Can a company predict turnover with an algorithm?

The WSJ reported an interesting story yesterday.  Google is apparently concerned about employee turnover so they’ve developed an algorithm to calculate the likelihood that a given employee will depart.

I assume stuff like:

– age

– gender

– education

– option vest schedule

– current position

– salary history

– life changes

– commute time

Some might say that developing an algorithm for turnover is sort of like the argument about how many angels can dance on the head of a pin.  Or maybe, like re-arranging deck chairs on the Titanic.

I think Google is a fantastic company.  Having been in multiple meetings across many different groups and departments there, I found the company is filled with really smart people that by and large are mostly interested in doing cool stuff and enabling partners like us to benefit.  And they generate nearly $2B a month in revenue!

But.. like any company, its real asset is people, and people are human.  Some are political, some are ladder-climbers, some are real go-getters, some like to coast, some only get jazzed by a thorny challenge, you get the point.  An algorithm is nifty, but I think most experienced managers that have been around awhile can predict an employee departure or likelihood of departure on gut feel as well as or better than any algorithm.  Isn’t that a managers primary job?  To hire, motivate, and retain the best employees?  Creating an algorithm like this, says to me that Google might have deeper systemic issues.