Opinion
CIV. NO. 3:07CV929 (WWE)
11-10-2014
A&R BODY SPECIALTY AND COLLISION WORKS, INC., FAMILY GARAGE, INC. and THE AUTO BODY ASSOCIATION OF CONNECTICUT on Behalf of Themselves and all Others Similarly Situated, v. PROGRESSIVE CASUALTY INSURANCE COMPANY and PROGRESSIVE DIRECT INSURANCE COMPANY.
RULING ON PLAINTIFFS' MOTION FOR RECONSIDERATION [DOC. #476]
Plaintiffs, A&R Body Specialty, Family Garage and the Auto Body Association of Connecticut, on behalf of themselves and all others similarly situated ("plaintiffs") seek reconsideration of the Court's September 9, 2014 discovery ruling pertaining to the production of estimating data. [Doc. #476]. Defendants oppose this motion. [Doc. #486]. For the reasons articulated below, the motion for reconsideration [Doc. #4476] is GRANTED, and the Court ADHERES to its previous ruling.
I. BACKGROUND
The issue of estimating data first arose in April 2013, when the Court held a telephone status conference concerning the extraction of estimating data from Progressive's third-party provider, Mitchell. [Doc. #283]. The Court ordered that the parties split the costs of extraction. [Id.]. Mitchell ultimately produced 157 "columns" of data, only some of which contained body shop-specific information. Progressive then produced a second set of data with more body shop-specific information. After comparing the two sets of data, plaintiffs found that Progressive used different field identifiers than Mitchell.
At the request of the parties, on June 2, 2014, the Court held a discovery conference where, among other issues listed on the joint agenda for discussion, the parties raised "Plaintiffs' request for Defendants to produce merged Mitchell and Progressive data." See Appendix "A" (Agenda for June 2, 2014 Status Conference with Magistrate Judge Fitzsimmons, at No. 10, sent to the Court by Attorney Robles via email dated 5/27/2014 at 4:08 P.M. E.S.T.). Addressing this issue at the June 2, 2014 conference, plaintiffs asserted that it would be too labor intensive for them to merge the data, and would further expose plaintiffs to potential challenges regarding the data's integrity. Plaintiffs further argued that they need "pristine" data for purposes of expert damage analysis. In response, defendants stated that they have no way to merge the data, and are uncomfortable manipulating a third-party's data set. On June 20, 2014, the Court issued a ruling addressing the matters raised at the discovery conference, including the issue of the merged data. [Doc. #472, 8-9]. The Court ordered that the parties meet and confer to discuss resolving this issue through the use of requests for admissions and/or other means, and to report back via letter brief. [Id. at 8].
Defendants submitted a letter brief dated July 7, 2014, "to address Plaintiffs' request that Progressive merge two separate data compilations from two distinct data sources, one from a non-party and the other Progressive, to enable Plaintiffs' experts to have a single 'pristine' data set to use in this case." See Doc. #478-5, at Ex. T, July 7, 2014 Ltr., p. 1. Plaintiffs responded via letter brief dated July 21, 2014 and requested that defendants be compelled to produce a "complete and accurate compilation of its estimating data without further delay." See id. at Ex. U, Pls. July 21, 2014 Ltr., pp. 1-4. Notably, plaintiffs failed to present any case law in support of their position.
The motion for reconsideration "provide[s] a more thorough history of the issue than was possible to set forth in the constraints of their letter brief." [Doc. #477, 2]. The Court notes that no page limits were imposed for the letter briefs. If plaintiffs believed that more information was relevant to the Court's consideration of this issue, they should have provided it in the first instance. If plaintiffs thought they would surpass the letter brief's "constraints," they should have sought leave of Court to exceed those "constraints."
On September 9, 2014, the Court issued a ruling disposing of issues raised in the parties' July 7 and 21, 2014 letter briefs, among others. [Doc. #472]. With respect to the "merging" of estimating data, the Court denied plaintiffs' request for a merged data set in light of the fact that neither the Rules of Civil Procedure, nor case law, supported plaintiffs' position. Plaintiffs now seek reconsideration of the portion of the ruling pertaining to "Plaintiffs' request to discover appraisal data for the automobile physical damage claims paid by [defendants] during the relevant period." [Doc. #476]. Specifically, Plaintiffs request the Court to "supplement its ruling to address Plaintiffs' rights to obtain a complete and accurate production of Progressive's data[...]" [Id.].
