A. Underlying facts and nature of the dispute.

This action was commenced by the EEOC on behalf of Linda Kim and a class of female job applicants who complain of sex discriminatory pre-employment testing. CRT developed a test to measure and collect data regarding knee, shoulder, and trunk strength, range of motion, and endurance in the form of a body index score (BIS) for individuals subjected to the test. The Complaint alleges that the use of the CRT test has a disparate impact on female job applicants, has caused Schuster to revoke conditional offers to Ms. Kim and others similarly situated, and deprived them of equal employment activities because of their sex. (Doc. 1 ¶¶ 17-18.)

The Complaint (Doc. 1) does not allege what type of business Defendant Schuster Company (“Schuster”) operates; however, it appears Schuster hires drivers in the trucking industry. Schuster uses the CRT test in its driver hiring decisions. A job applicant receives a job offer from Schuster conditioned on receiving a passing score on the CRT test. Schuster does not perform the testing or maintain the test data. Rather, a physical therapist administers the test using an apparatus that transmits data to CRT's structured query language (“SQL”) database.[1] Upon receipt of the data from the therapist's office, the BIS score[2] is then transmitted to the therapist's office.

This dispute involves certain subpoenas the EEOC issued to CRT “in attempt to understand what the BIS formula is and to obtain the relevant component score data the formula uses.” (Doc. 37-1 at 2.) While it is more implied than stated, the EEOC's evident concern is that the test is designed or scored in a manner that discriminates against female test-takers.[3]

The instant dispute arises after efforts by the parties have failed to resolve the dispute and a Rule 30(b)(6) deposition of CRT did not suffice (at least from the EEOC's perspective) to answer their questions about the CRT test. The EEOC states:

(Doc. 37-1 at 3.)

B. The subpoenas, responses, and objections

A February 4, 2020 subpoena to Brett Crosby, Cost Reduction Solutions, requested:

(Doc. 37-3.) There followed an email exchange between the parties' lawyers regarding upcoming depositions. (Doc. 37-4.) This exchange also set up one of the principal issues now before me as expressed by CRT's counsel:

(Doc. 37-4.)

A June 4, 2020 subpoena issued by the EEOC to Brett Crosby, Cost Reduction Solutions, requested:

(Doc. 37-5.) Document CRT000426, which identifies the subject matter of the subpoena, is an email from Erin Dunleavy at DCI Consulting Group to Jim Wildman at CRT seeking information “on the criterion validation front” relating to CRT's test. (Doc. 37-6.) On June 19, 2020, CRT objected as follows:

(Doc. 37-7.) The objection is contained in a document titled “Response to Request for Production of Documents,” and does not reference any subpoena. (Id.)

On August 17, 2020, the EEOC took the Rule 30(b)(6) deposition of CRT, apparently in the person of Brett Crosby. Counsel for the EEOC summarized what it learned out the accessibility of the information in the SQL database:

*3 (Doc. 37-12.)

C. The subject of the subpoenas

The EEOC prays that the Court “order CRT to produce data sufficient to show what the BIS formula is and the component data that go into the formula for each Schuster test-taker.” (Doc. 37-1 at 12.) This dispute is complicated by several factors relating to the data requested. First, the EEOC clearly requested some “data” from CRT. However, CRT, as discussed below, disputes that the EEOC's subpoenas requested the data it now seeks to compel. Second, apparently nobody at CRT knows the formula. The current president of CRT testified that the formula was developed by the company's founder, but he passed away. (Doc. 38-1, at 14.). Third, the dispute is also complicated by the way the data is maintained. CRT has an internal SQL server but CRT does not maintain the database containing test data itself. Rather, CRT has a vendor who maintains the database and server and who can write queries to pull data regarding test-takers. (Id. at 25-27.)

A. The Parties' Arguments

Having generally set out the nature of the dispute, I find it helpful to start with CRT's arguments, i.e., its objections to producing the items requested. CRT first argues that the formula does not fall within the scope of the subpoena as written. CRT does not dispute that the formula is “data,” (Doc. 41 at 3) however, CRT asserts the formula is not transmitted between CRT and the CRT leased device and, therefore, is not within the scope of the subpoena.

