Loop, LLC v. CDK Glob. (In re Dealer Mgmt. Sys. Antitrust Litig.)

Full title: IN RE DEALER MANAGEMENT SYSTEMS ANTITRUST LITIGATION, v. CDK GLOBAL, LLC…

Court: United States District Court, Northern District of Illinois

Date published: Jul 22, 2024

18-cv-2521 (N.D. Ill. Jul. 22, 2024)

Opinion

18-cv-2521 MDL 2817

07-22-2024

IN RE DEALER MANAGEMENT SYSTEMS ANTITRUST LITIGATION, v. CDK GLOBAL, LLC, Defendants. LOOP, LLC, d/b/a AUTOLOOP Plaintiff,

Rebecca R. Pallmeyer Judge

MEMORANDUM OPINION AND ORDER

Rebecca R. Pallmeyer Judge

This antitrust multidistrict litigation revolves around the back-end software essential for managing operations at our nation's car dealerships. At the center of the case are Defendants CDK Global, LLC (“CDK”) and The Reynolds and Reynolds Company (“Reynolds”), leading providers of this automotive software, who stand accused of conspiring to violate antitrust laws by restricting competitor access to their data systems-which allegedly has resulted in increased prices for a subset of their services. Plaintiff Loop, LLC (“AutoLoop”) has sued CDK on behalf of hundreds of automotive software application vendors that are direct purchasers of the services at issue in this case. Having survived various challenges throughout this litigation, including a pleadings challenge, Daubert briefing, and a summary judgment motion, AutoLoop now seeks certification of a class of software vendors (the “Vendor Class”) under § 1 of the Sherman Act allegedly harmed by the conspiracy. In support of its bid for class treatment, AutoLoop has submitted the report of its expert, Dr. Mark Israel. Defendant CDK opposes class certification and has moved to exclude portions of Dr. Israel's report. For the reasons detailed below, the court overrules CDK's challenges and certifies the Vendor Class.

BACKGROUND

I. Pre-Suit Factual Background

The court assumes familiarity with the facts of this case, which have been set out in greater detail in several prior opinions.¹To recap: Defendants CDK and Reynolds² are technology companies that operate in the automotive software space and sell “dealer management systems” (“DMS”) to retail automobile dealerships. In simple terms, a DMS is a software platform that enables car dealers to use the data they generate to manage the day-to-day functions of their business, including sales, financing, and service operations. CDK and Reynolds control roughly 70% of the United States franchise dealership market for these systems. Within a DMS, dealers use software applications (“apps”) that supplement the DMS's functionality and offer add-on tools to deal with tasks such as inventory management, customer relationship management, warranty services, repair orders, and electronic vehicle registration and titling. Most of these apps are sold by third-party developers, referred to here as “vendors,” although DMS providers like CDK and Reynolds also create and market apps to dealers.

1 See, e.g., In re Dealer Mgmt. Sys. Antitrust Litig. (“Summ. J. Op.”), 680 F.Supp.3d 919, 933 (N.D. Ill. 2023) [1381]; In re Dealer Mgmt. Sys. Antitrust Litig. (“Daubert Op.”), 581 F.Supp.3d 1029 (N.D. Ill. 2022) [1321]; In re Dealer Mgmt. Sys. Antitrust Litig. (“Dealers MTD Op.”), 362 F.Supp.3d 510 (N.D. Ill. 2019) [507]; In re Dealer Mgmt. Sys. Antitrust Litig. (“AutoLoop MTD Op.”), 362 F.Supp.3d 477 (N.D. Ill. 2019) [504]; In re Dealer Mgmt. Sys. Antitrust Litig. (“Authenticom MTD Op.”), 313 F.Supp.3d 931 (N.D. Ill. 2018) [176].

2 AutoLoop brought its case against CDK only, with the understanding that CDK is jointly and severally liable for all harm caused by Reynolds in furtherance of the alleged conspiracy. See In re Uranium Antitrust Litig., 552 F.Supp. 518, 522 (N.D. Ill. 1982) (“[A]n antitrust defendant is jointly and severally liable for the acts of its co-conspirators.”).

To make their apps commercially viable, vendors require access to dealers' (their clients') data, which is stored on the dealers' DMS. However, this data is kept on the DMS in a raw and unprocessed form that is generally unusable by vendors. So, vendors enlist data integration companies to “clean” and consolidate the data, making it functional for the vendor's needs-a process generally known as data syndication and referred to here as data integration services (“DIS”). CDK and Reynolds both sell “certified” DIS (called 3PA and RCI, respectively) within their own DMS platforms, but vendors can also purchase DIS from third-party data integrators. Notably, CDK has two subsidiary companies, DMI and IntegraLink, that sell DIS on the open market.

This antitrust case centers around the evolution of CDK's and Reynolds's policies regarding third-party access to dealers' DMS data. Initially, both companies had relatively “open” policies that allowed independent DIS providers to directly access dealers' DMS data. This meant vendors working with a dealer that used either of Defendants' DMSs had the option to use Defendants' certified DIS services or contract with an independent DIS provider. Beginning in 2006, however, Reynolds began preventing independent DIS providers from accessing its DMS, whereas CDK kept its DMS “open” to other DIS providers, and promoted this feature to dealers as a reason dealers should use CDK's system. In other words, for a period of time CDK and Reynolds competed in part on the “openness” of their DMSs. Then, around September 2013, CDK suddenly stopped marketing its DMS as open. Plaintiffs claim this about-face was the result of an oral agreement between senior executives at CDK and Reynolds to squeeze independent DIS providers out of the market. According to Plaintiffs, they achieved this by, among other things, coordinating a message that “data security” concerns required them to block independent DIS providers from accessing their respective DMSs. Consistent with oral agreement, CDK would begin to restrict access to its DMS, while Reynolds orally agreed that for a limited time DMI and IntegraLink would retain access to Reynolds's DMS-what Plaintiffs characterize as a form of quid pro quo.

In February 2015, following these alleged oral agreements, CDK and Reynolds entered into three written agreements. In one of these agreements, called the “Wind Down Agreement,” they effectively memorialized the alleged quid pro quo arrangement: CDK formally agreed that within approximately five years, DMI and IntegraLink would stop servicing dealers who had a Reynolds DMS. In return, Reynolds agreed to give DMI and IntegraLink access to its DMS during the winddown period. Reynolds and CDK also agreed that neither would any other independent DIS provider that attempted to access the other's DMS. Concurrent with the Wind Down Agreement, CDK and Reynolds also entered into the “3PA Agreement” and the “RCI Agreement.” These agreements provided Defendants with reciprocal access to each other's certified integration programs.

Thus, AutoLoop's theory of the case is that CDK and Reynolds conspired to eliminate competition from independent DIS providers, thereby forcing vendors to purchase DIS from CDK and Reynolds at an inflated price.

II. Initial Pretrial Proceedings

Dozens of auto dealerships, vendors, and independent data integrators sued CDK and Reynolds under the Sherman Act and various state antitrust and consumer protection laws. On February 1, 2018, the Judicial Panel on Multidistrict Litigation (“JPML”) consolidated the cases in this MDL. The parties filed various claims and counterclaims under federal and state antitrust laws, many of which were resolved early in the litigation via dismissal³ or settlement.⁴ The court's May 2018 scheduling order [166] laid out a merits discovery period for all plaintiffs through February 2019 (later extended to April 2019), followed by an initial round of Daubert and summary judgment motion practice, and then by motions for class certification plus a second round of Daubert briefing specifically directed towards the requirements of Rule 23. After fact discovery closed, the court ruled on the parties' initial Daubert motions in January 2022 and on their motions for summary judgment in June 2023. In its 2022 Daubert opinion, the court found AutoLoop's expert's merits report admissible in its entirety. (See generally Daubert Op. [1321].) Then, in 2023 the court ruled that AutoLoop's Sherman Act § 1 conspiracy claim and its parallel state law claim survived summary judgment, and that those claims will be subject to the rule-of-reason analysis at trial. (See generally Summ. J. Op. [1381].)

3 (See [176], [425], [504], [505], [506], [507], [749], [857].)

4 (See [502], [778], [1199], [1351].)

III. Post-Discovery Factual Developments

Given the long procedural history of this case, the facts underlying AutoLoop's substantive claims continue to evolve in tandem with the litigation. For example, CDK now highlights for the court's consideration several changes that have taken place in the DMS and DIS markets that were not presented at summary judgment. For one, CDK alleges that it now faces increased competition from other DMS providers, such as the firms Auto/Mate, Dealertrack, and Tekion, challenging its' and Reynolds's historic dominance over the DMS market. In addition, CDK pushed out a few new initiatives that it claims were intended to win over customers in the face of increased competition. These include a “Customer Rewards Program” (introduced in 2018 and discontinued a year later) that offered various benefits in the form of discounts on upfront costs and ongoing fees, including waivers of the DIS fees charged to participating dealers' vendors for use of their 3PA data. (Decl. of Leigh Ann Conver [1460-11] (hereinafter “Conver Decl.”) ¶¶ 18-21.) CDK also announced in 2022 that it would “sunset 3PA [CDK's DIS] by the end of 2024” in favor of a new product, “Fortellis,” which provides a centralized “app store” for dealers to purchase CDK and third-party apps as well as a more modern framework for vendors to integrate with dealers' DMS data. (Id. ¶¶ 37-39.) CDK launched Fortellis in early 2019. (Id. ¶ 37.) Finally, CDK created a new free service in September 2022 called “Data Your Way,” which allows dealers to export their own DMS data and share it with vendors directly without paying any DIS fees to CDK. (Id. ¶¶ 48-52.)

