PROCESS: Analysis
By Digital Discovery CLE Teaching Fellow Josh Solomon

Like all other components of the discovery process, the way material is actually produced changes as the object of discovery requests become more likely to be found in electronic formats. These changes present new problems, modify old ones, and, for the savvy participant, offer potential advantages.

The requesting party's concern about whether or not it has gotten all to which it is entitled is only exacerbated by the potential enormity of the amount of information that digital media may contain. One can never be sure that something has not been left out. Minimizing the chances of this happening will depend on the particular case and the type of information requested.

A good starting point for protecting oneself is in the framing of the discovery request itself. Given that many courts and opposing counsel may not be familiar with certain aspects of digital information, discovery requests must be carefully worded. While it is clear that digital information is discoverable, it is worth ensuring that discovery requests and orders are precise, and stating explicitly that electronic sources must be searched. Given the ambiguity surrounding these issues, it behooves the requestor to specify exactly what he or she is requesting: does the request ask for embedded data? does it seek residual data? does it explicitly request either the electronic or the paper version of the documents?

Requesters may also want to consider using interrogatories to get familiar with the other sides' systems, followed by expert analysis of the answers provided. By acquainting oneself with an opponent's technology, discovery requests can be formulated to make expectations clearer. Lacouture, Peter V. "Discovery and the use of computer-based information in litigation." Rhode Island Bar Journal 45 (December 1997), 9. It may also be valuable to inquire as to how the search will be conducted-and how it was conducted once the information has been provided. Was the "chain of custody" preserved, or is there a risk that data was lost, corrupted, or changed? Was the original version write-protected or otherwise secured against tampering as it passed through the search and discovery process? Is the copy of the information source complete? For example, a sector-by-sector copy of a system's drive (a.k.a. image copy) will yield more information than a file-by-file backup. Did the responding party search residual and embedded data for the applicable terms? All in all, requesting parties should examine the information they get and the process used with an eye toward determining whether the information obtained is reliable, able to be independently analyzed, and tamper-proof. Feldman, Joan E. and Kohn, Rodger I., "The Essentials of Computer Discovery," Practicing Law Institute, PLI Order No. G0-0051, (June 14-15, 1999), 62-3. Furthermore, since computerized information may be altered quickly and inadvertently, speedy action on the part of the requesting party is important. Johnson-Laird, Andrew. "Smoking guns and spinning disks." Computer Lawyer 11 (August 1994), 1.

The actual search terms used will generally be as important as the formulation of the discovery request and the technology behind the search process. Here the old maxim of "garbage in, garbage out" applies. Many who seek information from the other side will want to know on which terms the adversary built its search. The responding party may resist revealing such information for fear that doing so will expose work-product, or how they were thinking about the case. Some, on the other hand, may see providing notice of search terms in advance as a way to head off future challenges to a search. Thus, if one desires to know the search terms, asking may not always be in vain. There may even be an opportunity for the requester to dictate the terms or to reach agreement on terms. Where such an opportunity exists, the requesting party will have to decide whether to take advantage of it. While some may want to know the terms, they may not want to take responsibility for determining them. The producing party, after all, is likely in the best position to know which terms will produce key information from its own data sources.

The value of other techniques for ensuring adequacy of production, including deposing IT professionals and using neutral experts, will of course depend on such considerations as one's comfort with what one has received and cost-benefit analyses of how much to invest in the discovery phase. In a large, complicated case involving a defendant with a complex set of potential data sources, uncertainty about the defendant's compliance and significant investment in discovery may be warranted.

There will come a point in every case involving digital discovery when someone will have to decide the format (electronic or hard copy) in which information will be provided. That someone could be the judge through an order or, if neither the judge nor the requesting party pays notice, the responder by default. To make an educated choice between receiving the information in printed or electronic form, a party must understand the substantive differences between the two versions. It is an oft-stated (and true) principle that most "deleted" information is not really gone. Such "undeleted deleted" material, called residual data, may contain important information yet may never appear in a printed document. Similarly, printed copies will not show embedded data, the potentially valuable information stored automatically within computer files, which includes such things as time and date stamps, authors, and revisions. See, generally, Sheindlin, Shira A. and Rabkin, Jeffrey, "Electronic Discovery in Federal Civil Litigation: Is Rule 34 up to the Task?" 41 B.C. Law Review 327 (2000). It is not enough to make sure such "invisible" data is searched; the party making the request will also want to ensure that it can see the information. Furthermore, information in electronic format may make it easier to search and organize material once it has been received, particularly when the fruits of discovery are bountiful.

That said, it is not always the case that data in electronic format is more desirable than printed versions of it. Twenty-year-old backup tapes may be more inscrutable and costly to use than printed versions of what they contain. Depending upon the other side's systems, it is possible that a given party will not have the technology or capacity to make use of data in electronic format. In Sattar v. Motorola, 138 F.3d 1164 (7th Cir. 1997), for example, a plaintiff demanded hard copies of over 200,000 e-mails, since it lacked the technology to read the electronic format. This wrinkle demonstrates further the need for ensuring that one understands the opponent's system.

Finally, those who are forced to respond to requests for electronic data may wish to rely solely on electronic searches of that data, to make discovery more efficient and speedy. Their willingness to do so, however, may be conditioned on the requesting party's agreement to waive inadvertent privilege waivers and to allow producers to postpone privilege objections. On the one hand, doing so may sacrifice important rights. On the other, it may speed things up and possibly result in getting information that one wouldn't otherwise get. The final decision on whether or not to grant such waivers if asked will depend on a party's interests in the given case, and will undoubtedly be an issue of trading away one benefit or right for another.