CRT also objects under Rule 45(e)(1)(D) to providing component score information for Schuster test-takers because it is “electronically stored information from sources that the person identifies as not reasonably accessible because of undue burden or cost.” CRT asserts the EEOC subpoena places an undue burden upon it and the EEOC has not shown good cause to support the requested discovery. With respect to good cause, CRT contends the data regarding test-takers is not necessary “at this stage of the litigation” because of the burden shifting in cases of this nature. CRT points out that in disparate impact litigation, there is a three-step burden-shifting framework:

(Doc. 41 at 8) (citing Ricci v. DeStefano, 557 U.S. 557, 578, 129 S. Ct. 2658, 2673 (2009); EEOC v. Dial Corp., 469 F.3d 735, 742 (8th Cir. 2006)). CRT asserts that if the EEOC fails to make its prima facie case or Schuster fails to show the business necessity of the test, then the EEOC will never need the data to show a less discriminatory alternative exists.

*4 CRT also argues that the EEOC failed in its obligation to take reasonable steps to avoid imposing undue burden on it by refusing to accept a smaller sampling of Schuster test-takers as CRT offered. CRT argues that it should not be compelled to produce its data until the EEOC explains how it intends to create or provide an “alternative practice that has less disparate impact and serves the employer's legitimate needs.” Ricci, 129 S. Ct. at 2661.

Finally, CRT argues that cost shifting is appropriate because providing the data will impose significant costs on it. CRT does not assert that formulating and retrieving the requested data by an SQL query creates a burden. In his declaration, Mr. Crosby admits he does not know how difficult, time-consuming, or expensive it is to have its vendor, RSM, develop the SQL query. Rather, CRT contends its president, Brett Crosby, will need to take a time-intensive manual review of the data. The need for this manual review by CRT's president arises from the possibility that the clinician who performed the test may have entered in Schuster's company ID incorrectly. In other words, CRT argues this review is the only way CRT can be sure the EEOC received all of the Schuster tests and only the Schuster tests. Mr. Crosby estimates the value of his time to perform this necessary work is $6,000 to $8,000.

The EEOC argues that there is no reasonable argument that the raw results sent from the CRT machine leased by Schuster to CRT's database and the BIS score returned to Schuster are not “data.” Indeed, CRT seems to admit the information exchanged is not data. While the subpoena does not specifically mention the “formula” used to convert the raw data into a BIS, the EEOC asserts that it clarified in subsequent communications with CRT that it is seeking the formula as part of this request, even if the formula itself is not transmitted. The EEOC asserts that after six months of good faith negotiations trying to clarify its request and determine what CRT possesses, it should not have to start from scratch with a new subpoena that references the formula.

The EEOC asserts that the formula and component data are central to the issues in this case. The EEOC contends it requires the information to prove two elements of its claim and prepare experts on those issues; to wit: whether the selection procedure is job related and consistent with business necessity and whether there are less discriminatory alternatives. See EEOC v. Dial Corp., 469 F.3d 735, 742 (8th Cir. 2006).

The EEOC argues that it is an unreasonable expense to have Mr. Crosby perform his proposed manual review of the approximately 1,800 tests in question when its vendor could write a suitable query to generate the necessary report in less than an hour. The EEOC also contends that preparation of an SQL query will not be a significant expense, that CRT has an interest in the outcome of the litigation, and that the litigation is of public importance.

B. Analysis

Federal Rule of Civil Procedure 45 provides the framework for requiring third parties to “produce designated documents, electronically stored information, or tangible things in that person's possession, custody, or control.” Fed. R. Civ. P. 45(a)(1)(A)(iii).

1. Whether the formula falls within the scope of the subpoena.

If I were to look at the scope of a subpoena without considering the context that developed around this request, it would be difficult to determine whether the formula in question was ever transmitted. CRT provides some of the illuminating context. CRT states that because the formula never leaves its digital home on CRT's server, it is not transmitted and, therefore, it cannot be subject to the subpoena. I agree that, with this information alone, CRT could be forgiven for not producing the formula. Additional context, however, argues for a different conclusion.