IV. Class Certification

Before the court now is AutoLoop's motion for class certification. AutoLoop seeks certification of one damages class (the “Vendor Class”) under Federal Rule of Civil Procedure 23(b)(3), comprised of:

All automotive software vendors (i.e., persons or entities engaged in the sale of software solutions to automotive dealerships) located in the United States that, at any time since October 1, 2013, have purchased data integration services from CDK or Reynolds. Excluded from the class are (1) CDK, Reynolds, and any of their officers, directors, employees, subsidiaries, and affiliates; (2) Cox Automotive, Inc. and its subsidiaries and affiliates; and (3) automotive software vendors that first purchased data integration services from CDK after June 5, 2018.

(AutoLoop Class Certification Reply [1472] at 6.)⁵The Vendors bring a single Sherman Act § 1 conspiracy claim against CDK.⁶ AutoLoop requests that it be appointed as the class representative for the Vendor Class and that Kellogg Hansen and Professor Samuel Issacharoff be appointed class counsel.

5 AutoLoop modified its proposed class definition in its reply brief to exclude from the class any vendors that “first purchased data integration services from CDK after June 5, 2018.” This was done in response to CDK's argument that vendors who first contracted with CDK after June 5, 2018 signed agreements that included class-waiver provisions. Courts ordinarily allow plaintiffs to modify their proposed class definition in response to an argument made by the defendant. See Simpson v. Dart, No. 18 C 553, 2021 WL 2254969, at *1 (N.D. Ill. June 3, 2021) (allowing plaintiff to modify class definition in reply brief); see also Schorsch v. Hewlett-Packard Co., 417 F.3d 748, 750 (7th Cir. 2005) (noting that “litigants and judges regularly modify class definitions”). The court will do so in this case and addresses AutoLoop's recent proposed class definition, narrowed in light of the class action waiver.

6 As noted, AutoLoop's Florida-law claim also survived summary judgment; however, AutoLoop does not assert this claim on a classwide basis. (See AutoLoop Mem. in Supp. of Mot. for Class Certification [1422-1] at 2 n.2.)

The following discussion is divided into two sections. In the first section, the court addresses CDK's challenges to certain opinions of AutoLoop's expert, Dr. Mark Israel. Then, the court rules on AutoLoop's class certification motion.

DAUBERT MOTION

Federal Rule of Evidence 702 governs the admissibility of expert witness testimony. See Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579, 588 (1993). Before admitting expert testimony, a district court must determine that (1) the expert is proposing to testify to valid scientific, technical, or other specialized knowledge, and (2) the testimony will assist the trier of fact. See Robinson v. Davol Inc., 913 F.3d 690, 695 (7th Cir. 2019). When an expert's testimony is “critical” to class certification, a district court must resolve any challenge to that expert's qualifications or submissions before it rules on class certification. See Messner v. Northshore Univ. HealthSystem, 669 F.3d 802, 812 (7th Cir. 2012) (citing Am. Honda Motor Co. v. Allen, 600 F.3d 813, 815-16 (7th Cir. 2010)). In the context of this motion, expert testimony is critical if it is “important to an issue decisive for . . . a class certification decision.” Id. at 812. The party seeking to offer an expert's testimony bears the burden of establishing its admissibility under Daubert by a preponderance of the evidence. City of Rockford v. Mallinckrodt ARD, Inc., No. 3:17-CV-50107, 2024 WL 1363544, at *5 (N.D. Ill. Mar. 29, 2024) (citing Gopalratnam v. Hewlett-Packard Co., 877 F.3d 771, 782 (7th Cir. 2017)).

As alluded to above, the court has ruled on numerous challenges CDK has made against Dr. Israel's expert reports. (See generally Daubert Op. and Summ. J. Op.) These prior rulings are law of the case and will not be reconsidered here “absent exceptional circumstances such as a change in the law, new evidence, or compelling circumstances.” Zhang v. Gonzales, 434 F.3d 993, 998 (7th Cir. 2006). For this reason, the current review of Dr. Israel's reports can be more narrowly focused on whether he offers a reliable method to try the Vendors' case on a classwide basis. The court's analysis below homes in on this narrower question, and declines CDK's invitation to revisit past issues.

I. Dr. Mark Israel

Dr. Israel has submitted several expert reports in this case. In a nutshell, the central theory of his reports on the merits of the Vendors' claims is that

CDK's and Reynolds' actions have harmed competitors, competition, and consumers in DIS markets and App markets and have harmed competition and consumers in the DMS market.... [T]he mechanism of this harm was the joint blocking of independent DIS providers from accessing the CDK and Reynolds DMSs, which effectively monopolized CDK's and Reynolds' respective DIS markets, thus leading to higher DIS prices and higher DIS costs to App vendors, thereby also harming competition in App markets.

(Israel Initial Rep. [889-1] ¶ 10; see also Israel Initial Reply Rep. [889-2] ¶ 5.) To quantify how much the conspiracy increased DIS prices, he used a “difference-in-differences” (“DID”) model, which compared changes in prices for CDK's and Reynolds's integration services (3PA and RCI, respectively) against benchmark prices from an independent data integrator (in this case, Authenticom⁷) that were not altered by the alleged conspiracy. (See Israel Initial Rep. ¶¶ 193-99.) Dr. Israel checked the robustness of his DID results with a separate regression model, called a “before-during” (“B&D”) model, which compared 3PA and RCI prices before and during the alleged conspiracy period. (Id. ¶¶ 203-04.) Both models supported Dr. Israel's economic conclusion that CDK and Reynolds conspired to eliminate competition in their DIS markets, and that vendors incurred price overcharges as a result. (See id. ¶¶ 200-02, 204.) CDK vigorously challenged the admissibility of Dr. Israel's merits opinions on multiple grounds, but the court rejected those arguments in two separate opinions, ruling generally that the disputes amounted to a battle of the experts that must be left for the factfinder to resolve. (See Daubert Op. at 16-29; Summ. J. Op. at 6872.)

7 Authenticom is an independent data integrator. It was previously a party to this MDL but settled its claims against Defendant CDK in October 2020 [1199] and against Defendant Reynolds in May 2022 [1351].

Dr. Israel has submitted additional reports in support of AutoLoop's class certification motion, although these are more properly understood as addenda to his initial merits reports. (See Israel Suppl. Rep. [1422-2], Israel Suppl. Reply [1474-1], Israel Suppl. Sur-Reply [1513-1].) In these newer reports, Dr. Israel asserts that his initial merits analysis “was entirely done on a common basis, evaluating the alleged conduct as a whole and its impact on a market-wide basis, without the need to evaluate issues particular to individual class members” and thus “appl[ies] equally to all class members, and do[es] not require inquiry into issues particular to any class member.” (Israel Suppl. Rep. ¶ 13.) In other words, Dr. Israel opines, “there are common, class-wide economic methods to evaluate” market definition, market power, liability, and antitrust injury. (Id. ¶ 45; see id. ¶¶ 20, 24, 28, 34.) Further, he states that the models he used in his initial report provide a “common method to quantify damages” for the class as a whole and for each member of the class. (Id. ¶ 45; see Id. ¶¶ 35, 38.) In this most recent round of reports, Dr. Israel reruns his regression models with updated data through September 2023. (Israel Suppl. Reply ¶¶ 43-44, 96-97.) With this updated data, his DID model estimates $395.46 million in classwide damages and his B&D model estimates $406.94 million in classwide damages. (Id. at 49 tbl.1, 56 tbl.2.)

At class certification, CDK moves to exclude testimony of Dr. Israel that relates to “(1) any opinion that the fact of ‘monopolization' establishes class-wide antitrust impact; (2) all damages Dr. Israel computes using his before-during model; (3) all post-2019 damages computed using Dr. Israel's difference-in-differences model; and (4) any opinion that damages are continuing.” (CDK's Mot. to Exclude Ops. of Dr. Israel [1459] at 1.) CDK's motion rests in key part on the report of its own expert, Dr. Laila Haider, an economist and vice president at consulting firm Charles River Associates, whose opinions the court has not previously addressed in this case [1460-1]. The court deals with each of CDK's arguments in turn, ultimately rejecting CDK's motion to exclude any portion of Dr. Israel's class certification opinions.

A. Monopolization and Antitrust Injury

CDK first argues that the court should exclude “new” opinion testimony from Dr. Israel that, in CDK's view, conflates unreasonable restraint (i.e. closure of CDK's and Reynolds's DMS markets to independent data integrators) with antitrust injury (also referred to by some courts as “antitrust impact”). (CDK's Mem. in Supp. of Mot. to Exclude Ops. of Dr. Israel [1459-1] at 1.) Specifically, CDK notes that Dr. Israel testified in his deposition that the “fact of monopolization establishes injury.” (Jan. 19, 2014 Israel Dep. [1460-12] 73:10-11.) This means, CDK contends, that “according to Israel, the DMS closure automatically establishes injury to all vendors, regardless of any effect on price.” (CDK's Mem. in Supp. of Mot. to Exclude Ops. of Dr. Israel at 6.) CDK calls this view “legally flawed” because “proof of [an antitrust] violation and of antitrust injury are distinct matters that must be shown independently.” (Id. at 5-6 (emphasis in CDK's brief) (first quoting Meds. Co. v. Mylan Inc., No. 11-CV-1285, 2014 WL 1227214, at *5 (N.D. Ill. Mar. 25, 2014), and then quoting Atl. Richfield Co. v. USA Petrol. Co., 495 U.S. 328, 343-44 (1990).)

