Finding the Unindexable

The dirty little secret of discovery search is the myriad attachments and files that cannot be rendered into absolute text that meets the simplistic Boolean criteria that is the output of the typical attorney salvos. No one wants to talk about all the images, audio, video and other categories that give heartburn to the viewers and iFilters used by the search engines. Lawyers, service providers or even IT execute the searches using the agreed upon search criteria, deliver back the results and do not ask the critical question, "What did we miss?"

Back in the dark old days of purely paper productions, which is only 8-10 years ago, no one wanted to open the Pandora's box and talk about email and other ESI. Instead, they asked the custodians to simply print out everything relevant and deliver it with the rest of the 'records'. Imagine the inefficiency of a senior vice president staying late to print a year's worth of manually selected email. Then the vendor made three or more copies for internal, external and expert review. The story gets worse from there, but that is supposed to be old news. I remember a client telling me, "We don't ask about email and they don't ask about email. Get it?"

That brings us back around to the 'unindexable'. Now it is not really unindexable, it is just expensive and complicated to deal with in comparison to the typical MS Office files. In many cases there is a good argument that these file types are just not germane to the key issues in the matter. In some cases, they are clearly germane and critical to understanding the fact pattern at issue.

So if a company has invested in a centralized message archive and has enabled a 100% capture system like Estorian's LookingGlass, Exchange journaling or transaction log shipping, then it would understandably and rightfully rely on these systems as the 'true' repository of communications over the incomplete user collection. The tendency will be to negotiate search or collection criteria with the other side as the primary collection to be reviewed. If critical items are not found by the search, the odds are good that within the sea of ESI these items may not be manually turned over by the custodians that are interviewed. "After all, we have everything in the archive, don't we?"

Good validation testing of search methods, quality assurance checks and sharp counsel can catch these issues. They can make sure that descriptions of the search and collection systems clearly represent known exceptions and search limitations to protect the corporation against accusations of deliberate lack of compliance. But what if the other side singles out a series of pictures of an accident scene and demands that you find them, even if the file names and properties were altered as they were forwarded through the chain of command? Even worse, what if you find clearly inappropriate images during your review and the CEO demands that you find every copy within the firewall and get a list of the offenders to HR?