II. LEGAL STANDARD
The Second Circuit has held that "[t]he standard for granting [a motion for reconsideration] is strict, and reconsideration will generally be denied unless the moving party can point to controlling decisions or data that the court overlooked - matters, in other words, that might reasonably be expected to alter the conclusion reached by the court." Shrader v. CSX Transp., Inc., 70 F.3d 255, 257 (2d Cir. 1995) (citations omitted). "There are three grounds that justify granting a motion for reconsideration: (1) an intervening change in controlling law; (2) the availability of newly discovered evidence; and (3) the need to correct clear error or prevent manifest injustice." Whitserve, LLC v. GoDaddy.com , Inc., 3:11-CV-948 JCH, 2013 WL 1442449, at *1 (D. Conn. Apr. 9, 2013) (citing Virgin Atl. Airways, Ltd. v. Nat'l Mediation Bd., 956 F.2d 1245, 1255 (2d Cir. 1992)). "That the court overlooked controlling law or material facts may also entitle a party to succeed on a motion to reconsider." Whitserve, 2013 WL 1442449, at *1 (citing Eisemann v. Greene, 204 F.3d 393, 395 n. 2 (2d Cir. 2000)). However, it is not "appropriate to use a motion to reconsider solely to re-litigate an issue already decided." Conn. Com'r of Labor v. Chubb Grp. of Ins. Companies, 3:11CV00997 AWT, 2013 WL 836633, at *1 (D. Conn. Mar. 6, 2013) (quoting SPGGC, Inc. v. Blumenthal, 408 F. Supp. 2d 87, 91 (D. Conn. 2 006)); see also Lego A/S v. Best-Lock Const. Toys, Inc., No. 3:11cv1586 CSH, 2013 WL 1611462, at *2 (D. Conn. Apr. 15, 2013) (citation omitted) ("A motion for reconsideration is not simply a second bite at the apple for a party dissatisfied with a court's ruling...").
III. DISCUSSION
Plaintiffs do not challenge the Court's ruling denying their request for a merged data set. Rather, plaintiffs submit reconsideration is warranted for two reasons: "First, likely due to the constraints of the letter briefing format, the Court's ruling did not address the core of this crucial discovery dispute [namely, Plaintiffs' rights to obtain a complete and accurate production of Progressive's data]. Second, if Defendants are not required to cure the deficiencies in question, Plaintiffs will suffer manifest injustice." [Doc. #477, 10 (brackets added)]. Defendants oppose plaintiffs' motion and submit that it is an improper attempt to re-litigate an issue already properly decided by the Court.
A. Complete and Accurate Production of Progressive's Data
Plaintiffs submit that the Court erroneously focused on a single remedy proposed by plaintiffs and did not address the "core" of the discovery dispute, and again submit that defendants should be compelled to produce "a complete and accurate compilation of the required data both to fulfill their obligations and to prevent prejudice to plaintiffs." [Doc. #447, 12].
As an initial matter, the Court did not solely consider the production of a merged data set. On the contrary, the Court carefully reviewed plaintiffs' letter brief and explicitly noted their arguments that, "they have yet to receive a 'complete and accurate compilation of Progressive's estimating data,' and that 'unless the deficiencies are cured, the jury will be precluded from considering the full merits of Plaintiffs' case, which would result in considerable prejudice." [Doc. #472, 7]. Indeed, plaintiffs' letter brief mentions a "complete and accurate compilation of estimating data" no less than five (5) times. Moreover, during the June 2 discovery conference, plaintiffs not only requested that defendants produce a merged data set, but also asserted their need for a "pristine" data set for purposes of expert damage analysis. See, e.g., Doc. #463, 8.
To the extent that the Court's ruling "focused" on a "merged data set", this is the language that the parties had primarily used up until plaintiffs' July 21, 2014 letter brief. For example, the joint agenda for the June 2 discovery conference explicitly referenced, "Plaintiffs' request for Defendants to produce merged Mitchell and Progressive data;" plaintiffs' counsel argued his clients needed a "pristine" data set and proposed defendants produce a merged data set; and defendants' letter brief also referenced a "merged data set." To the extent that plaintiffs argue that a "merged data set" and a "compilation of estimating data" are materially different, this is a matter of semantics and an unconvincing argument in favor of reconsideration. Regardless of the label placed on plaintiffs' requested relief, the Court's ruling does not change. Whether plaintiffs seek a "data compilation" or "merged data set", the case law is clear that Rule 34 cannot be used to compel a party to produce a document that does not exist. See, e.g., Hallmark v. Cohen & Slamowitz, Midland Funding LLC, No. 11-CV-842S(F) 2014 WL 5017859, at *5 (W.D.N.Y. Oct. 8, 2014) (citation and internal quotation marks omitted) ("It is basic that in responding to a document production request, pursuant to Fed. R. Civ. P. 34(a), a party is not required to create documents meeting the document requests, only to produce documents already in existence."). For reasons stated in the Court's prior ruling, plaintiffs are essentially asking defendants to do just that. Again, as in the plaintiffs' July 21 letter brief, plaintiffs have failed to present any case law supporting this position.