The court reads Dr. Israel's testimony less broadly; it does not appear to the court that he has assumed that antitrust injury and liability necessarily rise and fall together. Instead, as he makes clear in his report, “[his] opinion is not that all monopolization cases in which [d]efendants are found liable necessarily have common antitrust injury; it is that the facts of this case demonstrate such common injury.” (Israel Suppl. Reply ¶ 40.) He explains that in other monopolization cases there may be debate as to whether buyers in the relevant market retain sufficient alternative options to avoid harm, “[b]ut based on the evidence in the present case, I see no room for debate on this question: The options for independent automated DIS services were eliminated, leaving a monopoly in each DIS market” that resulted in an increase in the price of CDK's and Reynolds's DIS, among other harms. (Id. ¶¶ 37-38; see also Israel Suppl. Rep. ¶ 33.) The quote from Dr. Israel's deposition that CDK points to is taken out of context; a more honest reading of his testimony shows that Dr. Israel was being asked a question about the evidence in this case, not being asked to opine on the general legal theory of antitrust injury. (See Jan. 19, 2014 Israel Dep. 72:18-74:11.)

Beyond this, CDK overstates the legal proposition that “proof of [an antitrust] violation and of antitrust injury are distinct matters that must be shown independently.” (CDK's Mem. in Supp. of Mot. to Exclude Ops. of Dr. Israel at 6 (emphasis in CDK's brief).) As an initial matter, each of the cases it cites for this proposition dealt with whether a competitor-as opposed to a customer, like each of the vendors is here-suffered an antitrust injury. See Atl. Richfield Co., 495 U.S. at 345 (holding competitor did not suffer antitrust injury); O.K. Sand & Gravel, Inc. v. Martin Marietta Techs., Inc., 36 F.3d 565, 573 (7th Cir. 1994) (same); Ind. Grocery, Inc. v. Super Valu Stores, Inc., 864 F.2d 1409, 1419 (7th Cir. 1989) (discussing antitrust injury in the context of harm to a competitor). The Supreme Court has explained that the distinction between consumer and competitor antitrust claims is relevant because “a competitor will be injured and hence motivated to sue only when [an antitrust violation] has a procompetitive impact on the market. Therefore, providing the competitor a cause of action would not protect the rights of . . . consumers under the antitrust laws.” Atl. Richfield Co., 495 U.S. at 34546 (emphasis in original). More to the point, none of these cited cases stand for the legal proposition that customers in a monopolized market cannot make out a prima facie case of antitrust injury based on evidence that a monopolist used its market power to increase prices. See Bradburn Parent/Tchr. Stores, Inc. v. 3M, No. CIV.A.02-7676, 2004 WL 1842987, at *13 (E.D. Pa. Aug. 18, 2004) (“[W]hen a monopolist unlawfully maintains its monopoly power . . . it is logical, at least as a general rule, to presume that all class members have suffered injury as a result of the conduct, in the form of supra-competitive prices.”). CDK's challenge has even less purchase here because Dr. Israel takes the next step and opines that his regression analysis provides a sound econometric method to discern the existence of antitrust injury on each class member in the form of supracompetitive DIS prices. Accordingly, to the extent that Dr. Israel is testifying that the alleged monopolization of CDK's and Reynolds's case resulted in antitrust injury to the vendors based on the evidence in this case, the court finds no basis to exclude his testimony.

B. Damages Related to Dr. Israel's B&D Model

Next, CDK argues that Dr. Israel's B&D model (which is his “secondary model” that he uses as a robustness check on his primary DID model) is unreliable because it (1) relies on pooled data- i.e., uses aggregated data for all vendors, rather than running the regression separately for each vendor-and (2) extrapolates from an allegedly unrepresentative sample. (See CDK's Mem. in Supp. of Mot. to Exclude Ops. of Dr. Israel at 7-10.) CDK concedes that “it is not unusual for economists to use pooled data to estimate aggregate damages in antitrust cases” but claims that it is necessary to first test the assumption that each customer was affected in the same way by the alleged conspiracy, specifically by using a “Chow test.”⁸ (Id. at 8.) CDK claims that when Dr. Haider performed a Chow test on Dr. Israel's B&D model, she found that the results “strongly reject[] the restrictive assumption that the price effects of the alleged conduct are the same for all vendors buying from a given Defendant in an alleged conduct period,” suggesting “that it is inappropriate to pool data for all vendors into a single regression model.” (Haider Rep. [1460-1] ¶¶ 241, 187.) In arguing that the secondary model uses an unrepresentative sample, CDK reasons that the vendors who purchased both before and during the alleged conspiracy (i.e., the subset of vendors that were used in the B&D model) are unrepresentative of the vendors as a whole because they “tended to buy more DIS, and more expensive DIS.” (CDK's Mem. in Supp. of Mot. to Exclude Ops. of Dr. Israel at 10 (emphasis in original).) Additionally, CDK takes issue with the model's inclusion of data from affiliates of Cox Automotive, Inc. as those vendors have settled their claims against CDK and are not part of the putative class. (Id.) AutoLoop and Dr. Israel respond to these challenges to the B&D model in their opposition papers and supplemental reply report. First, Dr. Israel defends his decision to rely on pooled data, contending that it was reasonable given the evidence of the alleged conspiracy's market-wide effects. He adds that pooling is particularly useful here because it generates results with greater “statistical significance,”⁹ whereas Dr. Haider's method of running separate regressions for each vendor produced many results that were not statistically significant. (Israel Suppl. Reply ¶¶ 112, 123, 127.) As for Dr. Haider's critique that Dr. Israel should have run a Chow test on his B&D model, AutoLoop denies that running a Chow test is a prerequisite to using pooled data and cites to several courts that have rejected similar arguments. See Olean Wholesale Grocery Coop., Inc. v. Bumble Bee Foods LLC, 31 F. 4th 651, 675-76 (9th Cir.) (en banc) (ruling that district court did not abuse discretion in concluding that “the failure of the Chow test did not require the court to reject the model” where “there was a rational basis for [expert's] use of the pooled regression model to demonstrate class-wide impact”), cert. denied sub nom. StarKist Co. v. Olean Wholesale Grocery Coop., Inc., 143 S.Ct. 424 (2022); In re Payment Card Interchange Fee & Merch. Disc. Antitrust Litig., No. 05-MD-1720 (MKB), 2022 WL 14862098, at *13 (E.D.N.Y. Oct. 8, 2022) (“[T]he case law does not support the contention that passing the Chow test is an absolute prerequisite.” (citation omitted)).

8 A Chow test is a testing principle that uses a mathematical formula to determine whether the independent variables have drastically different effects on two or more subsets of the data. (See Haider Rep. [1460-1] at 91 nn.340 & 341.)

9 Statistical significance is a measure used in research to determine if the results of a study are likely to be meaningful rather than occurring by chance. It helps researchers assess whether an observed effect or relationship between variables is likely to reflect a real pattern or if it could have happened randomly. See Matrixx Initiatives, Inc. v. Siracusano, 563 U.S. 27, 39 n.6 (2011) (citing Federal Judicial Center, Reference Manual on Scientific Evidence at 122-24, 354 (2d ed. 2000).)

Nonetheless, in response to Dr. Haider's critique, Dr. Israel concedes that he should have applied the Chow test to his B&D models that use producer price indices for the software industry (“software PPI”) or software PPI and auto sales as controls. Dr. Israel further admits that when he does apply the Chow test to these models, it rejects his assumption that the coefficients on the control variables are consistent across vendors. (Israel Suppl. Reply ¶ 110.) He contends, however, that the “appropriate reaction” to this result is “not to throw the model out, but rather to use a version that lets those coefficients vary by vendor ....” (Id.) In his reply report, Dr. Israel makes this adjustment and shows that-even when he allows the coefficients to vary by vendor-the estimated overcharges are similar to those in his original B&D model, and still higher than what his DID model estimates. (Id. at 55 tbl.2.) In other words, whether or not he accounts for the implications of the Chow test, his B&D model shows higher total overcharges than his DID model.

AutoLoop also dismisses CDK's argument that Dr. Israel's secondary model uses a nonrepresentative sample. As AutoLoop notes, the court has already addressed this issue, and has accepted Dr. Israel's conclusion that the alleged conspiracy affected all vendors in the same way. And, in any event, the group of vendors used in the B&D model represents a substantial portion of the putative class's transactions: roughly 70% of the Defendants' DIS sales volumes. (AutoLoop Opp'n to CDK's Mot. to Exclude Ops. of Dr. Israel [1470] at 11-12.) Finally, AutoLoop states that it would make “no economic sense” for Dr. Israel to remove Cox Automotive from the regression model just because it settled its case, since its transactions were still affected by the conspiracy. (Id. at 13.)