*5 First, this is not a situation where Mr. Crosby showed up for a deposition pursuant to a subpoena duces tecum without the formula. The EEOC is not trying to sanction him for failing to produce a document arguably beyond the scope of its vague subpoena. In the case at bar, the EEOC (perhaps unaware of exactly how the formula was maintained or where it resided) issued a subpoena that may not have been broad enough or specific enough to include the formula itself. The additional context that CRT ignores, however, is the approximately six months of discussions and negotiations between the EEOC and CRT regarding the information the EEOC was seeking. By July 3, 2020, CRT certainly knew what information the EEOC was after when its attorney stated, “You indicated that your aim was to learn the CRT algorithm[4] to understand how the data is processed and a BIS score derived.” (Doc. 37-9 at 15.)

Neither party cites any relevant authority regarding the interpretation of the language of a subpoena. CRT offers no support for the implicit assertion that a court is bound to ignore the EEOC's good faith efforts to explain its request and CRT's expressed understanding of it. CRT acknowledges these discussions can alter the scope of the request when it points out: “Although the Subpoena requests all information transmitted between CRT and the ‘Schuster leased CRT device,’ subsequent correspondence between the parties clarified the EEOC's request was for data related to Schuster test-takers.” (Doc. 41 at 4 n.3.) It would elevate form over substance and ignore the parties' efforts to resolve issues regarding the scope of the subpoena if I were to interpret it absent the context of these discussions. CRT has known the EEOC was seeking the formula since at least July. There seems little point in requiring service of a modified subpoena to formalize this request. This Court has acknowledged the important role of the parties' efforts to define the scope of a subpoena after issuance. Progressive Cas. Ins. Co. v. F.D.I.C., 302 F.R.D. 497, 505 (N.D. Iowa 2014) (“[W]hile the original scope of the subpoena was quite broad, [the defendant] later took reasonable steps to narrow that scope and eliminate any undue burden.”), aff'd, 49 F. Supp. 3d 545 (N.D. Iowa 2014). Thus, I overrule CRT's objection that the June 4, 2020 subpoena seeks information beyond its scope.

2. Whether Fed. R. Civ. P. 45(e)(1)(D) prevents the EEOC from obtaining discovery.

There is no dispute about whether the information requested is relevant. Am. Broad. Cos. v. Aereo, Inc., No. 13-MC-0059, 2013 WL 5276124, at *3 (N.D. Iowa Sept. 17, 2013) (“As a threshold matter, the Court must determine whether the discovery sought by Aereo is relevant to any issue in the underlying action. A party may obtain discovery regarding any nonprivileged matter that is relevant to any claim or defense.”). Instead, the dispute centers on the burden imposed on CRT in responding to the subpoena. “[F]ederal courts are particularly mindful of Rule 45's undue burden and expense limitations.” Progressive, 302 F.R.D. at 504-05 (citing Miscellaneous Docket Matter No. 1 v. Miscellaneous Docket Matter No. 2, 197 F.3d 922, 927 (8th Cir. 1999); American Broadcasting Cos., 2013 WL 5276124, at *7; Precourt v. Fairbank Reconstruction Corp., 280 F.R.D. 462, 467 (D.S.D. 2011)).

Rule 45(e)(1)(D) provides:

*6 The first stage of the inquiry is determining if the information is not reasonably accessible because of undue burden or cost. See Fed. R. Civ. P. 45(e)(1)(D). Whether data is “not reasonably accessible because of undue burden or cost” turns on “factors such as the complexity of the ESI and the nature of the media on which the ESI is stored.” KAIST IP US LLC v. Samsung Elecs. Co., No. 2:16-CV-01314-JRG-RSP, 2017 WL 9937760, at *2 (E.D. Tex. Dec. 21, 2017) (citing Zubulake v. UBS Warburg LLC, 217 F.R.D. 309, 318–19 (S.D.N.Y. 2003) (“[W]hether production of documents is unduly burdensome or expensive turns primarily on whether it is kept in an accessible or inaccessible format (a distinction that corresponds closely to the expense of production).”)).[5]

CRT admits that the database can be queried by having a third party write a query. While there is some dispute about how long this step would take and how much it would cost, CRT does not argue that this step alone creates undue burden. (Doc. 41 at 5.) Rather, CRT attributes the undue burden or cost to its insistence that Mr. Crosby will have to undertake a laborious review of all the data to weed out test-takers that were erroneously included or excluded. This process is necessary, CRT contends, because there have been instances where a human error at the time a test-taker's data was entered resulted in either being over-inclusive or under-inclusive with respect to Schuster test-takers.

