s conclusion improperly conflates the standard for whether material is protected attorney work product under Federal Rule of Civil Procedure 26 with the standard for when a party’s preservation obligation begins under Rule 37.” (Dkt. #585 at 14) (emphasis added). As a result, Google maintains, “the ‘sword/shield’ implied waiver analogy adopted by the Special Master is erroneous” because it “is perfectly legitimate for a party that is investigated in 2019 to preserve work product protection over a document created in 2006 related to an earlier matter involving different parties.” (Dkt. #585 at 14). In short, Google contends that it had no duty to preserve potential evidence for the States until Texas began its investigation in 2019, and therefore any litigation hold information predating the commencement of Texas’s investigation should not be discoverable.

B. The Duty to Preserve Evidence

“A party’s duty to preserve evidence comes into being when the party has notice that the evidence is relevant to the litigation or should have known that the evidence may be relevant.” Van Winkle v. Rogers, 82 F.4th 370, 375 (5th Cir. 2023) (quoting Guzman, 804 F.3d at 713). Put differently, the obligation to preserve evidence “kicks in” when the party charged with spoliation should have reasonably anticipated litigation. Schafer v. Darr, No. CV 23-1387, 2024 WL 183587, at *2 (E.D. La. Jan. 17, 2024). “Identifying the trigger for when a party should have reasonably anticipated litigation is challenging, as it varies based on the facts and circumstances.” Coastal Bridge Co., LLC v. Heatec, Inc., 833 F. App’x 565, 574 (5th Cir. 2020). But that trigger is activated as soon as a potential claim is identified or litigation is reasonably foreseeable. Apple v. Samsung, 888 F. Supp. 2d 976, 991 (N.D. Cal. 2012). Once a party’s duty to preserve takes effect, it is “required to suspend any existing policies related to deleting or destroying files and preserve all relevant documents related to the litigation.” Id.

So when should a party reasonably have anticipated litigation? Courts agree that the duty to preserve is triggered when the defendant is served with a complaint. See, e.g., Nucor Corp. v. Bell, 251 F.R.D. 191, 197 (D.S.C. 2008). And courts also generally agree that, prior to the commencement of litigation, the receipt of a demand letter, a request for evidence preservation, a threat of litigation, or a decision to pursue a claim will all trigger the duty to preserve evidence. See, e.g., Turner v. United States, 736 F.3d 274, 282 (4th Cir. 2013) (explaining that a document preservation letter or a letter threatening litigation triggers a duty to preserve evidence); Sampson v. City of Cambridge, 251 F.R.D. 172, 181 (D. Md. 2008) (noting that a defendant’s duty to preserve arose when plaintiff’s counsel made a written request to preserve evidence).

Absent obvious triggers like letters threatening litigation or demanding document preservation in advance of the filing of suit, courts must look to the specific circumstances of each case to determine when the duty to preserve arose. As aptly explained by one court, although the general rules governing the duty to preserve evidence are not controversial, “applying them to determine when a duty to preserve arises in a particular case and the extent of that duty requires careful analysis of the specific facts and circumstances.” Rimkus, 688 F.Supp.2d at 613. “It can be difficult to draw bright-line distinctions between acceptable and unacceptable conduct in preserving information and in conducting discovery, either prospectively or with the benefit (and distortion) of hindsight.” Id.

In the regulatory context, courts have held that a government investigation does not necessarily create a duty as a matter of law to preserve evidence with respect to later civil litigation. Rockman Co. (USA), Inc. v. Nong Shim Co., Ltd, 229 F. Supp. 3d 1109, 1123 (N.D. Cal. 2017). Thus, the “start of an investigation is not necessarily the date on which litigation is reasonably foreseeable, because it is common that investigations can result in no litigation being filed.” In re Soc. Media Adolescent Addiction/Pers. Inj. Prod. Liab. Litig., No. 22-MD-03047-, 2024 WL 3498235, at *7 (N.D. Cal. July 19, 2024).