The court need not referee the experts' battle on this issue for two reasons. First, CDK has effectively waived any challenge to Dr. Israel's secondary model. Dr. Israel produced the same basic model in 2019 during the merits phase of this litigation; CDK had an opportunity to challenge the B&D model then and chose not to. The court considered the reliability of Dr. Israel's DID and B&D models in two separate opinions and ruled that both were admissible. (See Daubert Op. at 25 (“Israel also considered other methods of testing collusion's impact, including one that substituted CDK's and Reynolds' own pre-conduct prices as the relevant baseline and one that used industry control variables.”); Summ. J. Op. at 68 (“Israel conducted robustness checks using pre-conspiracy 3PA and RCI prices as an alternative benchmark and reached a similar calculation under that approach.”).) CDK has not shown any “exceptional circumstances such as a change in the law, new evidence, or compelling circumstances” that compel this court to reexamine its previous determination. Zhang, 434 F.3d at 998. Second, and more to the point, CDK has not clearly articulated how its challenges to Dr. Israel's secondary model bear on class certification, which is the present concern of this court. The only case CDK cites on the purported Chow test issue, Reed Construction Data Inc. v. McGraw-Hill Cos., Inc., was decided at summary judgment, not under Rule 23 (and, notably, involved an expert who could not provide any rational basis for his decision beyond his own common-sense judgment). 49 F.Supp.3d 385, 405-06 (S.D.N.Y. 2014), affd, 638 Fed.Appx. 43 (2d Cir. 2016). As to Dr. Israel's alleged use of an unrepresentative sample, the relevance of any potential variation between the sampled and unsampled vendors is itself a common question that may, if needed, be resolved at trial via classwide evidence. (See Israel Suppl. Reply ¶¶ 125, 131, 135 (noting that distinctions such as simple versus complex DIS services can be adjusted for using a common, classwide methodology).) Accordingly, the court rejects CDK's request to revisit the reliability of Dr. Israel's B&D model at this stage.

C. Post-2019 Damages Related to Dr. Israel's DID Model

CDK's third challenge relates to Dr. Israel's updated damages calculations from his DID model-but this deserves little discussion. CDK argues that Dr. Israel improperly “projects” damages through 2023 by applying the overcharge figure he calculated in 2019 during the merits phase to the intervening five years of DIS transactions. (CDK's Mem. in Supp. of Mot. to Exclude Ops. of Dr. Israel at 11.) CDK contends that this type of projection is inappropriate because “there is no basis for Israel's assumption of a constant overcharge from February 2015 to the present day.” (Id. at 12.) Dr. Israel explained in his opening report that he could not update his DID model at the time his opening brief was due because he did not yet have updated Authenticom data.¹⁰He has since obtained updated Authenticom data and re-run his DID model with those figures in his most recent report. (See Israel Suppl. Reply ¶¶ 94-97.) Thus, this issue is moot.

10 Recall that Dr. Israel's DID model calculates damages by measuring how much 3PA and RCI prices increased as compared to the benchmark prices of Authenticom.

D. Continuing Damages

Finally, CDK asks the court to exclude any opinion that damages from the alleged conspiracy are continuing because Dr. Israel “reaches that conclusion by ignoring four years of market facts . . . [that] undermine[] any link to the alleged conduct.” (CDK's Mem. in Supp. of Mot. to Exclude Ops. of Dr. Israel at 13.) CDK lists several new programs, products, and events that it claims Dr. Israel did not consider based on his deposition testimony, such as the increased market share of new DMS providers like Tekion and the introduction of allegedly procompetitive programs like “Data Your Way.” (Id. at 14.) AutoLoop counters by first noting that CDK's challenge here does not get at whether post-2019 damages can be adjudicated on a classwide basis, but instead raises a common defense that CDK can levy against the vendors at trial. Nevertheless, AutoLoop also argues that “the economic evidence” supports Dr. Israel's conclusion that damages are ongoing because “the competition that was eliminated by the conspiracy has not been restored” and CDK and Reynolds continue to charge higher prices than they would have but for the conspiracy. (AutoLoop Opp'n to CDK's Mot. to Exclude Ops. of Dr. Israel at 14-15.) It also argues that Dr. Israel's overcharge analysis already accounts for several of the “new” market facts that CDK claims he overlooks, such as CDK's fee waiver program. (Id. at 15.)

The court agrees with AutoLoop that this challenge to Dr. Israel's testimony has little to no bearing on class certification. CDK's cherry-picking of Dr. Israel's deposition testimony does not in any way establish that Dr. Israel's updated damages calculations ignore post-2019 developments wholesale in a way that renders them fundamentally unreliable; indeed, Dr. Israel does appear to have accounted for many of the developments in this latter period, if not every single one that CDK identifies. In any event, the experts-each relying on economic principles and data from this case- have offered competing theories for when the appropriate cutoff date is for damages calculations. The difference in opinion here is not a basis to exclude Dr. Israel's testimony; rather, it represents one of the many battles over which the experts will confront each other during trial. See In re Broiler Chicken Antitrust Litig., No. 16 C 8637, 2023 WL 7220170, at *5 (N.D. Ill. Nov. 2, 2023) (“Just because there is an alternative explanation that is contrary to that proffered by [plaintiffs' expert] does not mean that [plaintiffs' expert's] opinion should be excluded.”).

In conclusion, the court denies CDK's Daubert challenges and finds that the entirety of Dr. Israel's class certification reports are admissible.

CLASS CERTIFICATION MOTION

The court now reaches AutoLoop's class certification motion. To certify a damages class, a plaintiff must satisfy the four requirements of Rule 23(a)-numerosity, commonality, typicality, and adequacy-plus Rule 23(b)(3)'s predominance and superiority requirements. See Kleen Prods. LLC v. Int'l Paper Co., 831 F.3d 919, 923 (7th Cir. 2016); Harper v. Sheriff of Cook Cnty., 581 F.3d 511, 513 (7th Cir. 2009). Rule 23 “does not set forth a mere pleading standard,” and the plaintiffs bear the burden of “satisfy[ing] through evidentiary proof” each of its elements. Comcast Corp. v. Behrend, 569 U.S. 27, 33 (2013) (internal quotation marks and citation omitted). Accordingly, the court is tasked with conducting a “rigorous analysis” to determine whether the plaintiffs have satisfied the Rule's requirements. Id. (citing Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 351 (2011); see also Am. Honda Motor Co., 600 F.3d at 815 (“[B]efore deciding whether a class should be certified, [the court must] . . . make whatever factual and legal inquiries are necessary to ensure that requirements for class certification are satisfied.”). However, the Seventh Circuit has warned against elevating class certification proceedings “into a dress rehearsal for the trial on the merits,” and has instructed district courts to simply consider the parties' evidence and determine whether the plaintiffs have proven each of Rule 23's elements by a preponderance of the evidence. Messner, 669 F.3d at 811 (citations omitted).

Aside from the textual requirements provided for in Rule 23, the Seventh Circuit also requires plaintiffs to prove that the proposed class is ascertainable. See Mullins v. Direct Digit., LLC, 795 F.3d 654, 659 (7th Cir. 2015). The “ascertainability” standard is satisfied if the class definitions are precise, defined by objective criteria, and not dependent on success on the merits. Id. at 659-60. The court's analysis here addresses each of Rule 23's requirements for class certification, though CDK only genuinely challenges class certification on predominance grounds.

A. Rule 23(a) Requirements

1. Numerosity and Ascertainability

Rule 23(a)(1) requires that a class be “so numerous that joinder of all members is impracticable.” FED. R. CIV. P. 23(a)(1). CDK does not contest that Plaintiff can satisfy this requirement. Indeed, but courts have found “a forty-member class . . . sufficient to meet the numerosity requirement.” Anderson v. Weinert Enters., Inc., 986 F.3d 773, 777 (7th Cir. 2021) (citing Orr v. Shicker, 953 F.3d 490, 498 (7th Cir. 2020)). The putative Vendor Class is made up of 244 class members. (See Israel Suppl. Reply ¶ 97 & n.156.) Joinder of these hundreds of vendors scattered across the country would be impractical and, as AutoLoop has identified each of the 244 class members based on CDK's and Reynolds's own objective data, its proposed class definition also easily satisfies the Seventh Circuit's ascertainability requirement. (See Israel Suppl. Reply at 88-92 tbl.9.)

2. Commonality

Rule 23(a)(2) requires that “there are questions of law or fact common to the class.” FED. R. CIV. P. 23(a)(2). For purposes of the commonality requirement, “even a single common question will do.” Moehrl v. Nat'l Ass'n of Realtors, No. 19 C 1610, 2023 WL 2683199, at *11 (N.D. Ill. Mar. 29, 2023) (quoting Wal-Mart, 564 U.S. at 359). For this reason, “[a] common nucleus of operative fact is usually enough to satisfy the commonality requirement of Rule 23(a)(2) and is typically manifest where the defendants have engaged in standardized conduct towards members of the proposed class.” Am. Council of the Blind of Metro. Chi. v. City of Chicago, 589 F.Supp.3d 904, 908 (N.D. Ill. 2022) (cleaned up) (citing Keele v. Wexler, 149 F.3d 589, 594 (7th Cir. 1998)). Moreover, courts within the Seventh Circuit have repeatedly held that “the question of the existence of a conspiracy in restraint of trade is one that is common to all potential plaintiffs.” In re Ready-Mixed Concrete Antitrust Litig., 261 F.R.D. 154, 167 (S.D. Ind. 2009) (quoting Sebo v. Rubenstein, 188 F.R.D. 310, 313 (N.D. Ill. 1999)). In light of this caselaw, AutoLoop argues that whether CDK and Reynolds conspired to coordinate their data access policies and block independent data integrators from their respective DMSs is a common question (among others) that warrants satisfaction of the commonality requirement. CDK does not contest that AutoLoop has satisfied the commonality requirement on this basis, and the court agrees that it is met here.

