The Epiq Angle

Court Allows UPS to Use Statistical Estimates Instead of Raw Data in Discovery Request —At Least for Now

May 5, 2017

UPS’s attempt to use statistical estimates in lieu of producing actual data to meet a discovery request got the green light from a federal appeals court in February. But the court also gave UPS’s opponent the chance to assess the sufficiency of the estimate and ask for the real data at a later time. The opinion highlights the importance of “cooperation and proportionality” in eDiscovery procedures, guided by the Advisory Committee notes to the 2015 amendments to the Federal Rules of Civil Procedure.


In the case, Solo v. United Parcel Services Co., a class of plaintiffs alleged that UPS had overcharged them for shipping packages valued at $300 or more. In response to plaintiffs’ request for information on individual shipments, UPS instead provided an estimate of the number of shipments, using statistical analysis of data from a more recent period (which was still available in live format). According to UPS, it would require an estimated six months and $120,000 just to retrieve the requested data, not including the time and cost of analyzing it after retrieval. UPS also said that the software required to make the data usable does not exist and would have to be developed in order to meet the plaintiffs’ request.

eDiscovery Proportionality and Rule 26(b)(1)

The plaintiffs moved to compel UPS to produce the package-specific data, as originally requested. But the court denied the motion to compel, citing 26(b)(2)(B) of the 2015 amendments, which provides that a party “need not provide discovery of electronically stored information from sources that the party identifies as not reasonably accessible because of undue burden or cost,” and Rule 26(b)(1), which states that discovery should be “proportional to the needs of the case … and whether the burden or expense of the proposed discovery outweighs the benefit.”

In this case, given that there were still certain outstanding questions (such as the applicability of an arbitration clause and various statutes of limitation pertaining to different class members), UPS argued that much of the data it was being asked to produce could turn out to be of little or no value to the case, which would be grossly outweighed by the burden and cost of production. The court agreed.

Meet and Confer in Discovery

However, the court went beyond simply weighing UPS’s potential costs against the plaintiffs’ potential benefits. Rather, it weighed the needs of both parties and emphasized their duty of cooperation, as articulated in the 2015 amendments. To this end, the order was issued without prejudice, leaving open plaintiffs’ ability to seek the package-specific data at a later time. Also, the court directed the parties to “meet, confer, and devise a mutually agreeable methodology for obtaining a sampling of the requested data”—presumably to allow the plaintiffs to be satisfied that the data produced by UPS serves the purposes of the discovery request.

This part of the ruling rests on the idea that parties have a “heightened duty of cooperation in procedural matters such as discovery.”  The Court looked to the Advisory Committee notes Rule 1, which says that litigants “share the responsibility [with the courts] to employ the rules” in a way that helps “secure the just, speedy, and inexpensive determination of every action.”

As the court explained: “Linking the concepts of cooperation and proportionality, the Advisory Committee's Notes to Rule 1 state, ‘Effective advocacy is consistent with—and indeed depends upon-cooperative and proportional use of procedure.’ These principles are particularly important when, as here, the discovery sought comprehends a broad-ranging and massive amount of data.”

Filed under: AI, artificial intelligence, class action, eDisclosure, ediscovery, ESI, predictive coding