There has been an ongoing argument as to the validity of concept search verses keyword search in discovery searches. The main arguments I have seen are:
- Keyword searches tend to miss relevant documents and are under-inclusive in their search results.
- Concept searches tend to produce too many non-responsive documents and are considered over-inclusive in their search results.
- The other argument against concept searches for eDiscovery is that concept searches are a “black box” and are therefore very hard to explain to the court as to their validity.
I have not been able to find any cases where the eDiscovery response was conducted via a concept search.
While at LegalTech 2010 in New York, I spoke to several conceptual search/clustering vendors that were positioning conceptual search as the next big thing…that keyword search was falling in favor. I don’t believe that to be the case but I am curious whether conceptual search technology has a future.
I do believe there is an interesting possibility to use conceptual search capabilities in the area of Early Case Assessments (ECA). For ECA, the discoveree wants to “data mine” potentially responsive ESI to determine their going forward strategy; should we settle or should we fight? To make the best decision about legal strategy, I believe having access to the most complete and relevant data set is a top priority. One of the wraps against concept search is it is over-inclusive; a benefit in making sure you have reviewed all potentially responsive ESI when performing ECA.