3. Typicality

Rule 23(a)(3) requires that “the claims or defenses of the representative parties are typical of the claims or defenses of the class.” FED. R. CIV. P. 23(a)(3). The inquiry here is whether AutoLoop's claims “arise from the same events or course of conduct that gives rise to the putative class members' claims.” Beaton v. SpeedyPC Software, 907 F.3d 1018, 1026 (7th Cir. 2018). This requirement “is meant to ensure that the named representative's claims have the same essential characteristics as the claims of the class at large,” under the logic that “a class representative will adequately pursue her own claims, and if those claims are typical of those of the rest of the class, then her pursuit of her own interest will necessarily benefit the class as well.” Howard v. Cook Cnty. Sheriffs Off., 989 F.3d 587, 605 (7th Cir. 2021) (internal quotation marks and citations omitted). “The issue of typicality is closely related to commonality and should be liberally construed.” In re Steel Antitrust Litig., No. 08 C 5214, 2015 WL 5304629, at *3 (N.D. Ill. Sept. 9, 2015) (quoting Saltzman v. Pella Corp., 257 F.R.D. 471, 475 (N.D. Ill. 2009), aff'd, 606 F.3d 391 (7th Cir. 2010)).

CDK does not challenge the typicality requirement, and the court finds is satisfied here. AutoLoop's claims arise from Defendants' alleged conspiracy to eliminate competition in the data integration market; such claims can be pursued by every member of the putative class, each of whom also purchased DIS from Defendants. AutoLoop also notes that it prevailed at summary judgment on CDK's counterclaims [1383], meaning that any potential trial would only involve its affirmative antitrust claims against CDK, which are typical of the class.

4. Adequacy of Representation

Finally, the fourth of 23(a)'s class certification prerequisites requires that “the representative parties will fairly and adequately protect the interests of the class.” FED. R. CIV. P. 23(a)(4). To be an adequate representative, “the named plaintiff must ‘be part of the class and possess the same interest and suffer the same injury as the class members.' ” Conrad v. Boiron, Inc., 869 F.3d 536, 539 (7th Cir. 2017) (quoting Amchem Prods., Inc. v. Windsor, 521 U.S. 591, 625-26 (1997)). In a similar vein, “the court must be satisfied that the plaintiff will keep the interests of the entire class at the forefront.” Id. Alternatively, a named plaintiff will not be able to adequately represent a class “if, for example, the proposed representative is subject to a defense to which other class members are not, or if the representative cannot prove the elements of the class's claim for reasons unique to the representative.” Ploss v. Kraft Foods Grp., Inc., 431 F.Supp.3d 1003, 1011 (N.D. Ill. 2020) (citing CE Design Ltd. v. King Architectural Metals, Inc., 637 F.3d 721, 724-25 (7th Cir. 2011)).

It is undisputed that AutoLoop and its counsel have adequately represented the class.¹¹AutoLoop's antitrust claims mirror those of the class, suggesting that its interest in “maximizing class- wide damages” is aligned with the class's general interest. Steel, 2015 WL 5304629, at *4. Additionally, with roughly $15 million in purported damages stemming from the conspiracy, AutoLoop holds a substantial stake in the successful outcome of the case. (See Israel Suppl. Reply at 88 tbl.9.) Furthermore, it has demonstrated its interest in the action by intensely litigating its antitrust claims through the discovery, Daubert, and summary judgment stages, all while its counsel have served as co-lead counsel in this MDL since 2018. In other words, adequacy is met here because AutoLoop's claims align with those of the class, it has a significant interest in the outcome of the case, and its lawyers have proven to be competent.

11 CDK did raise an adequacy challenge in its opposition brief to class certification; however, its challenge was narrowly related to the fact that the AutoLoop's initially proposed class definition included some class members that had signed enforceable class waivers with CDK. (CDK Opp'n to Class Certification [1456] at 12-16.) As noted, AutoLoop mooted this challenge by voluntarily narrowing the scope of its class definition to effectively exclude vendors that could have entered into enforceable class waiver agreements with CDK. (See AutoLoop Class Certification Reply at 6.)

The court concludes that AutoLoop has met Rule 23(a)'s preliminary class certification requirements, as well as the Seventh Circuit's ascertainability requirement. Accordingly, the court now turns to the heart of the parties' dispute-Rule 23(b)(3).

B. Rule 23(b)(3) Requirements

To certify a class under Rule 23(b)(3), AutoLoop must demonstrate that “questions of law or fact common to class members predominate over any questions affecting only individual members, and that a class action is superior to other available methods for fairly and efficiently adjudicating the controversy.” In re Allstate Corp. Sec. Litig., 966 F.3d 595, 603 (7th Cir. 2020) (quoting FED. R. CIV. P. 23(b)(3)). Courts have isolated two distinct but related requirements embedded in Rule 23(b)(3)'s text: the “predominance” and “superiority” requirements. See id. The court deals with each of these in turn.

1. Predominance

The predominance inquiry tasks district courts with scrutinizing the balance between common and individual questions in the case. Tyson Foods, Inc. v. Bouaphakeo, 577 U.S. 442, 453 (2016). Individual questions are those where “members of a proposed class will need to present evidence that varies from member to member,” whereas common questions are those where “the same evidence will suffice for each member to make a prima facie showing [or] the issue is susceptible to generalized, class-wide proof.” Id. (alteration in Tyson Foods, Inc.) (quoting 2 William B. Rubenstein, Newberg on Class Actions § 4:50 (5th ed. 2012)). When evaluating whether common issues predominate, “a qualitative assessment” is necessary. Steel, 2015 WL 5304629, at *5 (citing Parko v. Shell Oil Co., 739 F.3d 1083, 1086 (7th Cir. 2014)). This approach goes beyond merely tallying the number of common versus individual issues, and instead centers on the relative significance of the issues. Id. For this reason, “[i]ndividual questions need not be absent” to certify a class and “[t]he text of Rule 23(b)(3) itself contemplates that such individual questions will be present.” Messner, 669 F.3d at 815. Further, “the action may be considered proper under Rule 23(b)(3) even though other important matters will have to be tried separately, such as damages or some affirmative defenses peculiar to some individual class members.” Tyson Foods, Inc., 577 U.S. at 454 (quoting 7AA Charles Alan Wright, Arthur R. Miller & Mary Kay Kane, Federal Practice & Procedure § 1778 (3d ed. 2011)). Thus, above all, the question the court must answer is whether individual issues will “overwhelm the questions common to the class.” Amgen Inc. v. Conn. Ret. Plans & Tr. Funds, 568 U.S. 455, 468 (2013). While the Supreme Court has underscored that the predominance inquiry is “demanding,” it has also noted that “[p]redominance is a test readily met in certain cases alleging . . . violations of the antitrust laws.” Amchem, 521 U.S. at 623, 625; see also Messner, 669 F.3d at 814 (noting that antitrust cases “will frequently lead to [class] certification” even under Rule 23(b)(3)'s “rigorous[] appli[cation]”).

The court begins its predominance analysis “with the elements of the underlying cause of action.” Messner, 669 F.3d at 815 (quoting Erica P. John Fund, Inc. v. Halliburton Co., 563 U.S. 804, 809 (2011)). In the antitrust context, plaintiffs must prove (1) that the defendant violated federal antitrust law; and (2) that the antitrust violation caused them some injury. Kleen Prods. LLC, 831 F.3d at 925. Plaintiffs must also show damages, but “it is well established that the presence of individualized questions regarding damages does not prevent certification under Rule 23(b)(3).” Id. (quoting Messner, 669 F.3d at 815.) Proceeding as an individual plaintiff at summary judgment, AutoLoop has already shown that the record evidence is sufficient to create a triable issue for each of the elements of its Sherman Act § 1 claim. (Summ. J. Op. at 48-74.) Now, at the class certification stage, the question becomes whether AutoLoop can show that this evidence is common to the class, or in other words, that “common evidence and a single, reliable methodology will prove [these] elements on a simultaneous, class-wide basis.” Steel, 2015 WL 5304629, at *5 (citing Parko, 739 F.3d at 1085-86, and Butler v. Sears, 727 F.3d 796 (7th Cir. 2013)).

AutoLoop argues that predominance is “clearly satisfied” in this case because it survived at summary judgment based entirely on classwide evidence. (AutoLoop Mem. in Supp. of Mot. for Class Certification [1422-1] at 10.) In other words, AutoLoop contends that any vendor in the class could use the exact same record that it used at summary judgment to establish all the elements of its claim in an individual trial. Regarding the first element, it is undisputed that whether Defendants conspired to align their data access policies to drive independent DIS providers from the market is a question common to each class member. Indeed, the answer to this question is “central” to the class's claim and the trial “will focus overwhelmingly on common proof of conspiracy-witness testimony, documents, email and phone records, economic evidence, and other evidence relating to Defendants' conduct.” Steel, 2015 WL 5304629, at *6. AutoLoop's summary judgment evidence on this score included: background information on the relevant DMS and DIS markets; internal emails and communications between CDK and Reynolds evidencing a sudden collaborative relationship; notes from a CDK executive describing an apparent agreement between defendants to lock down their respective DMSs; testimony from Authenticom's CEO about conversations he had with Defendants' executives in 2015 and 2016 where they suggested that they had an agreement in place to remove independent DIS providers from their DMSs; and Dr. Israel's opinions that the economic evidence supported the existence of an unlawful conspiracy that had anticompetitive effects in the relevant DIS markets. (See Summ. J. Op. at 15-64.) Dr. Israel explains that each vendor in the putative class could rely on the exact same evidence to prove the first two elements of their claim, which concern several interrelated common questions, such as market definition, market power, whether there was an agreement between Defendants, and whether that agreement had anticompetitive effects in the relevant antitrust markets for DIS. (Israel Suppl. Rep. at 5-12.)

