Keyword searches seem easy enough to understand – after all, everyone is accustomed to using search engines online. However, when it comes to combing through potentially responsive documents to be reviewed for production, developing an effective search term list can be a little more complicated.
Case teams will want to sort through data responsibly, avoiding a burdensome production request while also being thorough. If you are drawing up an ESI agreement during the meet and confer, including a list of search terms can ensure both parties are on the same page. Ahead of this meeting, a good practice is to sample documents using your proposed search terms and appropriate search parameters. If you find responsive data, that will help you refine those search terms for use during collection and review.
Best Practices
Create search terms using a methodical approach beginning with your case’s narrative core: the who, what, where, and when of the matter. From there, you can look at specific language and limitations you want to apply to your search.
You may wish to build your search term strategy in layers, such as:
- Issue Analysis: define each issue that characterizes the claims and defenses of the case.
- Define Expressions: identify specific phrases or expressions that summarize each issue, keeping in mind that more than one expression may be needed per issue.
- Identify Components: in order to expand your search to capture related concepts and synonyms, it can be helpful to refine each expression into specific components, which can be run through an additional search.
- Sample Testing: with your case team, review sample data sets in order to enhance expression statements and improve precision on which documents you recall.
As you develop a search term strategy, be sure to also speak with your discovery team about how you plan to search for non-standard or structured data, such as spreadsheets and other files that are not text-based.
Take additional caution when searching for numbers, as these might return unwanted results. For example, searching for “1,000” can also return “1,000,000.” False positives are also hard to avoid; for example, searching for “comput*” will return “computer” but also “computational” although you can use proximity searches to limit some of these.
Another hurdle to overcome is learning the peculiarities of your specific search engine. Not every search engine uses the same logic: some will read commas as spaces, which would make “1,000” indexed as “1” and “000”. A surefire way to stay ahead of these issues is by working with an eDiscovery vendor, who can help you determine the search terms and related software needs specific to your case.
In Summary
Once they have been tested and developed over several iterations, keywords can be a helpful way to cull down a large volume of data. While the process can sound complex and daunting, developing an optimal search term list is easy with the right technology and team of experts at your side.
