Despite placing #1 on BusinessWeek’s annual ranking of “The Top 25 Ideal Employers” for the third year in a row, Google is facing its fair share of talent challenges. At 22,000 employees, it can no longer provide the start-up appeal it once did to its entreprenurial work force. Over the past few months, industry analysts and recruiters alike have watched as a number of Google employees have left the company to join other hot startups, including Facebook and Twitter. Other Google employees have gone on to start ventures of their own, some more successful than others.

So what do you do when your talent starts walking out the door?

Well, if you’re Google, you create an an algorithm.

As the Wall Street Journal recently pointed out, Google has created an algorithm to predict which employees are likely to leave the company, often times before the employee knows him or herself.

As the WSJ explains:

“The Internet search giant recently began crunching data from employee reviews and promotion and pay histories in a mathematical formula Google says can identify which of its 20,000 employees are most likely to quit.

Google officials are reluctant to share details of the formula, which is still being tested. The inputs include information from surveys and peer reviews, and Google says the algorithm already has identified employees who felt underused, a key complaint among those who contemplate leaving.”

From a managerial perspective, this insight could be quite useful depending on how those inputs are structured. It’s interesting to think about “employee engagement” as a key performance indicator, or an average “likely to leave” rating which could give a manager a close-to-real-time employee morale index.

In addition, could managers monitor the changes in employee morale as a result of implementing policy changes? For example, did a certain training program leave employees feeling more engaged over a certain time period?

If bundled with the latest social network analysis software , a manager might be able to visually see social clusters where morale was low, and potentially isolate the issue to the leader or some other characteristic specific to that group of employees.

Not everyone shares my optimistic view of the potential uses of such data. Some of the comments on the article:

ZAC HINKEL wrote: This reminds me of the movie ‘Minority Report.’ Just because you can, doesn’t mean you should.

Mark Shallcross wrote: There’s a precious quote buried in there: “… and some found the company’s human-resources programs too impersonal.” Now, why would they think that?

James Curry wrote: This very action is indicative of why Google is shedding talent. Math cannot manage human beings.

CHRIS GROTHKOPP wrote: “You will now be given your mathematically appropriate level of positive feedback today to keep your departure quotient in the acceptable variance range.”

What do you think? Is this an effective use of data? Will it help Google solve the ‘talent exodus’?