Given the importance of the conspiracy element to an antitrust action, courts and commentators regularly accept that common questions here will predominate over individual questions. See, e.g., Kleen Prods. LLC, 831 F.3d at 928-29 (finding conspiracy issues common to class predominated over individual issues in direct purchasers' Sherman Act claim where plaintiffs offered common evidence showing a conspiracy and expert report showing that relevant market was amenable to collusion, and expert offered a reliable method of measuring aggregate classwide damages); Ready-Mixed Concrete, 261 F.R.D. at 169 (holding that common questions predominated with respect to conspiracy element and remarking, “[a]lthough Defendants contend that the question of impact is too individualized to warrant class certification, the common question of the existence of a horizontal price-fixing conspiracy usually satisfies Rule 23(b)(3)”); see also 7AA Wright, Miller & Kane, supra, § 1778 (“[W]hether a conspiracy exists is a common question that is thought to predominate over the other issues in the case and has the effect of satisfying the first prerequisite in Rule 23(b)(3).”). Accordingly, AutoLoop contends that based on these common questions alone it has satisfied predominance in full. (AutoLoop Mem. in Supp. of Mot. for Class Certification at 10.) CDK does not address AutoLoop's arguments here; instead, it focuses exclusively on arguing that individual issues related to the antitrust injury element predominate over common questions in this case.

The predominance inquiry with respect to antitrust injury asks whether the plaintiffs can “show that it [is] possible to use common evidence to prove that [Defendant's antitrust violation] injured the members of the proposed class.” Messner, 669 F.3d at 816. Antitrust injuries are “the type [of injuries] the antitrust laws were intended to prevent and reflect the anticompetitive effect of either the violation or of anticompetitive acts made possible by the violation.” Tri-Gen Inc. v. Int'l Union of Operating Eng'rs, Local 150, 433 F.3d 1024, 1031 (7th Cir. 2006) (internal quotation marks omitted) (quoting Brunswick Corp. v. Pueblo Bowl-O-Mat, Inc., 429 U.S. 477, 489 (1977)). At the class certification stage, “a plaintiff is not required to actually prove this element” and instead “‘need only demonstrate that the element of antitrust impact is capable of proof at trial through evidence that is common to the class rather than individual to its members.' ” Ploss, 431 F.Supp.3d at 1018 (emphasis in Ploss) (citing Messner, 669 F.3d at 818); see also Suchanek v. Sturm Foods, Inc., 764 F.3d 750, 757 (7th Cir. 2014) (“If the [district] court thought that no class can be certified until proof exists that every member has been harmed, it was wrong.”)

Dr. Israel opines that the source of antitrust injury took many forms in this case, including: increased prices for DIS services at CDK and Reynolds; costs associated with forced switching to CDK and Reynolds and away from preferred DIS providers; the elimination of the option to switch away from CDK and Reynolds to a preferred DIS option; removal of the upward quality pressure on CDK's and Reynolds's DIS offerings that is generated by market competition; and the exclusion of outside options to use in negotiations with CDK and Reynolds. (Israel Suppl. Reply ¶ 16.) He supports his opinion that monopolization of the DIS markets in this case caused the above injuries with economic analysis (see Israel Initial Rep. § VI) and empirical evidence of injury (see id. § VII.A). The economic analysis spans dozens of pages in his initial report and details the mechanisms Defendants used to stretch their DMS market power into DIS market power; how competition between CDK and Reynolds initially prevented each of them from independently blocking DIS providers, thus limiting their ability to fully leverage their market power; and how collusion, facilitated through communication between their executives and formal agreements, allowed them to fully leverage joint market power. (See Israel Initial Rep. § VI.) The empirical analysis measures the anticompetitive effects of the alleged conspiracy, largely in the form of graphs evidencing that the alleged conspiracy lowered the extent of independent DIS provided at Defendants' dealers; raised DIS prices for apps sold to Defendants' dealers and correspondingly raised Defendants' DIS revenues and profits; did not result in lower prices for Defendants' DMS (which is what, according to Dr. Israel, would be expected to happen if DIS prices increased absent collusion); and ultimately led to pass-through of the DIS price increase into app prices. (See Israel Initial Rep. § VII.A.) All this evidence suggesting antitrust injury, according to Dr. Israel, is common to the class and can be analyzed on a classwide basis. (See Israel Suppl. Reply at 5-14.)

Having provided an economic basis for the antitrust injury allegedly suffered by the Vendors, Dr. Israel then goes on to quantify the injury suffered, i.e., damages, through two regression models. (See Israel Suppl. Reply §§ IV, V.) As explained in the court's Daubert section above, Dr. Israel's primary model is his DID regression, which quantifies damages for two of the categories of injury he identified in his economic analysis: (1) increased prices for DIS services at CDK and Reynolds and (2) costs associated with forced switching¹² to CDK and Reynolds and away from preferred DIS providers. (Id. at 49 tbl.1.) The DID model estimates $395.46 million in total classwide damages related to price elevation and $409.56 million in damages when his model also considers costs associated with forced switching. (Id.) He also breaks down the damages for each class member by multiplying the overcharge percentage from his DID model by the volume of commerce data from CDK and Reynolds. (Id. at 88-92 tbl.9.) In addition, Dr. Israel produced a secondary model, his B&D model, that also quantities damages on a classwide and individualized basis. (See id. at 56 tbl.2, 94 tbl.10.) This model estimates $406.94 million in classwide damages. (Id. at 94 tbl.10.) In short,

12 Dr. Israel describes the basis for the “forced switching” harm as follows:

Buyers who preferred to purchase DIS from Authenticom, or any other independent DIS provider, were forced to switch after CDK and Reynolds closed their DMSs. This forced switching directly harmed these buyers due to the loss of their preferred choices. One source of this harm-but not all of it- occurred if the buyer had to pay a higher price to Reynolds or CDK for DIS than it did to its preferred vendor, as most did. But even if a buyer did not have to pay a higher price post-switch (which would be an unusual case), that buyer was still harmed by being forced to switch away from its preferred supplier. I provided a conservative quantification of the higher price portion of the harm via my estimate of forced switching overcharge.

(Israel Suppl. Reply ¶ 48; see also Israel Initial Rep. ¶¶ 217-19.)

Dr. Israel contends that his regression models offer a method of determining the amount of classwide damages related to price overcharges and can also be used by each vendor to make a prima facie showing of injury and damages related to price overcharge.

To the contrary, CDK insists that the Vendor Class should not be certified because antitrust injury cannot be shown through common proof. (CDK Opp'n to Class Certification [1456] at 16-17.) The crux of its argument is that some vendors in the class were uninjured, which CDK claims dooms a finding of predominance. It reaches this conclusion in two parts. First, CDK argues that, as a matter of law, the Vendors' antitrust injury cannot take any form other than being charged an above-market price. (Id. at 17-18.) In other words, CDK posits that to show antitrust injury, the Vendors must present evidence that they suffered a price overcharge. (Id. at 17 (citing O.K. Sand & Gravel, Inc., 36 F.3d at 573).) This would mean that the Vendors may not rely on economic evidence of reduced quality of DIS or loss of choice to show antitrust injury. From here, CDK, through its expert Dr. Haider, claims that the Vendors also cannot rely on Dr. Israel's regression model to show injury in the form of overcharges because his analysis obscures relevant market factors, such as vendor size (in terms of quantity of DIS product purchased) (see Haider Rep. ¶¶ 141-46); differences between “complex and simple” DIS (see id. ¶¶ 147-52); and changes to CDK's DIS program that began in July 2019, such as its fee waiver program and introduction of the Fortellis and Data Your Way programs (see id. ¶¶ 182-90). For example, Dr. Haider argues that when she modifies Israel's B&D model to account for the differences between large and small vendors,¹³the average overcharge for small vendors buying from CDK is slight (0.2%) and not statistically significant in the initial conspiracy period, and negative (-4.5%) and not statistically significant in the post-February 2015 period. (Id. ¶¶ 142-43 & tbl.14.) CDK claims that this shows that “a great many” vendors were uninjured, thus precluding class certification. (CDK Opp'n to Class Certification at 18.) Finally, in a related argument, CDK maintains that Dr. Israel's models also cannot be used to measure damages because they fail to “account for material differences in . . . vendors' damages.” (Id. at 20 n.8.)

13 Dr. Haider defined small vendors as the “vendors whose total fees paid during the proposed class period are in the bottom 80 percent for each defendant” and large vendors as “vendors whose total fees paid during the proposed class period are in the top 20 percent for each defendant.” (See Haider Rep. at 83 tbl.14, nn. 2, 3.)

AutoLoop disputes each of these points. At a high level, AutoLoop asserts that CDK “improperly conflates the question whether a vendor incurred a measurable overcharge (damages) with the question whether the vendor suffered an antitrust injury.” (AutoLoop Class Certification Reply at 12.) AutoLoop urges that antitrust injury is not limited to price overcharges, and it cites to various other courts that have suggested as much. See, e.g., Viamedia, Inc. v. Comcast Corp., 951 F.3d 429, 475 (7th Cir. 2020) (“The [antitrust] harms that typically flow from a competitive market shifting to total control by a monopolist include potentially higher prices, lower output, and reduced innovation.”); Lucasys Inc. v. PowerPlan, Inc., 576 F.Supp.3d 1331, 1350-51 (N.D.Ga. 2021) (collecting cases) (finding plaintiff sufficiently alleged antitrust injury where defendant's “anticompetitive conduct harmed competition by depriving customers of choice, and thereby prevented them from accessing lower-cost, higher-quality options”); Nilavar v. Mercy Health Sys., 142 F.Supp.2d 859, 874 (S.D. Ohio 2000) (noting that “higher prices, lower quality services, and less choice for consumers” are “the kind of injuries that the antitrust laws were enacted to prevent”). Accordingly, AutoLoop asserts that CDK cannot disprove antitrust injury by jiggering with Dr. Israel's regression models because they quantify only the damages related to one form of antitrust injury that occurred in this case: price overcharges. Put differently, AutoLoop claims that even if a regression analysis were to show that a vendor did not incur an overcharge during the conspiracy period, that vendor still suffered an antitrust harm, based on Israel's economic analyses in this matter, because they were forced to purchase in a market where the Defendants had less incentive to innovate or improve the quality of their DIS. (See Israel Suppl. Reply ¶¶ 42-49.) Moreover, Dr. Israel explains that regression analyses simply show correlation between relevant variables (the variables here being, overcharges and the alleged conspiracy)-they cannot prove causation (injury), which requires a separate theoretical economic analysis. (See AutoLoop Sur-Resp. at 2-3.)¹⁴ In other words, he argues that Dr. Haider's attempt to show a lack of antitrust injury for some vendors by performing separate regression analyses was deficient because her regressions were untethered from an underlying economic analysis that explains why some vendors and not others were uninjured. (Id. at 3 n.2.)