In the end, the problem of knowing whether you got all to which you are entitled will persist in the digital-discovery age. Whether it will do so in a fashion that is more or less troubling will inevitably turn on one's technical savvy and understanding of the new issues raised by digital discovery.

For every decision to be made by and new challenge presented to the requesting party in a digital discovery situation, there are corresponding decisions and challenges facing the party forced to respond. Digitization of information may present greater options in how to respond to a discovery request. Responses can be organized in a manner that is similar to the traditional method, or in one that strays far from it. Will the responding side take advantage of electronic data's amenability to searching and sorting, or will that benefit be foregone by turning digital data into printed sheets of paper? To grasp the vast differences in potential processes, one need only imagine and contrast a warehouse full of paper-to be gone over page by page-with a box of disks to be subjected to a pre-written search string. Assuming data has been "retained" effectively, efficiency and cost considerations will generally weigh in favor of electronic processes-savings in legal fees alone from not having to hire an army of paralegals to cull through boxes of paper can be enormous.

As with the traditional approach, an effective electronic procedure must begin with a good search-term list. This list will serve as the heart of the process. Once the list is developed, the most efficient approach will incorporate it into a sophisticated software program designed to scan and sort the data. A good program will not just consist of a "Find" command, but will also have the capacity to index items, detect privilege, and exclude certain documents-or at least flag them for closer review. Short of a fully automated approach, electronic and traditional methods can be combined. A program may make an initial cut, for example, to be followed by manual reviews of selected documents. While this bifurcated approach may not provide as significant efficiency gains and cost savings, it may nevertheless be substantially better than a purely manual one. If the combined method involves printing information to be reviewed, however, one must be concerned about the other side's (and the judge's) expectations. When files are printed, even if eventually re-scanned, the residual and embedded data are lost, a contingency the discovery request may not permit.

Any diligent method for responding to a digital discovery request will ensure that certain standards are met. These standards will logically align in many cases with those mentioned above-those which the other side will examine when assessing the likelihood that it got what it asked for. Whether the process has the capacity for searching embedded and residual data, for example, will factor into any view of its quality. In general, responding parties will want to make sure their processes are reliable and verifiable if challenged, and that they protect themselves by making tamper-proof copies. Feldman and Kohn, 63.

The automated search and sort procedure will become more involved as a firm's technology increases in variety, age, and complexity. A well-designed program will be able to handle complex and difficult searches and indexing within a fraction of the time that its manual counterpart requires. Nevertheless, expenses will rise with complexity, as will the specter of error. This will lead to greater need for technical skill in designing an electronic process. For many, it may also lead to distrust of the electronic procedure. Will the computer adequately test relevance? Can it protect proprietary information and privileged exchanges? These are valid worries, the magnitudes of which will vary by case. To overcome these concerns, responding parties may seek concessions from those making requests, including rights to postpone privilege objections, and relief from inadvertent waivers. Such concessions could come in exchange for the promise of faster and more accurate responses. Of course, the concessions may not provide a fully supportive safety net for the producer. Even if privilege can still be asserted at trial, for example, one need be concerned about putting such information in the hand of an opponent.

For the same reasons that requesting parties will presumptively want data in its native format, respondents may have an interest in producing hard copies. Information in electronic format has the potential for revealing significantly more to an adversary. In some cases, through the use of embedded information, a sophisticated recipient of information will be able not only to study documents, but also glean document crafters' thought processes. Given that the design and capabilities of a firm's system may in and of itself constitute proprietary information, electronic submissions also run the risk of inadvertently giving away valuable trade secrets. See Sheindlin and Rabkin, 348-9. Despite these interests, convincing rationales for delivering only paper copies may be necessary, as several courts have mandated electronic delivery when such is requested. See Sheindlin and Rabkin, 355-56 and Anti-Monopoly, Inc. v. Hasbro, Inc., 1995 WL 649934. But see also Williams v. Owens-Illinois, Inc., 665 F.2d 918 (9th Cir. 1982) (holding that a party did not have to produce electronic versions of data it already produced in hard-copy format).

Cost will also be an issue in determining the best procedure for responding to a discovery request. While high-quality automated processes have the potential for great cost savings, if electronic production is required, hidden data simply means there is more to search. On balance, producing parties have been relatively unsuccessful at using this as an excuse for not producing data in its native format. They have also found that shifting of costs is generally no easier than with traditional discovery. A court is likely to feel that the choice to digitize was one that the party made, and as such one for which it should bear the costs. As one writer on the subject pointed out, "Courts appear to have been reluctant to force requesting parties to bear the costs of gathering and producing in usable form electronic evidence responsive to a Rule 34 document request. . . . The guiding principle in such cases appears to be the concern that technological advancements should not alter the framework of civil litigation by shifting costs of discovery." Sheindlin and Rabkin, 360. Moreover, whereas with paper files respondents have been able to reduce costs in certain situations by making information "open for inspection" rather than copying and delivering it, such an approach to cost-shifting may not always be available for digital data. Inspection of this sort may expose one's entire system to risk of intrusion. Taking into account these potential costs, it becomes clear that the worst-case scenario may be one in which a requirement of electronic submission combines with a poorly designed search and sort procedure that fails to allow one to recognize the cost benefits of electronic data.

Finally, just as a requesting party may find hard copies more valuable to it on occasion, a respondent may find electronic submission more desirable. This is most likely to occur when printing, sorting, copying, and disruption costs outweigh the benefits of keeping hidden information under wraps.

In the end, it seems evident that both the ability and the burden of responding to a digital discovery request will significantly alter response procedures. Those with the means to do so will be well served-from financial and efficacy standpoints-to make use of advanced search and indexing programs. As the judiciary does the same in trying to set guidelines, respondents will also need to puzzle through important new decisions about costs and formats in the digital context.