C. A Shifting Duty?

Another question raised by the parties’ dispute is, to whom is the duty owed to preserve evidence? The Court concludes that the duty to preserve potentially relevant evidence “arises when the party in possession of the evidence knows that litigation by the party seeking the evidence is pending or probable.” Kounelis v. Sherrer, 529 F.Supp.2d 503, 518 (D.N.J. 2008) (emphasis added); see also Point Blank Solutions, Inc. v. Toyobo Am., Inc., Case No. 09–61166–CIV, 2011 WL 1456029 at *24 (S.D. Fla. Apr. 5, 2011) (same); Realnetworks, Inc. v. DVD Copy Control Ass’n, Inc., 264 F.R.D. 517, 526–27 (N.D. Cal. 2009) (holding that a “general concern over litigation” concerning the product RealDVD did not “create a duty to preserve all documentation related to RealDVD” and that a duty to preserve arose only when “a potential claim was identified or future litigation was probable” with the parties in question).

The Court finds particularly instructive the analysis and decision in In re Delta/AirTran Baggage Fee Antitrust Litig., 770 F. Supp. 2d 1299 (N.D. Ga. 2011). In Delta, the United States District Court for the Northern District of Georgia considered whether a Civil Investigative Demand (“CID”) issued by the Department of Justice Antitrust Division (“DOJ”) triggered a duty on the part of Delta Airlines to preserve evidence for use in subsequent multidistrict litigation filed by passengers who were allegedly charged illegal first-bag fees. 770 F.Supp.2d at 1307–08. On February 2, 2009, the DOJ issued a CID requiring Delta to produce a variety of records relating to fees charged for checked baggage. Id. The following day, Delta’s assistant general counsel e-mailed a document preservation and litigation hold notice to a group of individuals identified as being potential document custodians. Id. Over the next few days, Delta collected relevant documents from the custodians and produced them to the DOJ. Id. Several months later, in May 2009, after a conversation with the DOJ regarding Delta’s e-mail system, Delta’s in-house counsel instructed its technology department to copy the hard drives of the custodians and place them on a separate server to prevent the automatic deletion of relevant e-mails. Id. In addition, Delta instructed a third-party vendor to suspend its overwriting of Delta’s back-up tapes, although the vendor did not implement this instruction until sometime after June 6, 2009. Id.

On May 22, 2009, the first case was filed in the multidistrict litigation. After conducting discovery, plaintiffs filed a motion seeking sanctions for spoliation of evidence. These private plaintiffs claimed that Delta’s improper three-month delay in suspending its standard electronic document destruction policy after receiving the DOJ’s CID resulted in the automatic deletion of relevant e-mails and documents dating back to July 2008. Id. As the Delta court observed, the plaintiffs’ spoliation motion “rest[ed] upon a critical assumption: that they (as private parties in civil litigation) [could] enforce the provisions of the February 2009 CID (a confidential investigation commenced by the DOJ) against Delta when there [was] no suggestion that the DOJ has taken action against Delta for failure to comply with the CID, and where the CID was issued more than three months before the first case in [the multidistrict litigation] was filed on May 22, 2009.” Id. at 1307.

The Delta court rejected this argument; specifically, it disagreed with the premise that the issuance of the CID gave rise to a general duty on the part of Delta to preserve evidence related to baggage fees. The court began by observing that plaintiffs failed to cite “any authority that would support such a sweeping and novel theory of spoliation.” Id. The Delta court went on to reject the suggestion that the duty to preserve documents “is a duty that does not attach to any party.” Id. Based on this theory, the Delta plaintiffs “posit[ed] that anyone (including them) could advance a spoliation argument against Delta for its alleged failure to comply with the DOJ’s CID.” Id. The Delta court disagreed, explaining that, “not only is such a suggestion unsupported by any case law, but it flies in the face of the legal definition of the word duty, which defines duty as a ‘legal obligation that is owed or due to another.’” Id. (quoting BLACK’S LAW DICTIONARY 580 (9th ed. 2009)). The court underscored the “sweeping and novel” nature of plaintiffs’ theory: “In essence, Plaintiffs ask this Court to hold that, as a matter of law, when a business is served with a CID, an irrebuttable presumption arises that civil litigation filed by one or more parties against the business receiving the CID is reasonably foreseeable. No court has so held, and this Court is unwilling to be the first.” Id. at 1308.

