The progression of electronic discovery has triggered corporations, law firms and vendors to adapt to -- some would say in order to survive -- modern litigation and the procedures and processes relating to electronically stored information.
Corporations have implemented e-discovery response teams, installed software and appliances to automate legal holds, and revised electronic retention policies to better prepare for the management of e-discovery. Law firms have created e-discovery practice groups, recruited IT specialists, fashioned applications and created review teams to better manage risks and costs associated with e-discovery. Vendors have advanced their operations to include the expertise -- including forensic collection, tape restoration and data processing -- to address corporate and law firm ESI needs.
After identification of both key custodians and related sources of electronic data -- and the preservation and collection of the same -- the law firm and its client must confront the costly tasks of processing and reviewing the data for responding to the current investigation, claim or litigation. The below phases describe the challenges and potential solutions that corporations and law firms need to address for managing risk, balancing costs and leveraging resources during the process and review stages of e-discovery.
DETERMINING WHAT TO PROCESS
Not all media identified and preserved may need to be processed. The type of case, sources of data and number of custodians will impact how best to determine what needs to be processed. Of course, how the organization maintains its data will also impact the processing stage. Early case assessment by the e-discovery team will help determine the scope of electronic data. The need for a "roadmap" of the organization's network infrastructure is critical in not only assisting counsel for the Rule 26(f) meet and confer, but also for determining types and sources of potential relevant information.
Reducing the number of pertinent custodians -- and the need to process and review data from their workstations, laptops, network directories and e-mail accounts -- will significantly impact processing and attorney review costs. However, such decisions should not be made unilaterally but in agreement with opposing counsel or by court consent.
Sampling of client's data will also enable you to limit the amount of gigabytes to process and thus review. You can eliminate entire directories of archived data or nonresponsive system files, for example, that exist on laptops, workstations or servers by reviewing a directory report that contains information on metadata fields such as modified date and file type. This sampling also provides insight into file sizes and the overall volume of potential ESI.
Determining what to process is a balancing act of costs and risks. Negotiating with law firms and vendors can reduce the review rates or per-gigabyte charges. But such reductions have a threshold. Therefore the goal is to limit the amount of ESI to process and ultimately review. The risks of reducing the amount of ESI to process include excluding potential key information that is beneficial to your case strategy and breaching discovery obligations resulting in sanctions, an adverse inference or worse -- even absent bad faith or gross negligence.
FILTERING OF IDENTIFIED ESI
Relevant information may have already been filtered during the identification and collection stages insomuch that date restrictors and key search terms were utilized to extract the relevant data from the ESI sources. Or the ESI may be entire laptops, network directories, workstations or other media. Regardless, the processing of ESI, either by third-party vendors or in-house technology staff, includes indexing of files, applying date restrictors, de-duplication, search term filtering, concept filtering and exporting the resulting population in an appropriate review database.
The ESI may be exported natively for review or rendered to image, text or HTML. Native review is the most beneficial as metadata can provide additional information regarding the document. Furthermore, expending resources and money to image files that most likely are not relevant is wasteful. Review format is typically guided by the type of review tool to be utilized by the review team or may be determined based on the agreed upon form of production under FRCP Rule 34(b)(ii).
The use of date restrictors and search terms should be discussed internally with the e-discovery team to guarantee the best possible results. Most resulting data sets will have "false positives," but by crafting queries based on the facts and issues of the case and by utilizing appropriate search methodologies, including Boolean operators, the false hits will be minimized. You should come to an agreement with opposing counsel regarding the search protocol in order to avoid discovery disputes.
The e-discovery team must address if the de-duplication is to be performed within custodian and ESI source or across all custodians and ESI sources. Typically the resulting de-duped data within custodian or source will be dropped from the export data. Conversely, resulting data that has been de-duped across custodians and ESI sources is tagged as duplicate and any responsiveness or substantive designation is applied to all records within the review database, assuring consistency among the review team.
The e-discovery team should communicate the desired specifications for de-duplication with vendors or in-house technology staff to guarantee there is no inadvertent omitted data and that the electronic files and corresponding attachments (the parent-child relationship) remain intact during the entirety of the processing, review and production stages.
