Multiple trends in e-discovery intensify use of technology
The market for e-discovery technology and services will grow at a rate of approximately 12 percent per year from 2013 to 2018, according to Gartner. Among the drivers for that growth are a greater volume of data and an increasing number of data types. Several changes in rules for discovery, including new rules for proportionality and shortened timeframes for some actions, are likely to provide further support for fast growth of e-discovery solutions.
A Supreme Court ruling on April 29, 2015, adopted amendments to the Federal Rules of Civil Procedure (FRCP) to require proportionality in the discovery responses. This requirement means that discovery should take into account the importance of the issues, the amount in dispute, access to relevant information and available resources, as well as the likely benefits as compared to the expense of the proposed discovery. In addition, a note to the amendment says that parties should consider “reliable technology” as a means of reducing costs when large amounts of electronic information are involved.
“Rules for proportionality have existed for some time, but the new rules put a stronger spotlight on this issue,” says Hal Marcus, discovery counsel and director of product marketing at Recommind. “We think these rules will lay the foundation for a broader adoption of analytics technology.” Recommind’s Axcelerate software has a broad range of capabilities that include legal hold, early case assessment, review, analysis and production.
Through a combination of automated review (“predictive coding”) and sampling, it is possible to demonstrate after a portion of a data set has been reviewed that only a limited return would come from more analysis. One approach is to provide input that the software uses to categorize documents as responsive, non-responsive or unknown. After human review of some portion of the categorized documents, any revisions to the categories are input to refine the training. “The training stops when the process is ‘stabilized’ in terms of having that set of data classified to a pre-agreed level of accuracy, and the remainder of documents are not reviewed,” Marcus adds.
The other approach to predictive coding is to use a flexible learning process, searching the documents for the most relevant ones and then using those to train the system to find “more like this.” The suggested new documents are reviewed and used to train the system on an ongoing basis. Eventually no new relevant documents are suggested by the system. “If, say, 20 percent of the total data set has been reviewed and the sampling shows that most of the relevant material has been found, then a further manual check by humans may not be required and the other 80 percent of the material may not be reviewed,” Marcus explains.
The impact of the changes in FRCP rules will take some time to be felt. “There is some change but not a seismic shift,” Marcus says. Judges typically expect that each party shares responsibility to provide transparency and to cooperate with other parties. However, some companies are concerned about excessive transparency. They do not want to reveal any more than they have to.
“Back in the time of paper discovery, some document sets got a quick look and others got a detailed review, but there was not generally a requirement to explain the logic that directed that process,” Marcus says. “In the case of e-discovery, companies have sometimes had to share many details when they wanted to use predictive coding, including what search terms, review workflow steps and training sets they used.” Each shared piece of information about review strategy may provide an opening for the other party to challenge. “Although virtually all review is done on electronic versions of documents,” Marcus says, “a relatively small number of organizations have adopted predictive coding as a standard practice, so this area is not yet mature.”
Proliferating data sources
The explosion in the amount of data and the number of sources is well recognized. “It’s not just that companies are using Microsoft Office with documents in file shares and other documents in SharePoint,” says Bill Piwonka, CMO of Exterro. “There are many new community and collaboration platforms, some of which are in the cloud, collaboration tools such as Dropbox, Slack and HipChat, as well as mobile devices and social media sources.” The Internet of Things (IoT) will produce another burst of content. “Of all that data, an estimated 60 to 80 percent is redundant, obsolete or trivial, and companies sometimes don’t know where to start,” Piwonka says.
Exterro’s e-discovery suite includes legal hold, project management and e-discovery data management (collection, processing, review and production), as well as data mapping. It enables users to identify the most critical information prior to collection, which saves time and money. Exterro also offers information governance solutions that enable clients to gain insight into their unstructured data throughout the organization. The Exterro Project Management solution is designed to manage workflows associated with e-discovery and other legal processes.
One factor to consider, given the many data sources that may be subject to e-discovery, is how well each solution integrates with various file types. Although the majority, including Exterro, can consume a wide variety of file types, some are limited to a smaller group such as Microsoft files only. In those cases, another product will be needed to accommodate the other files. If the solution can also integrate with other enterprise applications such as HR software, finding the custodians by department during an e-discovery may be much easier.
“Increasingly, e-discovery is being treated as a business process,” Piwonka says, “which means companies can manage, measure and optimize it.” One thing that has been missing from e-discovery is the ability to orchestrate all the steps, so that the handoff at each stage is seamless. Piwonka explains, “When all aspects of e-discovery are coordinated, reminders are sent, acknowledgements are tracked and the output is an auditable, defensible record to prove that the company has met preservation requirements.” Another dimension of that coordination is the need to provide access to both internal and external stakeholders with appropriate role-based access.