Other courts have reached similar conclusions. For example, the District Court for the Southern District of Florida concluded that neither a government investigation, nor a prior lawsuit, triggered the duty to preserve evidence for use in a later proceeding. Point Blank Solutions, Inc., 2011 WL 1456029, at *24. In Point Blank Solutions, plaintiff, a manufacturer of body armor, contracted with defendant, Toyobo America, Inc. (“Toyobo”) to supply Zylon, a fiber used in making bullet-proof vests. In June 2003, an officer was shot while wearing a Zylon vest and was injured by bullets that penetrated the fiber. Id., at *13. Following this incident, the DOJ launched an investigation and a number of suits were filed across the country involving the Zylon-containing vests, including a suit against Toyobo by manufacturer Second Chance.

Years later, in 2009, Point Blank sued Toyobo, and ultimately sought spoliation sanctions, arguing that Toyobo was under a duty to preserve evidence that began in 2003, with the first round of litigation and the government investigation. The Point Blank Solutions court disagreed. In rejecting Point Blank’s argument that the 2003 investigation and litigation created a duty to preserve evidence that “shifted” from the government and Second Chance to Point Blank when it filed suit, the court found the concept of a shifting duty to be “incompatible with the basic rule that a duty is owed to a specific party.” Id., at *24. As the court explained: “A lawsuit filed by Second Chance and a subpoena issued by the government do not necessarily mean that Toyobo should have at that time anticipated litigation with Point Blank or that Point Blank can take advantage of an evidence-preservation duty owed by Toyobo to Second Chance or different government agencies.” Id., at *26.

The Court agrees with these decisions and similar cases recognizing that the duty to preserve evidence is owed to a specific party and does not “shift” over time.[3] Under the contrary approach urged by Plaintiff States, “It is difficult . . . to imagine how a party could ever dispose of information under such a broad [shifting] duty because of the potential for some distantly related litigation that may arise years into the future.” Brigham Young Univ. v. Pfizer, Inc., 282 F.R.D. 566, 572 (D. Utah 2012). Such an approach is unreasonable, unfair, and unworkable.

For their part, the States primarily reference In re Actos (Pioglitazone) Prods. Liab. Litig., No. 6:11-MD-2299, 2014 WL 2872299, at *21-22 (W.D. La. June 23, 2014), to support their argument that “a duty to preserve can arise when a party anticipates litigation on similar facts or issues, even if the opposing party may be different.” (Dkt. #600 at 14). The States’ reliance on In re Actos is misplaced. Indeed, the In re Actos court expressly acknowledged the “shifting duty” concerns articulated in the Point Blank Solutions and Delta cases, among others, and did not disagree with their analysis. See In re Actos, 2014 WL 2872299, at *21-22 & n.72. To the contrary, the In re Actos court distinguished the facts at issue in that case from the “shifting duty” line of precedent. The court noted that, unlike Point Blank Solutions and similar cases, which “involve[d] a claimant seeking the benefit of a previously-issued hold not issued for his or her benefit, the In re Actos case involved a prior document-retention hold that, “by its own language, was prompted by the filing of a lawsuit alleging ‘personal injury or wrongful death . . . resulting from the use of . . . [the drug] Actos,’” and “Plaintiffs in this [multidistrict litigation] each allege personal injuries and/or wrongful death ‘resulting from the use of . . . Actos.’” Id., at *22 (emphasis in original). As the In re Actos court further explained, “[t]he scenario presented in Point Blank Solutions is simply not similar to the scenario presented in the instant case.” Id. “Here, a defendant manufacturer, in 2002, issued and ‘refreshed’ multiple times thereafter, by its own admission, a general products liability hold sweeping in its breadth, and applicable to a specific drug, Actos. The cases within this [multidistrict litigation] all grow out of and/or relate to products liability claims for wrongful death or personal injury allegedly resulting from the use of the same drug (Actos), thus, clearly distinguishing the instant matter from the [“shifting duty”] cases argued by [the Takeda defendants].” Id. (emphases in original) (cleaned up).

