Corporations are racing against time to create archives that allow retrieval of e-mails in response to increasingly common civil suit discovery motions.

Losing that race can be costly, as Morgan Stanley learned when it lost US$1.44 billion -- most of that as penalties assessed by an annoyed judge -- in the Sunbeam case, after it failed to produce e-mails required by the court.

E-mail is becoming a standard part of civil discovery. Opposing counsel is well aware that e-mail archives are rich veins to mine, and last year unstructured data was specifically added to the discovery list in Rule 26 of the US Federal Rules of Civil Procedure. And the threat isn't confined to Global 2,000 corporations. Any organization can face a civil suit, and any civil suit can include the requirement to produce relevant e-mail. CIOs in organizations that are as yet unaware of this legal trend should alert their CEOs to the danger.

For most enterprises today, the challenge is to find the requested e-mail at all. In most organizations the archive is little more than a pile of forgotten tapes originally intended as backup in case of a computer crash. How many IT organizations today can even find the tapes from two years ago? In the Sunbeam case, Morgan Stanley annoyed the court by dribbling in tapes as it found them, sometimes in closets and other forgotten corners in various offices nationwide.

However, say the experts at Wikibon.org, even as enterprises scramble to find those lost tapes and create a real archive that will allow them to produce sets of historic e-mails if required, they should think beyond this purely defensive requirement to active use of the archived data.

E-mail contains valuable data

"Most unstructured data has very little metadata associated with it," said Wikibon.org co-founder David Floyer at a public Peer Incite meeting held by the organization last week. "E-mail is an exception. E-mail headers contain a great deal of useful information about who is talking to whom and how organizations really work.

"There is a huge requirement for automated data classification for semantic search," he said. "These are essential for organizations to protect themselves."

For example, he says, the human resources department may need to search the archive for early evidence of emotional or sexual harassment, sales may want to search for indications of contract abuse, and purchasing can search for evidence of employee theft. Executives then can intervene early to avoid problems before they reach the stage of civil or criminal action.

Martin Tuip, manager of business development at e-mail archiving house Mimosa Systems and a meeting participant, suggested that analysis could go far beyond the initial concern over legal issues to supply insight into how the organization actually works, who influences whom and what the unofficial lines of communications are.

This could provide insights, for example, into why some projects succeed while others fail. He envisions organizations creating a single repository of unstructured data that can then be mined by business applications as needed for information.