DiscoveryMining

I just read Craig Ball’s post about using keywords to search for privilege, then producing the resultant set. He lays out a nice argument for eDiscovery keyword search, which he calls the “blunt instrument”. Search is a great tool for many purposes, but I prefer to frame it in terms of Analytics versus Review.

I take Review to mean exactly that, reading documents. Users look at the documents, review them (read them) and then make a decision based on strategy or obligation. Analytics is also the reference point I would use for conceptual or topical searches. Both functionalities analyze the data and allow you to logically group the documents into ‘like’ buckets. Discovery Mining offers the latest in all of these technologies, semantic and keyword, plus the analytics for grouping documents, yet I would not call this Review.

It is tough for me to admit, because I am a search engineer at heart, but nothing beats a good set of eyes on a page. There are studies which make the case that computers are just as effective at making a determination of relevance as humans. But with everything that is at stake, I understand why Attorneys would want to ‘eyeball’ documents before they go out. At Discovery Mining, our clustering and semantic analysis combined with keyword search is a powerful method to trim a collection, but this is only part of the eDiscovery process.

We continue to design with our intention toward streamlining the review of massive volumes of data because increasing data volume truly is the core issue for eDiscovery and why you need analytics in the first place.