In sum, In re Actos involved an unusually broad products liability hold as to a specific drug, Actos, that encompassed subsequent litigation concerning injuries resulting from use of the same drug. The unique facts of the case, featuring an expansive litigation hold that was repeatedly refreshed, drove the result in In re Actos, not any disagreement with the analysis and conclusions in Point Blank Solutions, Delta, and other cases rejecting the “shifting duty” theory.

D. Application

As Google acknowledges, when Texas announced its antitrust investigation into Google’s ad tech products in a CID issued on September 9, 2019, Google could reasonably anticipate that this litigation would follow. Thus, on that date its document retention responsibilities began. The problem with picking an earlier date is that there is no reason to believe that Google “reasonably anticipated” a lawsuit of this nature, scope, and magnitude at any earlier moment in time.

As the Court has previously observed, this case is exceptional by any measure, both in its scope and the complexity of the issues presented. Broadly speaking, the case concerns Google’s “alleged monopolization and suppression of competition in online display advertising – essentially, the marketplace for the placement of digital display ads on websites and mobile apps.” In re Digit. Advert. Antitrust Litig., 555 F.Supp.3d 1372, 1373 (J.P.M.L. 2021). Like so many aspects of modern life, the internet brought seismic changes to the advertising industry, including the advent of “display advertising,” a form of tailored digital advertising, displayed on websites and mobile applications, that allows advertisers to direct ads to specific web users based on their browsing history and characteristics. Publishers and advertisers use sophisticated web tools to conduct auctions, each lasting a fraction of a second, to sell ads that are targeted to likely purchasers. In this market, and as described in the current complaint, advertisers use so-called “ad buying tools” to purchase publisher “inventory”—the space on websites where ads are shown. These ad buying tools let advertisers specify how much they are willing to pay for inventory and the audiences that they want their ads to reach. On the other side of the transaction, publishers use “ad servers” to manage and sell their web display inventory by specifying where advertisers can purchase that inventory and the minimum prices that publishers will accept for it, among other functions. Situated in the middle of these transactions are “ad exchanges” that serve to connect advertisers’ buying tools and publishers’ ad servers. The ad exchanges conduct real-time auctions in which advertisers—through ad buying tools—bid on publisher inventory.

Challenging Google’s alleged misconduct in the display advertising market, the Plaintiff States have asserted federal and state antitrust claims, as well as claims under the various States’ deceptive trade practices laws. The Plaintiff States’ Fourth Amended Complaint, which spans approximately 236 pages, includes the following alleged violations of federal law:

(1) Google has violated Section II of the Sherman Act, 15 U.S.C. § 2, by willfully acquiring or maintaining a monopoly in the market for ad servers, ad exchanges, and ad buying tools through anticompetitive conduct, (4th Am. Compl. ¶¶ 598–601);

(2) Google has monopoly power, or in the alternative, a dangerous probability of acquiring monopoly power in the market for ad exchanges and ad buying tools. According to the Plaintiff States, Google also has willfully, knowingly, and with specific intent, attempted to monopolize the market for ad exchanges and the market for ad buying tools, (4th Am. Compl. ¶¶ 602–06); and

(3) Google has violated Section I and II of the Sherman Act by tying its ad exchange and ad server together to coerce publishers to use both products, (4th Am. Compl. ¶¶ 607–13).[4] 

The Plaintiff States also maintain that Google’s conduct in the display advertising market was deceptive and violated the States’ deceptive trade practices laws. (4th Am. Compl. ¶¶ 526–97, 674–758). Specifically, among other claims, the Plaintiff States assert that Google deceived publishers and advertisers about its use of programs designed to benefit Google at the expense of publishers and advertisers, failed to properly disclose programs, and misrepresented that it runs a transparent marketplace where participants compete on equal footing. (4th Am. Compl. ¶¶ 526– 97).

