Dealing with Discoverable Documents in the Digital Age

Bruce W. Felmly
Director, Litigation Department
Published: New Hampshire Bar News
December 17, 2014
News - Featured Image - Trial-Practice-and-Litigation

Tips and traps for document-heavy cases and some thoughts on the future of predictive coding.

It was not that long ago that commercial litigation discovery involved assembling scores, hundreds, or even thousands of banker’s boxes of hard copy documents into a warehouse where platoons of paralegals or lawyers would physically and painstakingly review the documents, making notes or abstracts for possible use of the nuggets supposedly found in this treasure trove.
Today, physical inspection of hard copy documents is reserved for relatively document-thin cases, and even then the actual production likely arrives on a disk. You will typically receive a series of disks containing downloaded documents, generally in a word-searchable format, and hopefully in a “load file” that can be used with your firm’s discovery search engine. Production of “native format” documents can involve significant potential risk and is reserved for unique discovery requirements.
The following steps or considerations are important for conducting the search or inspection of the document universe accurately and effectively.
Searching
Develop a search plan for the universe of documents that will enable you to identify those with the potential to lead to the discovery of admissible evidence. Presumably the search will be defined initially by your opponent’s request for production of documents. It is helpful, indeed required by many rules of civil procedure, to meet and confer, and stipulate on a plan for actual searching of the documents. Can you agree on a list of 25–75 search terms, including project-specific titles, keywords in the dispute, names of key participants, and related tags that will likely assure substantial identification of relevant documents?
Privilege
Protect attorney-client privilege. Review specific searches against the universe to cull out attorney-client privilege or intellectual property. Be sensitive to persons working collaterally with in-house lawyers (paralegals, administrative staff). You’ll need to determine from in-house counsel or the client the reality of “privileged communications.” Were these communications undertaken for legal advice, and were they kept truly confidential? You will need to screen the production for attorney-client/work product, most likely using search term software. Since that is not foolproof, evaluate the use of a stipulation/claw-back agreement with your opponent to protect against waiver of privilege by means of inadvertent disclosure.
Discovery Volume
Assess the volume and burden to actually review the documents produced by the computer search of the universe. Reassess if it is unrealistic, mistaken, or overbroad (e.g., 155,000 documents are pulled, referencing “the project” – mostly not dealing with our project).
If reduction in the scope of the production cannot be achieved, it will generally require a negotiation with the requesting counsel, or possible motion practice to shift the costs of such overbroad discovery. Court rules and case law permit such cost-shifting.
Determine how the review of the searchable documents will result in efficient abstracting, creation of witness-specific groupings of documents for later use, preparing client privilege logs, supporting motions in limine for various evidentiary issues, and other specific purposes. Get maximum bang for your buck.
Predictive Coding
In the same way that word searching has largely replaced individually inspecting the hard copy of every document in the production universe, it is likely that “predictive coding” will replace much of the word searching currently done in major cases.
In its simplest sense, predictive coding is a search methodology where the computer model is presented or coded with a selection of responsive or representative documents and then searches the universe to pull out or identify similar or related documents. This is done based on complicated algorithms and relationships, and of course, the key is to familiarize the computer with an accurate sample of representative documents.
Many lawyers are beginning to use this process and it is gaining significant traction in the courts. If you are planning to use this method, it will require opposing parties and the court to agree to its employment, a process that will be enhanced by the increasingly available studies that describe a much higher accuracy rate for predictive coding in comparison to either hard copy physical inspection or prior word searching approaches. It will clearly be a significant tool and become commonplace within relatively few years.