E-Discovery - Proper Preparation Requires Data Retention Policies

The second of a six-part series on e-discovery from the Taneja Group .

There is a cautionary tale for storage managers in the court case known as Zubulake vs. UBS Warburg.

Most people know about the multibillion-dollar judgment against Morgan Stanley for poor e-discovery practices. Another classic example of the risks of poor discovery is the infamous Zubulake case, which should have been a routine employee discrimination case. But as e-discovery continued, the court found that the defendant failed to produce evidence from backup tapes in a timely manner and, worse, overwrote entire backup tapes when the original data should have been placed under litigation hold. The court slapped sanctions and costs on the defendant for their missteps. Even worse for the defendant, the judge instructed the jury that it was free to infer that the missing tapes would have been unfavorable to UBS. The judge was careful to say that the jury was not required to do so, but was clearly permitted to draw that inference. Of course, the jury did just that. The result was a $29.3 million judgment against hapless USB.

Compliant and consistent data retention policies would have saved USB millions of dollars in sanctions, attorney costs and judgments. And USB is not alone. Most businesses with $100 million or more in annual revenue are involved in some type of active litigation, and large corporations handle hundreds of active cases. This level of active litigation makes the e-discovery process ongoing and expensive in terms of time and resources. One common estimate is that manual e-discovery procedures average $2,000 per Gbyte of reviewed data. Keeping in mind that a single e-discovery procedure can easily involve multiple terabytes of data, $2,000 per Gbyte is an expensive proposition.

Many attorneys plead the burdensome costs of data e-discovery to the court, and federal rules of civil procedure (FRCP) do protect parties against unreasonable burden and cost in discovery motions. However, if the party has poor data retention practices in place -- say with badly structured data storage or poor litigation holds -- then those rules will likely not apply. In Farmers Ins. Co. vs. Peterson, the judge stated that the "unilateral decision on how it stores information cannot, by itself, be a sufficient reason for placing discoverable matter outside the scope of discovery." In other words, poor data storage will not save a company from heavy e-discovery costs, and will cost it much more to search in the long run.Good data retention policies help a corporation to hold the data they should and to delete the data they do not need. Several e-discovery applications, such as the Atlas platform from PSS Systems Inc. , serve this function by classifying discovered data and retaining it for business value, regulatory matters, or current litigation.

Still, it is incumbent on users to establish the definition of relevancy. Questions to ask include: What retention periods for what data does regulatory compliance require? What do internal governance and investigations need for oversight? Where is my data with business value and when does it cease to be valuable? And what data do I reasonably need to keep available for litigation purposes? With the majority of organizations having over 95 percent of their data in electronic format, these are serious questions to ask.

To answer these questions, organizations must standardize retention and data management policies. This procedure has many important benefits, including lower costs of storage and storage management, lower energy costs from less stored data, better data value, and much less risk for data loss or non-compliance. As for litigation e-discovery, such policies make it possible to quickly and accurately run e-discovery tools across large and distributed storage locations and make it relatively easy for organizations to be in compliance with court orders. For example, StoredIQ Corp. operates across a wide range of storage platforms and applications to classify data types. Autonomy Corp. 's Zantaz has similar capabilities using its Idol engine.

Is this easy? No. Many present-day retention tools are application-specific and data can be widely scattered across many storage devices and locations within the same organization. However, by establishing and communicating retention policies, centralizing and consolidating data, and working closely with legal and compliance officers, IT can close the gap between a high-risk data free-for-all and a finely tuned enterprise information infrastructure.

Another challenge is reactive versus proactive tendencies when it comes to e-discovery tools. Even powerful e-discovery software takes time to initially crawl through all enterprise storage areas. Proactive deployment gives tools the time they need to fully index data sources. When an e-discovery demand comes up -- and it will -- corporate content is already indexed, ideally de-duplicated and ready to yield compact results sets.In the preparation phase, e-discovery classification software crawls through multiple storage destinations and data types. Kazeon Systems Inc. , whose technology Symantec Corp. (Nasdaq: SYMC) uses for Enterprise Vault, offers this capability. The software indexes data in the initial phase, and then runs transparently in the background to keep indexes current as data is created, modified, and retired. And this is where a partnership between IT and legal comes in very handy. IT can create a well-organized data infrastructure not only for e-discovery but also for resource management, budgetary reasons, and compliance. And the better structured the data, the faster e-discovery searches will go. Huge amounts of company data reside in backup and archival environments, and the e-discovery product should have integrated access to existing back-end storage systems.

Next, the second step in the e-discovery workflow: Collection.

Part one of this series can be found here

Christine Taylor is an analyst with Taneja Group , which provides research and analysis to the storage, server, and knowledge management industries.