Can We Prevent Top Talent From Walking Out the Door?

Biographical information (biodata) has been used to predict turnover and performance for a long time.  The idea is that certain verifiable aspects of a person’s life are indicators of future behavior.  To use an adage from an earlier time, if a person has changed jobs frequently in the past, s/he is not likely to stay with you very long.

We can now fast forward this idea to employee retention.  Or, put another way, can we predict which people are going to leave a company?  This article seems to indicate yes and it should not come as a surprise.

Putting aside privacy concerns for a moment, this approach goes beyond determining if a person is the right fit for a job due to their personality or values.  Rather, it potentially blends ideas that we would always think of creating turnover (bad boss, less pay compared to peers, length of commute, etc.) as well as those (number of startups in area, change in housing costs, etc.) that perhaps we had not thought of.    The added layer to the analysis is that it allows HR to say, “Wow, this is a person we don’t want to lose (or could not replace), let’s make some adjustments” or “Eh, that person is an underperformer anyway, so good riddance.”

More importantly, it treats retention as a dynamic, rather than static, state.  Previous biodata models would say, “This person has a 70% chance of staying 2 years or more.”  This data model might say, “Right now, this person has a 90% chance of staying through the end of the year.”  But, 6 months from now, if things change in the organization or in the person’s role, the model may say, “Right now, this person has a 60% chance of staying through the end of the year.”  This puts the onus on HR to work with managers to address the potential issues of valuable employees on a proactive basis.

From an employee’s perspective, I think there is opportunity here as well.  Imagine if they could pull up his/her “propensity to leave” score at any time.  Think of it as part of an employee engagement indicator.  This person could then see those things that may be causing them anxiety at work that might lead them to leave.  If it’s something minor that s/he feels could be easily addressed s/he could take it to a manager.  If it looks insurmountable, s/he would know that a new job search is a good idea.

The privacy issue here is real and, as with all concerns, depends how you feel about others using your data.  If I’m Amazon, Google, etc, there is a great temptation to link a person’s customer data with their employment application/employment status data to refine algorithms.  I have no idea if I’ve signed off on this when accepting their terms and conditions.  Do you?

Regardless of the “hotness” of the job market, this approach to dynamically tracking retention probabilities can be a very useful tool.  It can lead HR to being ahead of the game when trying to retain talent rather than offering the promises of a jilted lover as someone valuable walks out the door.

Thanks for coming by!

Please provide this information so we can stay in touch.

CLOSE