Besides defendants' suggesting the use of requests for admissions to provide the information plaintiffs seek, the only remedy proposed by plaintiffs was the production of a "merged" data set.
Merriam-Webster defines merge as, "to cause (two or more things, []) to come together and become one thing: to join or unite one thing with another." http://www.merriam-webster.com/dictionary/merge (date last visited: November 3, 2014). Similarly, a "compilation" is defined as, "a group of things (such as songs or pieces of writing) that have been gathered into a collection." http://www.merriam-webster.com/dictionary/compilation?show=0&t=1415046662 (date last visited: November 3, 2014). To the extent plaintiffs use the term "data compilation" as that phrase is used in Federal Rule of Civil Procedure 34, it is clear to the Court that the relief plaintiffs' seek does not encompass a "data compilation" as is kept in the usual course.
As to completeness, plaintiffs take issue with the Court's statement that, "Defendants have already produced the data sought." [Doc. #477, 10]. Plaintiffs claim this is not correct as "[d]efendants still have not identified the requested repair channel data for 11,269 claims. Moreover, contrary to the Court's characterization, Defendants did not produce the bulk of the appraisal data - Mitchell did, only after Defendants insisted Mitchell was the best place to obtain it and the Court so ordered." [Id.]. Although the Court could have been clearer in its statement that defendants and Mitchell have already produced the data sought, the fact does not change that plaintiffs have the "bulk" of the estimating data sought. Plaintiffs, however, again argue that the Mitchell production was incomplete with respect to repair shop data. Plaintiffs also assert that defendants admitted their supplemental production was incomplete with respect to 11,269 claims and have further raised doubts about the accuracy of their data by virtue of the statement that, "there are records where the Mitchell data contains one shop name and Progressive's data another."
See Doc. #478-5, at Ex. U, Pls. July 21, 2014 Ltr., p. 2 ("Upon receiving the Mitchell production, Plaintiffs noticed that the name of the repair shop was missing from many entries and asked Progressive to fill in the gaps. Specifically, Plaintiffs asked Progressive to identify the name of the repair shop, whether the shop was in Progressive's Network and whether the repair was ultimately performed. Progressive supplied some of this missing data in a separate production but reported that 11,269 claim numbers were 'not found' in its database. It speculated that these might have been training files, not actual claims. Plaintiffs disproved this theory by cross-referencing one of the "not found' numbers with the Mitchell production. When confronted with this example, Progressive acknowledged that there might be discrepancies between its production and the Mitchell and refused to confirm the accuracy of either production.").
--------
Defendants maintain the position that, "there is no set format for the repair shop data to be entered or even a requirement that the repair shop be identified at all." [Doc. #486, 4]. Defendants further submit that,
The only way to determine with certainty what repair shop completed each customer's repair is to conduct a file-by-file review of each claim. This review would involve an examination of tens of thousands of claims and the creation of a new data set based on that review. Furthermore, even a file-by-file review would still not provide Plaintiffs with the "complete and accurate" data they request.[Id. at 5]. Defendants also noted in their July 7 letter brief that the two data sets have a common field identifier, namely the claim number, and that plaintiffs are "free to compare the data in the repair shop fields based on claim number." See Doc. #478-5, at Ex. T, Defs. July 7, 2014 Ltr., p. 3. Aside from the burden of undertaking such a review, plaintiffs are concerned that this sets the stage for defendants to attack the integrity of the data, and consequently to undermine or seek preclusion of plaintiffs' expert analysis.
Again, the relief plaintiffs seek implicates the creation of a document for the purposes of production. Plaintiffs previously demanded from defendants, and ostensibly seek here, that defendants update the Mitchell's file with the additional shop specific information. Plaintiffs do not just seek the production of the repair shop names, they seek a data set in which the missing repair shop names are correlated to claim numbers. See, e.g., Doc. #477, 17 ("[Defendants] should be compelled to rectify any incompleteness or inaccuracy in their data by directly producing a pristine set. To the extent that compiling the data and certifying its accuracy might be burdensome on Defendants, they are hoist with their own petard."). As the Court previously found, plaintiffs are not entitled to receive ESI in their preferred format. Nor are defendants required to create a document for production. Plaintiffs also apparently have the ability, albeit with much effort, to ascertain the shop data by searching a claim number listed as "not found in the database" against the original Mitchell's estimate level database, without using the three digit extension used by Progressive. [Doc. #477, 8].
