Optimized Keyword Selection – Another Component of ECA?

Migrated from eDJGroupInc.com. Author: Barry Murphy. Published: 2010-06-23 10:59:15  As I work on some early case assessment (ECA) research, I realize just how complicated and multi-faceted ECA actually is even though our industry refers to it as a point solution.  In an earlier entry, I wrote about ECA being the high-end of legal decision support, but perhaps the better way to state it is that ECA is part of the spectrum of legal decision support.  In fact, ECA is a spectrum in and of itself, ranging from keyword filtering through powerful content analytics like conceptual clustering.What really got me thinking about how ECA is made up of several components is a briefing I had on RenewData’s new offering called Anagram.  Basically, this solution helps legal teams come up with keyword strategies.  Not being a lawyer, I hadn’t thought much about keyword strategies.  However, mapping out what keywords to include or not for identification can be a long, expensive process.  RenewData’s Chief Scientist is Andy Kraftsow, founder of Dolphin Search and Digital Mandate (since acquired by RenewData), who has experience as a linguist.  What RenewData hopes to do with Anagram is save time in the keyword selection process by looking at the process backwards.When lawyers begin looking at what keywords are important for identifying potentially responsive information, the process seems fairly straightforward.  Look at the key issues; brainstorm on concepts, words, and synonyms; select the keywords; conduct the collection.  Tracking this process, however, can be cumbersome.  Imagine trying to explain how you generated a keyword list based on several hours of brainstorming without detailed notes or a virtual recording of the brainstorming sessions.  What I find interesting about RenewData’s Anagram is that it aims to make this very necessary process more efficient.  Anagram looks at all the words in a collection and then presents potential keywords in ordered lists based on linguistic analysis.  It helps to make a cumbersome process more transparent and scientific.As I listened to the briefing, my mind of course went to all the potential applications of such a solution – creating keyword lists for compliance and data loss prevention applications; getting a jump start on the hot keywords for downstream first-pass review, etc.  It then hit me that optimized keyword selection is really yet another component of ECA.  It can make the process of identification, collection, preservation, and first-pass review go faster – to me, that is one of the big goals of ECA.eDiscoveryJournal will be producing a more in-depth report on ECA and all the components, including optimized keyword selection.  If you know of other approaches or solutions for this, please do let us know by commenting.