Practical Advice for Reducing Privileged Review Burden (and Costs).

Back to Blog

When conducting privilege review in litigation, law firms and e-discovery providers often design keyword lists to identify a subset of potentially privileged documents that attorneys then review.  The keyword lists used typically cast a broad net in targeting potentially privileged documents to protect against inadvertent disclosure.  While this approach may protect privileged materials, it also captures many more documents than necessary, extending the time and raising the cost of downstream privilege review.

An alternative to the broad net approach—one that addresses the disclosure concerns while reducing downstream costs—is to conduct an upfront analysis of the keyword list to refine the keywords that tend to over-capture so the resulting potentially privileged set requiring attorney review can be reduced.

Usually, a basic keyword search for privilege will contain attorney and law firm names and keywords such as “attorney*”, “lawyer*”, “counsel”, “law firm*”, and “privilege*”.  These generic terms, however, can return a large number of documents that aren’t privileged. If you isolate the set of documents that are uniquely tagged by those particular keywords and see that they’re adding significant volume to the overall privilege set, it makes sense to conduct a further analysis to see if the volume can be reduced. How?

  • First, select a random sample of 200-500 documents hit by those keywords for review.
  • Tag the sampled documents as either likely privileged (applied quite broadly) or not likely privileged.
  • If a document is found to be likely privileged, copy paste the text that indicates why into a word document. This will inform how to refine further search strings to identify similar documents and exclude others.
  • Note trends for documents or document types hit by the keywords that turn out to be not likely privileged (e.g., newsletters, public documents, etc.) or particular portions of a document that are triggering a hit  but aren’t really identifying privileged material (e.g., auto-generated email disclaimer footers).

The information from the sample review can then be used to refine the privilege keyword list. This typically takes the form of grounding the original terms with other search terms to improve the precision of the privilege hits. It also has the effect of expanding the number of search terms overall.  For example, an initial search for “attorney*”, which could result in a number of non-privileged documents being returned, may be replaced by multiple refined terms such as  (contact OR discuss*) with w/1 (attorney{s} OR counsel) and (counsel OR attorney{s}) w/5 involv* and so on.

To further reduce volume, create a separate set of search terms that will exclude certain types of documents that have a high probability of not being privileged (e.g., newsletters, footers).

Iterate this process one to two more times with refined keywords at each stage. You should find that there is now a significant reduction in the number of non-privileged documents hit by the privilege keywords, but you haven’t sacrificed privileged documents in the process. Once you’ve reached an acceptable result, the refined privilege keywords can be used along with the more specific attorney name and law firm names (and all their variations) for a comprehensive but substantially reduced privilege review set. You may even be able to leverage this list for use in finding privileged material for other matters.



Thank you for subscribing to the H5 blog, True North.

We strive to provide quality content on a variety of topics related to search, eDiscovery and the legal realm.

Please check your email inbox for your subscription confirmation!