Our old friend keyword search has recently re-established its seat at the e-discovery table. Just two weeks ago, in the latest of a series of twists and turns in Kleen Products v. Packaging Corp. of America, the plaintiffs – who originally proposed to force the defendants to utilize a technology-assisted review strategy (which they called Content Based Advanced Analytics, or CBAA) that ignited the first real “predictive coding” case – withdrew their demand and agreed that the defendants could instead continue to apply good old fashioned keyword search methodology in the first request for production.
Rather than just “lamenting the end of the predictive coding debate in Kleen Products” like our other friends (although it’s not over until the fat lady sings, or, in this case, after October 2013 when the plaintiffs can re-surface the use of technology-assisted review in a meet-and-confer), we’re pleased to see that keyword search is back in the game. In our opinion, it never left the game.
Whether as a standalone search and retrieval method or applied in conjunction with other approaches such as technology-assisted review (as discussed in “Revisiting the Relationship Between Keyword Search and Technology-Assisted Review in Judge Scheindlin’s National Day Laborer Opinion” and “Why Machines Can’t Replace Lawyers”), the latest twist in Kleen Products shows that keyword search isn’t (and shouldn’t be) going anywhere anytime soon.
When keyword search is designed and executed both appropriately and strategically – that is, e-discovery teams leverage an iterative approach that is driven by technical expertise, informed and validated by attorneys, and tested and documented to defensibly eliminate significant volumes of non-relevant data without sacrificing relevant data – the advantages are many:
1. Reduced hosting and review costs. Strategic search can result in a deeper defensible culling – reducing hosting costs and the time and cost spent by attorneys reviewing non-relevant information.
2. A smaller, richer set of documents. Well-designed search terms, built with linguistic insights and thoroughly tested for precision and comprehensiveness, ensure rich and focused document collections – leading to a more expedient and informative path to case strategy and production.
3. Better search term negotiations outcome with opposing counsel. Statistical data and firsthand knowledge of the quality of your review population positions litigants for a better outcome in Rule 26(f) negotiations with opposing counsel.
4. Defensibility throughout the process. By utilizing a principled and well-documented approach, technical expertise called for by the courts and statistically valid measurement of results, search can “right size” document review and efficiently fulfill legal obligations for due diligence.
5. Focus on case strategy. Technical expertise can complement one’s search term development process, freeing up legal teams to focus on case strategy.
6. Flexibility. Strategic search can act as a machete to significantly cull large data populations, and simultaneously serve as a scalpel – identifying finer trends and patterns.
Combining legal and technical expertise with sound scientific search and information retrieval methodology can yield more cost-effective and defensible document review for e-discovery. It requires litigants to change their approach to discovery in some ways – engaging in more planning and design activities early in the process. Adoption of search best practices – again, either as a standalone methodology or in combination with others like technology-assisted review, can bring tangible improvements in the quality, efficiency and cost of review and analysis of documents.
We are not finished when it comes to innovative information retrieval in e-discovery, but keyword search will continue to have an important role in practitioners’ toolboxes.
Sheila Mackay is Senior Director, E-Discovery Consulting at Conduent. She regularly advises in-house legal teams and outside counsel on best practices for utilizing advanced search techniques as well as CategoriX, Conduent’s technology-assisted review offering. She can be reached at firstname.lastname@example.org.
About the Author