CRT's attention to detail during this litigation, if not at the time of data entry, is laudable. Nevertheless, the possibility of such an error is insufficient justification for the review CRT proposes. First, no one has alleged the data is confidential. While CRT claims its formula is proprietary, there has been no concern raised that some sort of privilege or confidentiality is breached when the EEOC reviews any of the personal data from a test-taker. Nor has anyone claimed such a breach occurs if a non-Schuster test-taker's data is inadvertently disclosed to the EEOC. Thus, even if an error attributable to data entry occurs and a non-Schuster test-taker's data is shared, the harm resulting from the error is minimal.

Second, CRT relies upon one instance of a data entry error to justify its position:

(Doc. 41 at 6 n. 5.) This incident does not make Mr. Crosby's proposed review of the data necessary. Rather, it shows the minimal time and effort he might need to expend in the event additional errors are discovered after the data has been produced. Presumably Schuster knows the names of its test-takers. Thus, the parties can confer regarding anomalous data before troubling Mr. Crosby to query and correct the production.

Finally, imperfect production of discovery materials is fairly commonplace in modern litigation, despite the parties' best efforts to avoid it. In document intensive cases, parties draft search terms or queries that often result in overinclusive and underinclusive production of documents. This seldom results in the intensive pre-production data review that CRT contemplates. Rather, potential errors are dealt with by attempts to refine the queries to produce relevant documents and by clawback provisions. See, e.g., In re Tier 1 Jeg Telecomms. Cases, No. 4:07-CV-00043, 2013 WL 12158598, at *3 (S.D. Iowa Nov. 25, 2013) (“Clawback provisions of this kind serve the salutary purpose of promoting expeditious production in document-laden litigation without the expensive and time-consuming effort of a rigorous attorney-client/work-product review. The producing party may make document drops secure in the knowledge the protection for inadvertently produced privileged or protected documents will be preserved.”). Here, CRT's desire for a perfect response to the subpoena has become the enemy of a good response, i.e., one that is unburdensome, timely, and cost-effective. I conclude CRT has not shown the items sought are not reasonably accessible because of undue burden or cost.

*7 If CRT had established the data sought is not reasonably accessible, I would need to establish if there was good cause to order the production under Rule 45(e)(1)(D) “considering the limitations of Rule 26(b)(2)(C).” I will address this issue for the sake of completeness in the event of review. Neither party references the limitations of Rule 26(b)(2)(C) which states:

I find no basis to conclude that any of these limitations applies. Indeed, there seems to be no source for the subpoenaed information, other than from CRT. Moreover, as stated above, the information sought is within the scope of Rule 26(b)(1).

CRT takes three different but related tacks to show there is no good cause to order the discovery. First, CRT argues the information may never be needed pursuant to the applicable burden-shifting framework because the EEOC may never be able to show a prima facie case of disparate impact and Schuster may not be able to show business necessity. (Doc. 41 at 8) (citing Ricci v. DeStefano, 557 U.S. 557, 578 (2009); EEOC v. Dial Corp., 469 F.3d 735, 742 (8th Cir. 2006)).

Discovery in this matter has not been ordered to be conducted in stages. The parties are required to conduct discovery on all aspects of their claims and defenses. Thus, the parties do not have the ability to await a determination on the EEOC's prima facie case before turning to the issue of whether there is a less discriminatory alternative. Moreover, I am not convinced that the requested data relates only to the existence of a less discriminatory alternative. The data may also support the parties' respective positions regarding the prima facie case and business necessity.