On their face, the claims in this case implicate the application of federal antitrust law and various States’ antitrust laws, as well as various States’ deceptive trade practices laws, to a variety of alleged acts and practices of Google spanning nearly fifteen years in a complex digital advertising marketplace. For example, according to the States, a non-exhaustive list of apparently key milestones in Google’s alleged misconduct giving rise to this suit include the following:

• 2006—Google starts exploring an acquisition of DoubleClick.
• 2008—Google acquires DoubleClick and begins using it to unlawfully limit exchange competition.
• 2010—Beginning in 2010, Google restricts the ability of publishers using a non-Google ad server to trade through Google’s ad exchange (AdX), only allowing publishers that license Google’s ad server to receive live, competitive bids from AdX.
• 2014—Google launches “Dynamic Revenue Sharing,” which the States allege harmed competition in the ad-exchange market by manipulating floors after bids were received and after “peeking” at the bids of rival exchanges.
• 2015—Google implements “Reserve Price Optimization,” a program that the States maintain overrode and increased the bidding floors set by publishers and thereby deceptively increased the amount that advertisers paid for impressions on AdX.
• 2018—Google and Facebook enter into a “Network Bidding Agreement,” a key part of what the States allege was part of an “unlawful agreement” under which Facebook substantially curtailed its use of a practice known as “header bidding” in return for Google giving Facebook an advantage in publishers’ web display and developers’ in-app ad auctions.

These are just a handful of examples of various Google programs, agreements, and policies identified by the States, spanning well over a decade, that the States maintain were unlawful. Does it make sense for the Court to spin the wheel and select any of these earlier dates as the point in time where Google should have “reasonably anticipated” this massive, complex lawsuit, from these litigants, addressing the above-referenced and many other purportedly illegal actions undertaken by Google? The States provide no meaningful guidance on this point, but the answer is plain: Of course not. Picking any point on the timeline of this case prior to September 2019, as the date that Google could “reasonably anticipate” this lawsuit, would be, at best, arbitrary, and, at worst, nonsensical. The Court will not go down that road.

The States contend that the Court must wade into the mangrove swamp of picking an earlier duty-to-preserve date because Google has asserted work-product privilege for certain documents that predate Texas’s announcement of its investigation in September 2019. This argument appears to turn on the assumption that the commencement of the duty to preserve must coincide with the period of time for which Google has asserted any work-product protection. Not so. As Google correctly notes, documents reflecting attorney work product remain protected from disclosure in subsequent litigation matters even after the circumstances that originally gave rise to the protection no longer exist. See, e.g., In re Grand Jury Proceedings, 43 F.3d 966, 971 (5th Cir. 1994); Anderson Energy Grp. (Ohio), LLC v. Endeavor Ohio, LLC, No. 3:13-CV-1784, 2014 WL 12580471, at *2 (N.D. Tex. Sept. 4, 2014); In re Ethicon, Inc., Pelvic Repair Sys. Prod. Liab. Lit., 299 F.R.D. 502, 513 n.8 (S.D. W.Va. 2014). Google has confirmed that none of its work-product privilege claims are based on reasonably anticipated litigation in this case. (Dkt. #585). If the States have a substantive disagreement as to any such claims, they can request an in camera review process.

IV. CONCLUSION

For the foregoing reasons, Google’s Motion for Review of the Special Master’s July 15, 2024 Order Granting in Part Plaintiffs’ Motion to Compel Written Discovery Regarding Google Chats, (Dkt. #585), is GRANTED. The Order, (Dkt. #585), is MODIFIED to reflect that the Court has made no determination on spoliation at this time, preliminary or otherwise. The Order is further MODIFIED to require that Google must produce the start and end dates for any legal holds applicable to any custodians identified in this matter only from September 9, 2019 onwards.

So ORDERED and SIGNED this 17th day of January, 2025.