Although not cited by either party, the Court finds the Northern District of California case, Apple Inc. v. Samsung Electronics Co. Ltd., Case No.: 12-CV-0630-LHK(PSG), 2013 WL 4426512 (N.D. Cal. Aug. 14, 2013), both instructive and distinguishable from the present facts. There, defendant Samsung moved to compel Apple to produce further responses to its requests for production seeking financial documents regarding: "(1) units sold, gross and net revenue, gross and net margin, and gross net profits [for certain Apple products]; reports and projections of U.S. sales, profitability margins, and financial performance for each version of the iPhone and iPad[]; and all costs comprising costs of goods sold and all costs other than standard costs for each of the accused products." Id. at *1. Apple produced documents in response, to which Samsung objected, claiming the production was deficient because most of the data presented was worldwide rather than U.S. specific and the data was compiled at the "product-line level" rather than the "model level." Id. at *2. Apple refused to produce the financial data requested, "claiming it does not maintain reports of such data in its 'ordinary course of business.'" Id. As to the relevancy of the data sought, the Court noted that,
Apple does not dispute that the data is relevant under the generous Rule 26 standard. Indeed, the U.S.-specific, model level data is highly relevant to both Samsung's own damages claims as well as Samsung's defenses against Apple's damages claim[...]Samsung also requires data specific to each accused product; generalized data spread across several different models will not suffice[...] For similar reasons, the data sought by Samsung are also relevant to rebut Apple's damages claim[.]Id. (footnotes containing citations to the record omitted). From these arguments, the Court found that, "Apple does have financial databases that it could query to generate at least some of the reports sought by Samsung." Id. at *3 (referring to footnote 10, which cites a docket entry reflecting Apple's Rule 30(b)(6) witness who, "provided a roadmap for querying the database to generate the financial figures."). The Court further went on to find that,
It is therefore uncontroversial that if Apple had reports of the data at issue, or could generate such reports with only reasonable efforts, Apple would have to produce them. But Apple insists that it does not have reports of the nature Samsung would like and only a herculean effort could produce even a subset of the reports demanded. Samsung presents evidence insinuating otherwise, submitting Apple documents showing that it does report at least some of the data at issue that is specific to the U.S. market and to individual product models. Apple also admits that it maintains "systems billings" and other "revenue line items" on a model-specific basis, but argue these are imperfect representations of revenue.
While this Court has held a party should not [be] required to create completely new documents, that is not the same as requiring a party to query an existing dynamic database for relevant information. Courts regularly require parties to produce reports from dynamic databases, holding that "the technical burden... of creating a new dataset for the instant litigation does not excuse production. Compelling production here would therefore not violate any established discovery principles.Apple, 2013 WL 4426512, at *3 (internal footnotes containing citations omitted). Despite this finding, the Court ultimately did not require Apple to produce the documents sought because the burden or expense of the proposed discovery outweighed its likely benefit in light of the fact that the parties had already submitted their expert damages report, and the financial documents would therefore be of limited value to Samsung. Id. The Court further noted that, "Although counsel was not able to shed light on exactly what was done, Samsung's experts were clearly somehow able to apportion the worldwide, product line inclusive data to estimate U.S. and product-specific damages." Id. Accordingly, the Court did not require the production of data that Samsung "is able to do without." Id.
Finally, the Court also took steps to protect Samsung from any undue prejudice arising from Apple's reporting limitations:
To be sure, Samsung's damages experts are still open to attack by Apple for their failure to use more granular financial data, either at pretrial hearings or at the trial itself. While Apple clearly could not impeach Samsung with any newly-compiled financial data that was not produced, in fairness it also is precluded more broadly from challenging Samsung's damages experts for failing to allocate geographically or by product model in any way that could have been supported by the reports disputed here that were requested but not produced.Id.; see also id. at 4 ("Apple need not produce the financial documents sought by Samsung. But Apple is estopped from challenging Samsung's experts on any ground that would be rebutted by reference to documents that Samsung requested but did not receive.").
