It’s well known that litigation is expensive and review costs can add up quickly, especially when there’s plenty of ESI associated with a case. Corporations are looking to minimize costs wherever they can, and this is certainly true when litigation is on the horizon. A crucial place to cut down on costs early on is during keyword development by more quickly identifying the responsive document population.
Not only can having a good understanding of the necessary keywords reduce the costs of review by decreasing the total number of documents to be reviewed, it can also improve the precision and recall of your search to confirm you have the documents you need. By avoiding missing any documents now, you could avoid costly sanctions later.
While traditional search term definition provides a sound starting point, analytics tools can help counsel take it to the next level. These tools speed the identification of the terms that will provide a narrow but highly inclusive responsive document population.
Concept search is a powerful tool that allows you to identify previously unknown search terms and documents by discovering ones that are conceptually similar in content.
It works by identifying a particular concept — document, paragraph or search terms — and then finding conceptually similar documents to that material. For example, you’re reviewing a document that is found responsive and want to find additional similar documents. In this case, concept search would work like the “More Like This” functionality many online shopping carts have and allow you to display all related documents. You may then identify within these new documents other terms that would be beneficial to include in your keyword list.
Keyword expansion helps to identify previously unknown, important search terms in your document population. It works similarly to concept search but specifically for search terms to ensure that no key term is left out. Another example, you have multiple keywords already identified, for instance ones the EDRM site gives: BMW 1150, brakes, accident and maintenance. After running a keyword expansion on those terms, you might see the word stoppies. This word was not on your original keyword list but was frequently used when your other keywords were, which allows you to expand your keywords and catch additional responsive documents you may have otherwise missed.
Clustering is a way to visually organize documents in conceptually similar groups so reviewers have a better idea of what exists in the document population. This can be helpful not only in understanding additional search terms to include, but also which terms to exclude.
A final example, you’re reviewing documents from a class-action lawsuit brought against an automaker whose defendants claim their cars are lemons. If you only pulled results with the word lemon(s), and there were documents with references to the food lemon, you would retrieve documents pertaining to those as well. Clustering allows you to see there’s a cluster of documents with the word lemon and car as well as a cluster of documents with the words lemon and cake. This additional information offers several courses of action. The non-relevant cluster of documents (lemon and cake) could simply be excluded from review. The non-relevant cluster could be used to generate keyword terms that would exclude documents (lemon not within X words of cake) or the keyword terms could become more specific to only include the relevant cluster (lemon within X words of car).
An important step in any keyword list strategy is testing and evaluating the effectiveness and reasonableness of your search terms. Reviewing the results of keyword searches generally or more rigorously through sampling can validate your list. Analytics features like keyword expansion, clustering and concept search provide another means of testing your list and refining it based on the results.
During review, it’s important to get your hands on the most responsive documents as quickly as possible. Using keyword analytics not only helps you in strategy and finding these documents quickly, it can also help keep costs down and avoid expensive sanctions.