Concept Searching

August 11th, 2009 | By admin

(This is a guest post by Rich Turner, Vice President of Content Analyst Company, LLC)

We hear a lot about “concept searching” and “advanced analytics” but very often, the definitions are confusing, and the situations where you would use features like these aren’t often clear.

We’re all familiar with keyword searching – as well as its limitations. Take for a very simplistic example, you’re working on a case involving leases in a 600-unit housing complex. The lease agreements aren’t standardized, so you’ll need to make sure your search terms include lease, leases, leasing, rental, renting, and since there are subleases involved, sublease, sublet, sub-let, and sub-leasing as well. You could use Boolean operatives (like wildcards, if/then statements) but those can be unforgiving.

Will you get all the relevant documents? The answer is maybe, maybe not – there are many ways to lease property. However, you are likely to find many irrelevant documents with any basic keyword search. Say that this building also provided leases for satellite dishes. Your set of keywords will return those as well – hundreds of irrelevant documents.

Few lawsuits present discovery searches which are this simple. This is why concept searching and advanced analytics are so useful. Most concept search engines can work right alongside keyword searches, and give you the ability to make a representative “lease” or even cut-and-paste the relevant paragraphs from that lease into a search window. The search engine, in turn, will find and rank all documents that use similar terminology and talk about similar concepts. Yes, it may also find the documents about satellite dishes – but those will be ranked very low on a scale of relevance because, while they’re definitely leases, they’re not about property.

Defensibility is often a question about keyword searching (in fact, keyword itself is only a de facto standard). Some judges are pushing hard for using advanced techniques, and many have been quick to point out the pitfalls of keyword-only searches. So what do you do?

Consider the use of concept searching as another tool in your arsenal. In the example above, if a document gets returned that is relevant but doesn’t contain any of your variant of keywords, perhaps that’s an indication that your list is too exclusive. If you’re reviewing the ESI that you have received, then a concept search can allow you to quickly sort “responsive” documents based on the keyword searches into truly relevant ones that actually address the issues you’re reviewing.

Most concept search engines also provide a number of other features as well, all driven by concepts. These include categorizing or grouping documents by concepts, and even clustering conceptually-related documents automatically into groups by concept. Properly implemented, these tools can greatly enhance your ability to bring an orderly process to a mountain of ESI and find relevant information more quickly. In today’s economy, speed and efficiency are more important than ever – and concept searching is one tool at your disposal with tremendous benefits.

About the Author:

Rich Turner is a Vice President of Content Analyst Company, LLC. His team manages all outward marketing programs, corporate communications, and partner programs. Rich holds a Bachelor’s of Arts degree in English Literature from the University of Bridgeport, and a Master’s Degree in International Business from Columbia University.

Content Analyst Company

Headquartered in Reston, Virginia, Content Analyst® is a leader of advanced search tools and technology. Content Analyst’s software known as CAAT, includes patented Latent Semantic Indexing technology, and provides advanced, conceptual-based search and document analysis for a wide range of customers, from highly-classified intelligence installations to world-class publishers to cutting-edge eDiscovery and litigation and support providers. CAAT exponentially reduces the time needed to discern relevant information from large volumes of documents and data. To learn more about how Content Analyst, visit their website at www.contentanalyst.com, or email info@contentanalyst.com.

Leave a Comment »

---