Storage Tip: Identify data pattern recognition

by David Hill

August 15, 2007 —

Send your Storage question to David Hill today! | See other Storage tips

What seems to be the problem? If you had your druthers, you would never want to produce a copy of key parts of your enterprise's data and deliver it to a third party to satisfy a subpoena as part of a lawsuit. But neither you nor your company has that option. You must produce that data. Remember that the changes to the Federal Rules of Civil Procedure (FRCP) are more explicit about what you need to do with electronically-stored information (ESI). But your obligations do not end with producing the required information. Your obligation is also to help put your legal department in the best possible position given the available evidence. And that may not be easy.

What do you need to know? You can help your legal department select and use the proper software to review the data. There are a large number of search engines that do a capable job in doing key word searches of a large number of documents. You must examine the different tools, but you should be able to find one that meets your needs. And in many cases the tools may be sufficient to meet your needs.

However, there are some cases where that approach is insufficient. The traditional approach is a document-by-document search. But what if information is missing or there is other relevant information than that subpoenaed? What if e-mail threads are not sufficient sources of information?

Cataphora is a software company that helps enterprises deal with high-end investigations where recognition of behavioral patterns -- not just document-by-document search -- is necessary. (Cataphora briefed me as an industry analyst on its product and service.) Cataphora doesn't look for the smoking gun (which may not be there), but rather the smoking hole. If you read "The Adventure of the Silver Blaze," Sherlock Holmes commented on the dog that barked in the night. Dr. Watson was puzzled as no dog had barked in the night. And that was the strange thing according to Holmes. For example, an employee of interest neither received nor sent e-mails for several days, but created new calendar events. That would be a strange thing. It would not be a smoking gun, but it would be a smoking pattern.

Cataphora provides its software as a service (SaaS). Sophisticated multigraph mathematics, such as is used in social networking, is at the heart of its analytical capabilities. Cataphora uses mathematics and computational linguistics to help defendants in complex civil litigation investigations.

Cataphora may request additional information -- such as phone records and expense account information -- that may not have been subpoenaed. Cataphora wants to collect all the necessary information from which highly indicative patterns may emerge. For example, no e-mails are missing, but all of a sudden, e-mails stop and phone records show that conversations occurred between two parties that up to then had been e-mailing each other. Why did that happen and what does it mean? Keyword search alone will not help you find the answer.

Understanding patterns can help attorneys better determine the quality of their case as well as provide an aid for positioning that case.

What can you do about it? Not all litigation requires a sophisticated investigation process; straightforward e-discovery tools for reviewing data may be sufficient. Complex investigative work is an add-on expense, but it is not always needed or appropriate. However, cases that could have a significant impact on an enterprise, such as a possible large dollar amount award, may very well justify looking into a company such as Cataphora. But even if the case does not justify the use of SaaS, you may want to look for any obvious patterns that you can detect now that you know to look for them.

Mesabi Group