Microsoft Wants the Government to Let It Get Rid of Its Data
In August 2011, Microsoft sent a letter to the U.S. District Court's Advisory Committee on Civil Rules, which provides a fascinating look at the vision of biggest of the big business technology players for the evolution of law regarding e discovery, and its own internal e discovery practices. The letter can be found here.
Microsoft called for the amended rules to provide sanctions for the spoilation only in cases of willful destruction and prejudice to the requesting party.
Microsoft revealed its approach to electronic discovery was custodian based. In 2011, Microsoft was collecting an average of 17.5 GB from each custodian for litigation - a big jump from 7 GB in 2008. There were an average of 45 custodians involved in each litigation hold, and a total of almost 15,000 litigation holds around the time the letter was writen in 2011 - these holds effected 12.5% of its U.S. workforce. Importantly, only a third of these holds concerned cases filed in a court. The majority were issued in reaction to some other circumstance which gave Microsoft notice that litigation was a possibility. In 2011, Microsoft was holding the collected data needed for cases that were actually filed in central 'datacenter'. Most of the on-hold data was never actually reviewed for the purposes of litigation. Further, 95% of the data that is collected was filtered out using date ranges, search terms, de-duplication, and data minimization techniques. This filtering was performed by third party vendors whose work was overseen by only four Microsoft employees. Usually less than a quarter of the post-filtering data set was actually produced. Microsoft's study concludes with startling point that, "For every 2.3 MB of data that are actually used in litigation, Microsoft preserves 787.5 GB of data - a ratio of 340,000 to 1." When they say 'used' they mean entered as an exhibit at a trial. About 150,000 pages were produced for every 48 million that were preserved.
Microsoft emphasized not just the financial cost that comes from the holds on so much data, but also the disruption to its regular business caused by the holds. It anticipated that greater clarity about the nature of trigger events and the potential for sanctions would greatly reduce the time and money it spends on electronic discovery.