CRT also contends there is not good cause to provide all the Schuster test-taker data when it has offered data for a smaller sample of test-takers. CRT asserts that under Rule 45(d)(1) the EEOC has an obligation to take reasonable steps to avoid imposing undue burdens or expense on CRT. This is undoubtedly true; however, for the reasons set forth above, the burden placed on CRT is caused by its insistence on a manual review of the data. CRT argues “not only does Pepsi not need Coke's components to produce what it believes is a better product, it certainly does not need every can of Coke!” (Doc. 41 at 10.) This argument might be more persuasive if the only issue was discovering CRT's (or Coke's) formula. The EEOC responds,

(Doc. 42 at 3.) Thus, while the formula might be derived from a smaller subset, the full set of data is relevant to determine whether the test is discriminatory, if there is a business necessity for it, or if there are less discriminatory alternatives.

*8 Similarly, CRT argues no good cause exists to allow the discovery because the EEOC has not explained how it will demonstrate the existence of an available alternative practice that has less disparate impact and serves the employer's legitimate needs. Again, this argument ignores the fact that discovery in this matter is not staged. The parties do not have the luxury of performing discovery in stages and presenting their cases piecemeal. Moreover, an understanding of the data in question is related to the existence and scope of a disparate impact and a determination of how the CRT test serves Schuster's legitimate business needs. Thus, good cause exists to obtain the data.

C. Whether Cost Shifting is Appropriate.

Rule 45(d)(2)(B)(ii) provides:

I agree with the following statement of the applicable standards:

Cedar Rapids Lodge & Suites, LLC v. Seibert, No. 0:14-CV-04839-SRN-KMM, 2018 WL 3019899, at *2 (D. Minn. June 18, 2018).

While the proposed review of the data by Mr. Crosby could result in some expense, I have previously determined that review is not necessary to facilitate the production. Therefore, it should not be considered part of the expense resulting from the compliance. Mr. Crosby does not know how much it would cost a third party to construct the query. (Doc. 41-2 ¶ 23.) The EEOC estimate, based on statements by CRT regarding prior work by its vendor, is that the work would take less than an hour at $200 per hour. Even if writing the query costs three times that estimate, it is not a significant expense.

*9 CRT argues, “To determine the proper allocation of costs, the Court should consider (1) whether the non-party actually has an interest in the outcome of the litigation; (2) whether the non-party can more readily bear the costs than the requesting party; and (3) whether the litigation is of public importance.” (Doc. 41 at 13 quoting Wells Fargo Bank, N.A. v. Konover, 259 F.R.D. 206, 207 (D. Conn. 2009) (citations omitted)). CRT does not take issue with the EEOC's position that CRT has an interest in this matter and that the litigation is of public importance. I agree that the litigation clearly has implications for CRT's business; thus, CRT has an interest in the litigation. Moreover, the litigation could have impact on the use of this and similar tests. Thus, the litigation is of public importance.

Regarding whether CRT can more readily bear the costs of production than the EEOC, once again I return to my conclusion that the manner of production proposed by CRT is unnecessarily time-consuming and expensive. CRT can readily bear the necessary cost of compliance with the subpoena.

For all of the reasons set forth above, Plaintiff EEOC's motion to compel is granted and CRT's objections are overruled. CRT's request for cost shifting is denied. CRT shall produce the information consistent with the foregoing within 14 days of this order.

IT IS SO ORDERED this 7th day of October, 2020.

Schuster leases the equipment necessary for the CRT test from CRT and places it in the possession of the physical therapist for such testing. (Doc. 37-11.)

Both parties use the term “BIS score.” The term bears the same sort of redundancy as “GPS System,” “ATM machine,” and “PIN number.”

In a June 22, 2020 letter from EEOC counsel to CRT counsel, the EEOC stated, “the data sought is directly related to the EEOC's claim that Schuster's use of the CRT test has a disparate impact on female applicants and that it is not job related and consistent with business necessity.” (Doc. 37-8.)

The parties appear to use the terms “algorithm” and “formula” interchangeably.

KAIST involved the identical language (i.e., “not reasonably accessible because of undue burden or cost” contained in Fed. R. Civ. P. 26. 2017 WL 9937760, at *2.