14 Citing ABA Section of Antitrust Law, Proving Antitrust Damages: Legal and Economic Issues 131 (3d ed. 2017), and EEOC v. Sears, 839 F.2d 302, 360 (7th Cir. 1988) (Cudahy, J., concurring in part and dissenting in part) (“Regression statistics by themselves only demonstrate correlations between variables; to move from correlation to causation, there must be some independent theory about the causal relationships of the variables.”)

In this vein, AutoLoop contends that even if the court were to focus exclusively on price overcharges for antitrust injury generally, Dr. Israel's economic analysis and damages models show a price overcharge for virtually every vendor. (See Israel Suppl. Reply at 88 tbl.9 (DID model showing 241 out of 244 vendors experiencing an average price overcharge); id. at 94 tbl.10 (B&D model showing all vendors experiencing an average price overcharge).) Dr. Israel also addresses Dr. Haider's challenges to the methodological decisions Dr. Israel made in his regression analyses. He defends his decision to not divide the sample into different subgroups, explaining that the economic evidence in this case shows that the impact of the conspiracy was market-wide; thus, the most appropriate way to estimate overcharges for the class members is to calculate the average overcharge in Defendants' DIS markets. (See id. ¶¶ 116-34.) He adds that the modifications Dr. Haider has made to his regression model, which she claims shows many vendors were uninjured, do not establish that any vendors were in fact not overcharged because by her own admission they lack statistical significance. (Id. ¶¶ 13132.) Dr. Israel further points out that her results can be read as showing positive overcharges even for the small vendors.¹⁵(Id.) He also opines that CDK's actions since April 2019, i.e. granting fee waivers from some DIS fees and introducing the Fortellis and Data Your Way programs, do not undermine his conclusion that antitrust injury is common to the class because his regression model takes into account the fee waivers and because CDK's DIS platform remains monopolized. (Id. ¶¶ 57-63, 13742.) Finally, AutoLoop stresses that Dr. Haider's attacks on Dr. Israel's regression models are unrelated to the question of class certification; indeed, neither Dr. Haider's nor Dr. Israel's regression models necessitate individualized inquiries into specific class members' circumstances-both use classwide statistical methods to estimate damages for all class members. (Id. ¶ 24.)

15 Dr. Israel explains that “Dr. Haider finds a -4.5% overcharge with an 11.9% standard error for her bottom 80% of CDK vendors, and thus the top end of the 90% confidence interval is 15.1% (-4.5% plus product of the critical value of 1.645 multiplied by 11.9%).” (Israel Suppl. Reply ¶ 132 n.205.)

Notwithstanding the extensive briefing the parties devote to this issue, the court is not concerned that any potential individualized issues related to antitrust injury will “overwhelm” the bevy of questions common to the Vendors. Their sparring on antitrust injury largely ignores the relatively narrow inquiry the court must decide: is it possible for the Vendors to prove injury with evidence that is common to the class? The Vendors have proven that they can. The court has ruled numerous times now, and has done so again today, that Dr. Israel's economic analyses related to antitrust injury are reliable. His opinions on antitrust injury in this case (see Israel Initial Rep. §§ VI, VII.A) are undoubtedly common to the class, and each vendor can rely on them to make out their prima facie showing of injury. Indeed, one vendor, AutoLoop, has already relied on Dr. Israel's opinions to defeat a motion for summary judgment and has given the court a sneak peek at what trial might look like. Tellingly, of the challenges CDK raised against AutoLoop at summary judgment, virtually none were individual to AutoLoop's specific claim and instead all concerned issues that would be common to each vendor. (See Summ. J. Op. at 48-74.) CDK has not shown why any other class member could not make out their case with the same evidence.

The debate between the parties about whether antitrust injury is confined to price overcharges or instead includes more abstract effects (decreases in quality or loss of choice) is largely academic because Dr. Israel provides economic evidence of classwide price overcharges and quantifies this injury with his regression analyses. (Id.; Israel Suppl. Reply §§ IV, V.) And, Dr. Haider's attempt to undermine Dr. Israel's antitrust injury conclusions related to overcharge by attacking his damages model is misplaced. As AutoLoop and Dr. Israel explain, while regression models can provide strong evidence of antitrust injury by demonstrating that a plaintiff incurred net positive damages, their inability to account for mitigation measures taken by either party means that the absence of damages does not automatically imply the absence of injury in the form of an overcharge at some point in the class period. See In re Ethylene Propylene Diene Monomer (EPDM) Antitrust Litig., 256 F.R.D. 82, 88-89 (D. Conn. 2009) (“Proving damages proves injury because damages necessarily indicate that the plaintiff has been impacted or injured by the antitrust violation; the converse, however, is not necessarily true .... [I]t is possible for a plaintiff to suffer antitrust injury-in-fact and yet have no damages because it has taken steps to mitigate the actual price paid through rebates, discounts, and other non-price factors such as lowered shipping costs, technical services, or any other type of purchase incentive.”)

Notably, Dr. Haider does not even challenge the ability of Dr. Israel's DID model to assess classwide damages; instead, she quibbles about the precision at which it can allocate individual damages. However, the Seventh Circuit has counseled that “[t]he determination of the aggregate classwide damages is something that can be handled most efficiently as a class action, and the allocation of that total sum among the class members can be managed individually, should the case ever reach that point.” Kleen Prods LLC, 831 F.3d at 929. It has also squarely rejected the notion that plaintiffs must prove each class member was injured to certify a class, which is essentially the burden CDK asks the court to impose on the Vendors. Id. at 927 (“While we have no quarrel with the proposition that each and every class member would need to make such a showing in order ultimately to recover, we have not insisted on this level of proof at the class certification stage.”). To be sure, predominance issues do arise when a proposed class includes members “who could not have been harmed,” as opposed to a class that includes members “who were not harmed.” Messner, 669 F.3d at 825 (emphasis in original). But CDK does not argue that any vendor could not have been harmed, and nothing in the record suggests that any of the class members were immune from harm-each class member purchased DIS from Defendants during the alleged conspiracy period where prices were inflated. The fact remains that courts handling antitrust class actions routinely endorse the practice of using regression models to estimate an average overcharge that can establish injury and measure damages on a classwide basis, so long as the regression analysis is based on a rigorous application of economic theory. See, e.g., In re NorthShore Univ. HealthSystem Antitrust Litig., 657 F.Supp.3d 1077, 1090-92 (N.D. Ill. 2023); Broiler Chicken, 2022 WL 1720468, at *20; In re Pork Antitrust Litig., 665 F.Supp.3d 967, 1005-06 (D. Minn. 2023); Bumble Bee Foods LLC, 31 F.4th at 677-78; In re Disposable Contact Lens Antitrust, 329 F.R.D. 336, 389 (M.D. Fla. 2018).

Acknowledging that Messner and Kleen counsel in favor of certifying the Vendor Class, CDK asks this district court to reexamine the holdings in those cases in light of the Supreme Court's ruling in TransUnion LLC v. Ramirez, 594 U.S. 413, 431 (2021). (CDK Sur-Reply in Opp. to Class Certification [1510] at 3.) According to CDK, “TransUnion and Rule 23 require a practical analysis: Every class member must have Article III standing and thus concrete injury, so predominance is only met if common evidence can sort out members who in fact suffered no injury.” (Id. (cleaned up).) The court need not devote much time to this challenge for two reasons. First, this court is bound by Seventh Circuit precedent and lacks authority to overturn its rulings unless it is “almost certain that the higher court would repudiate the doctrine if given the chance to do so.” Olson v. Paine, Webber, Jackson & Curtis, Inc., 806 F.2d 731, 734 (7th Cir. 1986). This is not one of those instances. CDK's contention that TransUnion altered the class certification analysis is wrong-the Supreme Court expressly declined addressing whether every class member must show standing before certification- thereby leaving intact the holdings of Messner and Kleen. TransUnion LLC, 594 U.S. at 431 n.4. (“We do not here address the distinct question whether every class member must demonstrate standing before a court certifies a class.”) (emphasis in original). The Vendors have provided evidence capable of establishing concrete injury (price overcharges) on a classwide basis; this “is sufficient to show an injury-in-fact traceable to the defendants and redressable by a favorable ruling,” which is all that is required for the Vendors to show Article III standing at the class certification stage. Bumble Bee Foods LLC, 31 F.4th at 682 (rejecting similar standing argument raised by antitrust defendant); see also Huber v. Simon's Agency, Inc., 84 F.4th 132, 155 (3d Cir. 2023) (“TransUnion suggests that the need for unnamed class members to demonstrate Article III standing depends on the stage of litigation[,] . . . at certification . . . it is not necessary for each member to prove his or her standing for the class action to be justiciable.”); Lujan v. Defs. of Wildlife, 504 U.S. 555, 561 (1992) (explaining that standing must be shown “with the manner and degree of evidence required at the successive stages of the litigation”). Second, even under CDK's fantasized conception of standing, the Vendor Class withstands scrutiny. As the court understands it, CDK argues that TransUnion requires AutoLoop to establish, at class certification, not only that every vendor has suffered concrete injury, but also that common evidence can filter out uninjured class members. Dr. Israel's regression models do exactly that. His DID and B&D models rely on common evidence to establish damages related to price overcharges and can filter out uninjured class members. (See Israel Suppl. Reply at 88 tbl.9; id. 94 tbl.10.) At bottom, CDK's standing challenge is a Hail Mary attempt that falls way short.

