Don’t Stop Believin’: The Staying Power of Search Term Optimization

  • Keyword search
    September 4, 2018

    Some things never outlive their usefulness. Just as a steering wheel remains the preferred method of driving a vehicle, the use of keyword search terms in ediscovery remains a powerful technique in the ediscovery world. In recent years there has been considerable attention on artificial intelligence (AI) and machine-learning techniques being applied to the challenges of ediscovery document review. Technology assisted review (TAR) has become an evolving trend in the industry, with ediscovery vendors and law firms abandoning search term strategies to jump feet first into TAR (pun intended).

    Keyword Search

    While these emerging technologies and new techniques have a prominent place in the eDiscovery toolbox, keyword search will never go out of style as an effective and defensible process. Additionally, search term negotiation is easier to conduct than TAR agreements because the process is transparent, simple to execute and easily justifiable to the courts.

    Optimizing the process of developing keyword searches is no easy task. Because few understand the science of search, the use of outside expertise is beneficial in building an effective and comprehensive search term strategy. Without such guidance, the efficacy of the approach can be difficult to measure. Additionally, negotiations over proper usage of search terms may become onerous and contentious. Thus, judges are often tasked with making determinations regarding the aptness of the methodology, and many are reluctant to do so.

    The judiciary recognizes that the skills required to create an effective search term strategy vary widely from those required to practice law. Consequently, they are recommending that attorneys use individuals with the appropriate expertise to advise and guide them. Enter the ediscovery liaison.

    For example, in Assured Guaranty Municipal v. UBS Real Estate Securities, U.S. Magistrate Judge James C. Francis IV explained, “A court should be hesitant to resolve issues that demand technical expertise…” and adamantly recommends, “… the appointment of a neutral consultant who will design a search strategy.” He explains that the use of such experts, “would be necessary for [the Judge] to offer an opinion as to the most efficient search protocol.” 

    Other courts agree. In Victor Stanley v. Creative Pipe, U.S. District Court Judge Paul Grimm explains, “Selection of the appropriate search and information retrieval technique requires careful advance planning by persons qualified to design effective search methodology.” It is clear that the judiciary recognizes that the expertise required in developing and defining an effective search term strategy resides with information retrieval (IR), linguistic and data scientists.

    An IR approach to ediscovery would use the same systematic and scientific approach as a research project. First, define what you’re looking for by building a narrative around the who, what, when, where and how. It’s imperative to have a complete understanding of the essential topics and to approach the identification of documents that address those topics in a logical and systematic manner.

    A major shortfall when crafting search terms is the tendency toward overly broad and ambiguous singular terms and phrases. Typically, this results in a large percentage of false positives in the result set. Disambiguating keyword search terms by adding appropriate context and synonyms is a common best practice with information retrieval consultants. These professionals understand how to build searches that can be broadened or narrowed, depending on the precision requirements of their research. In addition to a thorough understanding of the use of proximity to assign weight to selected components of the narrative, these professionals also understand the role of stop words, misspellings, and other search engine and linguistic issues that must be considered.

    Prism’s Search Term Optimization process is akin to constructing a building. First, lay the foundation with a clear understanding of the issues of the case and the types of documents that will support your strategy. Once a solid foundation is built, the structure of language, logical expressions, and metadata are blended as necessary to create the appropriate set of robust Boolean searches. These searches then target the retrieval of responsive documents.

    Prism’s consultants are highly skilled in linguistics, search lexicon, statistical validation, as well as various search and retrieval technology. They have developed and refined this innovative approach through many years of providing advisory services to both law firms and corporate clients. In addition, Prism has frequently been appointed by the Court as an ediscovery liaison to assist in the negotiation and resolution of search term disputes.

    Download the White Paper

    Read more about this workflow in the white paper entitled, Don’t Stop Believin’: The Staying Power of Search Term Optimization.

    Assured Guaranty Municipal Corp. v. UBS Real Estate Sec. Inc., No. 1:2012cv01579 (S.D.N.Y. March 25, 2013).

    Victor Stanley, Inc. v. Creative Pipe, Inc., 250 F.R.D. 251 (D. Md. 2008).