Here, similar to Apple, defendants do not contest the relevancy of the data sought. Indeed, such data is not only relevant to plaintiffs' claims of steering but also to defendants' defenses against these claims. Accordingly, as in Apple, if defendants (or Mitchell) had the data plaintiffs seek, or could produce the "data compilation" with only reasonable efforts, then defendants would have to produce the information sought. However, where the similarities end is that Apple had databases it could query to generate the reports sought by Samsung. Here by contrast, to produce the data compilation plaintiffs seek, defendants would have to conduct "a file-by-file review of claims information beyond the data itself in order to verify whether each customer's vehicle was repaired and, if so, what body shop repaired the vehicle, and could require Progressive to reach out to customers." [Doc. #486, 7]. This starkly contrasts to a technological giant, such as Apple, whose Rule 30(b)(6) witness explained the "roadmap" for querying the information sought from an existing database. Here, there is no evidence of record that Progressive could "query" the information sought through the use of data extraction programs or otherwise. Indeed, defendants' representations are to the contrary; namely, that they would have to manually create the data sought.
The Court further notes the cases upon which the Apple court relies in support of compelling the production of Apple's data are also further distinguishable from the present facts. For example, the case of Gonzales v. Google, Inc., 234 F.R.D. 674, 683 (N.D. Cal. 2006), cited in footnote 12 of the Apple ruling, addressed Google's argument that it faced an undue burden responding to a subpoena request because "it does not maintain search query or URL information in the ordinary course of business in the format requested by the Government." Gonzales, 234 F.R.D. at 683. The Court noted that,
[A]s a general rule, non-parties are not required to create documents that do not exist, simply for the purposes of discovery. In this case, however, Google has not represented that it is unable to extract the information requested from its existing systems. Google contends that it must create new code to format and extract query and URL data from many computer banks, in total requiring up to eight full time days of engineering time. Because the Government has agreed to compensate Google for the reasonable costs of production, and given the extremely scaled-down scope of the subpoena as modified, the Court does not find that the technical burden of production excuses Google from complying with the subpoena.Id . (internal citation omitted). Again here, there is no indication that defendants can extract the information requested by means of a computer code, or otherwise "compile" the data sought using available technology. The fact that the government offered to pay for such extraction further distinguishes Gonzales from the present facts.
Accordingly, the Court will not require defendants to produce the information plaintiffs seek. However, because it appears that it would be in both parties' interests to have the data plaintiffs seek and for both parties' experts to work from the same data set, the Court suggests that the parties cooperate in hiring a neutral third party to conduct the comparison defendants state plaintiffs are free to perform. This would provide plaintiffs, to some extent, the data they seek, while at the same time preventing an attack on the data's integrity.
To the extent that the parties do not wish to proceed as the Court suggests above, and plaintiffs undertake their own comparison of the repair shop fields, defendants are cautioned that they will not be able to impeach plaintiffs with any newly-compiled data that is not produced. In that regard, fairness may also dictate a broader preclusion of defendants' challenging plaintiffs' experts for relying on data which defendants produced and encouraged plaintiffs to manipulate.
Finally, with respect to the fact that, "there are records where the Mitchell data contains one shop name and Progressive's data another," there is no evidence before the Court as to how many claims contain such conflicting data. When this information does become available, either by virtue of a third party's comparison of the data or plaintiffs' comparison of the data, the Court will not hesitate to permit plaintiffs, within reason, to propound narrowly tailored discovery requests, such as requests for admission, to ascertain which shop name is correct.
B. Manifest Prejudice
Finally, plaintiffs argue that they will face manifest prejudice without obtaining the relief requested in the motion for reconsideration. The prejudice faced by plaintiffs without the information sought has already been brought to the Court's attention in plaintiffs' July 21 letter brief. Plaintiffs have not presented any new facts or case law that was unavailable at that time. The Court further credits defendants' argument that they and plaintiffs will be working from the same set of data as a basis for their experts' analysis, among other issues. The Court's directive above and any appropriate preclusion orders should also ameliorate any prejudice faced by plaintiffs.
IV. CONCLUSION
Accordingly, the motion for reconsideration [Doc. #476] is GRANTED, and the Court ADHERES to its previous ruling, as further articulated above.
This is not a Recommended Ruling. This is a discovery ruling or order which is reviewable pursuant to the "clearly erroneous" statutory standard of review. 28 U.S.C. § 636(b)(1)(A); Fed. R. Civ. P. 72(a); and D. Conn. L. Civ. R. 72.2. As such, it is an order of the Court unless reversed or modified by the district judge upon motion timely made.
ENTERED at Bridgeport this 10th day of November, 2014.
/s/_________
HOLLY B. FITZSIMMONS
UNITED STATES MAGISTRATE JUDGE