Why the Smart Money Checks the Analytics Engine
In both senses of the word, fortune favors the prepared when it comes to litigation: fortune as treasure and fortune as positive outcome.
Three elements characterize superior preparation for e-discovery: good document housekeeping; precise extraction and distillation capabilities; and unified workflow from the left through to the right of the EDRM model. This article talks about the latter two, but it's useful to quickly touch on housekeeping.
Good housekeeping means saving, and properly categorizing, only what makes sense from the practical, regulatory and legal points of view. The cleaner the corpus of documents, the better you are able to respond as discovery progresses through early case assessment, review, production and litigation.
E-discovery capabilities and the quality of coordination as the process moves from left to right are the most impactful in terms of comprehensive discovery and cost efficiency. (While poor document housekeeping can complicate matters, it's comparatively easy, though expensive, to correct.)
The State of the Art Has Shifted
The combination of structural advances in computer-assisted review and growing realization of the value it delivers has dramatically changed how organizations approach e-discovery and adapt their procedures to take advantage of evolving analytical capabilities. For example, Gartner predicts in its May 2012 Magic Quadrant for E-Discovery Software report that e-discovery techniques "will become ever more widely accepted, and within five years they will be part of standard operating procedure."
Moving beyond the model of stringing together discrete best-of-breed point solutions, companies now see the workflow from left to right as a continuum. Integrated e-discovery suites designed to work together provide a consistent user experience throughout the processing and analysis phases—early case analysis, email threading, text near-dup, clustering, concept search, categorization, etc.
A consistent user experience is important to improving the efficiency of people's work. But the underlying technology—the analytics engine that does the heavy lifting—is the key element governing the speed, accuracy and effectiveness of today's e-discovery solutions. A proven integrated analytics engine that provides a wide spectrum of analysis options also provides a solid foundation on which to build a consistent suite. Its components have been developed to work together, and can cleanly hand off data from one step in the workflow to the next without having to re-factor it.
In addition, an integrated advanced analytics engine also simplifies maintenance, upgrades, improvements and support requirements. Consistent analytics technology in combination with an intuitive e-discovery application makes both training and knowledge simultaneously more focused and more broadly useful. This contributes to lower cost and higher productivity.
Analytics Engines Make All The Difference
The goal in discovery is to find all of the relevant information without the process costing two arms and three legs. Comprehensive, defensible computer-assisted review is a key part of achieving this since it substantially reduces, and more effectively focuses, costly human-review time.
Today's combination of robust integrated analysis options combined with improvements in power and performance enable companies to cost-effectively derive a higher-quality, concise review set from continuously growing document collections. Different, integrated cuts at the source information produce the best results, particularly in early case analysis when you're figuring out whether the action has any merit.
As advanced analytics technologies continue to scale in speed and volume, consider one example of how a company can change its operational approach to respond better to an action: eliminating the up-front culling of information in order to reduce the initial volume of documents for analysis.
The value of culling information prior to running advanced analytics is rapidly diminishing. It makes more sense now in terms of both quality and cost to evaluate the larger universe of documents. Although analyzing 10 million documents instead of a million adds time, if you do it right, you might find useful information that you never expected or wouldn't have found otherwise.
With greater insight upfront and clearer focus on responsive information, you can make more informed decisions much sooner—for instance, about whether to settle before going to court, or to push your case harder supported by more facts.
Zeroing-in Without Breaking the Bank
Companies deploying an integrated e-discovery suite powered by a comprehensive advanced analytics engine have multiple, complementary options for analysis. These capabilities go far beyond the familiar keyword searches and text-comparison.
Conceptual evaluation, for example, uses concepts reflected in example documents to find those with similar content that traditional keyword search would likely miss because the words aren't exactly the same. Clustering groups documents that appear conceptually related—very useful when you're not sure what's in the corpus. Categorization puts documents into different folders by comparing them to example documents, and powers technology assisted review. The difference? Clustering = "Show me what's there;" Categorization = "Find me things like this."
Automated analysis, such as email threading, can also reduce a number of related emails down to a concise review set that includes all the relevant information. Consider a 10-email thread in which the 6th and 7th contain all the relevant information from 1 though 5; and 8, 9 and 10 simply add "Got it," "Thanks," and "See you!" Analysis will indicate that reviewing 6 and 7 will show everything of relevance, eliminating the need for a person to review eight others.
Near-duplicate analysis based on concepts, and not simply text comparison, further reduces the collection that represents the entire body of case-related information. For example, it will recognize two documents as conceptually nearly identical even though a few paragraphs have been rearranged. More literal text-comparison analysis will see them as different documents to be reviewed by a human.
The net result of analyses like these is to produce the smallest possible collection of likely relevant documents for human review. E-discovery suites powered by an intelligently integrated advanced analytics engine will give you a range of options to perform comprehensive, effective and cost-efficient discovery.
Content Analyst's software provides advanced, conceptual-based search, classification and document analysis for a wide range of customers. For further information on the capabilities and value of advanced analytics, please visit ContentAnalyst.com or info@contentanalyst.com.