Cubicle Meets Pandora's Box

by Conor Friedersdorf

In Slate, Michael Agger has an interesting article on the information a company can glean from the electronic footprints of its employees. What does filtering software look for as it scans terabytes of data? Signs that an employee is unhappy include e-mails written in all caps, "call me" events where e-mail conversations are taken offline, communication with other employees in distant parts of the org chart, and updates to the resume.

Signs of a good employee: blocks of text that they produce are widely shared or forwarded, lots of people are interacting with them via e-mail, and a tendency to answer open-ended questions posed by other employees or managers.

Mr. Agger concludes:

Yes, it's lame if a manager needs to rely on an algorithm to figure out who her most valued employees are. Yes, the Big Brother-ish aspect of all of this gives one pause. But if you set aside that reaction, most of what Charnock is talking about is common sense. Are you in the mainstream of your workplace or off in a little eddy of your own? If so, why? Are you being productive in your own time and style or just getting really good at Desktop Tower Defense and wishing you did something else? Your electronic tracks don't indicate your true value as an employeeWho cracks better jokes in the weekly meeting? No one!but it's naive to think they don't reveal anything at all.

My suspicion is that most companies would find engaging in this sort of analysis to be an unmitigated disaster. As Mr. Agger notes, it might prove useful for a big company buying a promising startup, but if an existing company's managers started running algorithms, flagging electronic content, and reading it, they'd quickly find themselves awash in information they're unqualified to interpret and unable to handle maturely. 

But the bigger problem would come when some employees inevitably grew savvy enough to start gaming this kind of analysis. Imagine if you knew you'd be evaluated by its metrics. How much work time would you waste writing calculated e-mails?