The use of concept filters or content analysis differs from search term filtering because they focus on document and term "clusters," among other criteria. Although these filtering techniques are sometimes used in processing the raw ESI, more commonly they are used to group or folder documents already filtered within the review database as a way to analyze and develop case strategy.
REVIEW OF PROCESSED ESI
The review platform should be determined early in the litigation and consideration should be given to a review tool that fits the needs of the review team, provides for customization, allows for multiple access points and is cost-effective. The review database may be an in-house application or a third-party-hosted solution. Whatever database is designated, make sure the review team is trained on the application and is knowledgeable with regard to the functionalities.
Review teams are typically designated once the data analysis is completed but prior to or during the processing stage. The teams can consist of in-house attorneys, outside counsel or contract attorneys. The amount of resulting data that is filtered is typically less than one-fourth of the overall raw ESI processed. The costs for attorney review of the resulting data can be double or triple the costs of the processing fees depending on time constrains and number of available resources to dedicate to the review.
Clearly the team should be informed of the pertinent issues, key events, witnesses, strategy and the potential privileges that may arise. The review may consist of a responsiveness "first pass" and a substantive "second pass." In a perfect discovery world, only one pass would be necessary. However, condensed discovery deadlines and available resources are critical factors. Tagging documents for first-pass responsiveness should include database tags for non-relevant, responsive, privilege, redact, corrupt/password protected, hot and further review.
Conducting search queries and categorizing results within the database may be beneficial to quickly organize the documents, but it is not meant to be the end all, especially when using such methodologies to create the production set. Recently, in Victor Stanley Inc. v. Creative Pipe Inc., Judge Paul Grimm issued a discovery order holding, among other things, that privilege was waived where the producing party failed to seek a search protocol for privilege files, failed to ensure reasonable search protocols for privilege and failed to secure a "clawback" agreement.
There are a few other practice points to consider when performing the review of processed data. Care should be taken so as to not break the relationship between documents and attachments during the characterization of records. Most databases provide a mechanism to ensure that such separation does not occur. Duplicate records should be linked in order to make certain that tagging is consistent among the reviewers. Maintain native file integrity regardless of the production format. Confirm that there are no conflicts in tagging before allowing the vendor or in-house support staff to prepare the production. This is especially true in coping with the production of redacted images.
PRODUCTION OF ESI
With the responsiveness review complete, attention turns to the actual production of electronically stored information. If form of production was not requested, FRCP Rule 34(b)(ii) permits the responding party to produce in a form or forms in which the ESI is ordinarily maintained or in a reasonably usable form or forms. If producing in image format, in almost all instances Bates numbers should be branded on the images with any confidential or other restricted designation.
It would also benefit the parties to provide a "sampling" of the images and load files with document breaks to assert that the files are compatible with the respected parties' image viewing software. It may be necessary to supplement some native files that were not printable and thus impossible to render to image.
If producing natively, the procedure for producing "parent" documents without their "child" attachments, or vice versa, must be addressed since any modification, especially with regard to e-mails, would alter the integrity of the metadata. For instance, in the event that a privilege claim applies to the attachment(s) but not the e-mail, the most popular solution is to render the entirety of the document to image for production, less the privileged attachment(s).
With native files, Bates numbering cannot be accomplished on the page level without altering the contents of the file, although the MD5 hash value can be used as a document identifier. Another solution is to rename the file by applying Bates number prefixes, but there is vulnerability in possibly altering the metadata including the MD5 hash value.
There are pros and cons concerning form of production -- yet another test of balancing risks and costs in the e-discovery process. Some organizations only want to produce in native form because of the cost involved in an image production. Others prefer not to produce in native form except when specifically requested because they do not want the opposing party to have the metadata information.
The growing nature of e-discovery continues to plague organizations and their counsel. Escalating costs and increasing volumes of ESI, in orders of magnitude, are creating new challenges for those organizations and the law firms that represent them. New techniques for processing and reviewing data are constantly evolving to mitigate risks, lower costs and improve efficiency.
Christopher Starr is the director of practice technology at Wolf Block and provides assessments and solutions to the firm's practice groups and external clients on practice-related applications, e-discovery issues and technical services.