The court is satisfied that the Vendors have provided common evidence that, if believed by a jury, can prove each element of their § 1 Sherman Act claim, and that any individualized inquiries will not overwhelm the questions of law and fact that are common to the class.

2. Superiority

When a court finds that common questions predominate, they generally also find that superiority is also met. See Moehrl, 2023 WL 2683199, at *22. “Superiority will be found ‘when a class action would achieve economies of time, effort, and expense, and promote uniformity of decision as to persons similarly situated, without sacrificing procedural fairness or bringing about other undesirable results.' ” Id. at 21 (quoting Wilkins v. Just Energy Grp., Inc., 308 F.R.D. 170, 190 (N.D. Ill. 2015). AutoLoop argues that a class action is superior here because the alternative would require every plaintiff to pursue a separate trial with “needless repetition of the same evidence on all of the common issues addressed by the [c]ourt at summary judgment: the existence of Defendants' conspiracy, its impact on competition and DIS prices, CDK's alleged procompetitive justifications, and Dr. Israel's damages model.” (AutoLoop Mem. in Supp. of Mot. for Class Certification at 14.) It also argues that class treatment is preferrable because forcing vendors to pursue their claims individually would be cost-prohibitive for those with low damages, especially given the expense of litigating an antitrust case. See Murray v. GMAC Mortg. Corp., 434 F.3d 948, 953 (7th Cir. 2006) (noting that class actions were designed for situations “in which the potential recovery is too slight to support individual suits, but injury is substantial in the aggregate”). Finally, it notes that absent class members have had years to file their own cases but only one other vendor has done so (Cox Automotive), suggesting that a class action is a superior means to obtain relief for the class.

CDK does not squarely address AutoLoop's superiority argument, effectively waiving any challenge to class certification on this basis. See Bonte v. U.S. Bank, N.A., 624 F.3d 461, 466 (7th Cir. 2010) (“Failure to respond to an argument . . . results in waiver.”). Either way, the court is confident that certifying the Vendor Class is the superior option here given the forgoing discussion.

C. Class Period

Finally, CDK argues that even if the Vendors' Class is ultimately certified, the court should shorten their proposed class period. Autoloop proposes a class period covering all DIS sales from October 1, 2013 to the present. (See AutoLoop Class Certification Reply at 6.) In CDK's view, Plaintiffs have not shown that they are entitled to class treatment for the latter half of this period. Notably, Reynolds-one of the two accused coconspirators-settled with the indirect purchaser plaintiffs (a group of retail car dealerships, referred to in this MDL as “Dealers”) in late 2018, which should, according to CDK, be enough to end the class period on its own. Moreover, CDK identifies purportedly significant changes in the market that have taken place since this settlement and the close of fact discovery in mid-2019, including the rise of third-party DMS competitors and its own introduction of “new and pro-competitive initiatives” in response to this shift, such as the fee waiver promotion and the Fortellis and Data Your Way programs. These shifts, it argues, render AutoLoop's expert analyses-first completed in 2019, based solely on data from merits discovery-outdated for assessing whether common questions of injury and damages predominate among the proposed Vendor Class over the past five years. Accordingly, CDK proposes two alternatives for an earlier end date to the class period: (1) October 2018, the month of the Reynolds settlement; or (2) December 2019, the month when Plaintiffs submitted their initial merits expert reports.

The court rejects CDK's claim that the Reynolds settlement is by itself sufficient to end the class period. As an initial matter, AutoLoop correctly points out that the length of Defendants' conspiracy is both a merits issue and a common question that affects the class as a whole. See Kaplan v. IA.C. Cap. Advisors, L.P, 146 F.Supp.3d 588, 589 (S.D.N.Y. 2015). This does not mean that it is per se inappropriate to consider on class certification: “[c]ourts have long recognized that they ‘must evaluate some aspects of the merits of plaintiffs' proposed class period to determine the appropriate endpoints.' ” W. Virginia Pipe Trades Health & Welfare Fund v. Medtronic, Inc., 325 F.R.D. 280, 293 (D. Minn. 2018) (quoting In re Data Access Sys. Sec. Litig., 103 F.R.D. 130, 143 (D.N.J. 1984)). But in this case, the parties have already thoroughly litigated the merits at summary judgment, and if CDK wanted to raise the Reynolds settlement as a potential cutoff date for the underlying conspiracy, it should have done so then.

Even if the court were to reopen the summary judgment analysis for a “peek” at CDK's argument on this front, Schleicher v. Wendt, 618 F.3d 679, 685 (7th Cir. 2010), the mere fact that Reynolds settled with the Dealers says nothing about whether it ceased conspiring with CDK-or whether its actions have continued to affect the market to this day. To withdraw from a conspiracy, a defendant must show “[a]ffirmative acts inconsistent with the object of the conspiracy,” such as alerting authorities or resuming procompetitive behavior. Drug Mart Pharm. Corp. v. Am. Home Prods. Corp., 288 F.Supp.2d 325, 329 (E.D.N.Y. 2003) (citing United States v. U.S. Gypsum Co., 438 U.S. 422, 464 (1978)). A settlement agreement may, in some cases, constitute such an “affirmative act” if it requires the defendant to cease their accused anticompetitive conduct. See id. at 330-33 (finding settlement agreement with term controlling defendant's future pricing practices sufficient to end conspiracy). But the Dealers' October 2018 settlement with Reynolds contains no injunctive provisions requiring Reynolds to do anything of the sort.¹⁶ And even supposing Reynolds had successfully withdrawn, if the “effect of the conspiracy lingers beyond the end of the formal collusion . . . [t]his also taints the post-conspiracy prices and may give rise to extended damages.” Areeda & Hovenkamp, supra, ¶ 395b2; see In re Rail Freight Fuel Surcharge Antitrust Litig. (No. II), No. 1:19-CV-03379 (BAH), 2020 WL 5016922, at *24 (D.D.C. Aug. 25, 2020) (“[L]ingering effects of a completed conspiracy after a class period may be remediated upon successful proof of the underlying anticompetitive conduct.”); In re Polyurethane Foam Antitrust Litig., 314 F.R.D. 226, 268-69 (N.D. Ohio 2014).

16 (See Ex. 1 to Decl. of Peggy Wedgworth [427-2] at 10, 12-19 (detailing $29.5 million payment to Dealership Class in exchange for release of claims, but no injunctive provisions).)

Nor is the end of fact discovery a sufficient basis to justify cutting off AutoLoop's proposed class period. CDK is correct that trimming the class period may be appropriate where the evidence supports Rule 23's requirements for some, but not all, years. See, e.g., In re TFT-LCD (Flat Panel) Antitrust Litig., 267 F.R.D. 291, 303 (N.D. Cal. 2010) (shortening proposed class period from eleven to seven years on typicality and predominance grounds where class representatives made no purchases in the first several years of the proposed period and evidence largely post-dated these years), abrogated on other grounds by In re ATM Fee Antitrust Litig., 686 F.3d 741 (9th Cir. 2012). Here, however, AutoLoop has met its burden of justifying, by a preponderance of the evidence, a class period covering the full eleven years (and counting) since the onset of the alleged conspiracy in 2013. As discussed in the Daubert section above, Dr. Israel has updated his model through 2023 using new Authenticom data, and his renewed analysis shows continued overcharges over the latter five years of the proposed class period-notwithstanding the various changes in the market that CDK identifies. CDK will be free to make its best case at trial for why these market changes defeat the Vendors' claims for ongoing antitrust harms up to the present day.¹⁷

17 The Vendors present various arguments to the contrary. As Dr. Israel notes in his supplemental reply, the fact that competition may have increased in the DMS market says nothing about whether Defendants' practices regarding their DIS systems-from which they continue to exclude third-party integrators-remain anticompetitive. (Israel Suppl. Rep. ¶ 58.)

CONCLUSION

CDK's motion to exclude certain opinions of Dr. Mark Israel [1459] is denied and AutoLoop's motion for class certification [1422] is granted. The following Vendor Class is certified under Rule 23(b)(3):

All automotive software vendors (i.e., persons or entities engaged in the sale of software solutions to automotive dealerships) located in the United States that, at any time since October 1, 2013, have purchased data integration services from CDK or Reynolds. Excluded from the class are (1) CDK, Reynolds, and any of their officers, directors, employees, subsidiaries, and affiliates; (2) Cox Automotive, Inc. and its subsidiaries and affiliates; and (3) automotive software vendors that first purchased data integration services from CDK after June 5, 2018; (4) any federal, state governmental entities, any judicial officer presiding over this action and the members of his/her immediate family and judicial staff, and any juror assigned to this action.

Having completed all pretrial matters, the court recommends that the JPML remand this case for trial to the Western District of Wisconsin pursuant to § 1407. (See Mem. Op. and Order [18] ¶ 18-cv-2521.)