Taking Recruiting to the Big Data Level

In the end, recruiters are sales people.  They find prospects (potential candidates), qualify them, and hopefully get the best into the hiring pipeline.  Before job posting websites, to find candidates recruiters had to have great personal networks.  There is still no substitute for that for specialty jobs, but access to candidates has never been easier than it is today.

Of course, when candidates come to you they are going to be of varying quality, which gives recruiters the big task of sifting through online resumes, etc.  What if there was a way to set your skill based qualifications and then search for them online to identify candidates?  Gild claims to do that for programmers.  They have an algorithm for searching the web for programmers’ code, reputation among other programmers, and other job related (merit, as they put it) information.

I am intrigued and interested by this approach.  In essence, football teams just got done doing this for their draft.  Teams looked at every piece of game film available on every college player, organized workouts for those who showed the most promise, then held private workouts and interviews for the best of the best.  Of course, the universe of eligible college football players is much smaller than that of computer programmers, and someone still had to look through the film.  Guild suggests a process that is completely done by machine.

One criticism I can see is people questioning whether or not this kind of process captures the personality issues which can be important to success.  I don’t think anyone is proposing that such a process be the only part of recruiting and hiring.  However, if you could scour the internet for all of a person’s writing, blog posts, etc, I think you can get a pretty good idea about their personality.  You’d better believe someone is going to figure out a way to get a machine to score those things.

My concerns are around how what Gild looks for would be validated.  Is the algorithm built based on what should, or what has been shown, to predict performance?  Clearly, finding someone who wrote good code on a similar project should be a good predictor (a work sample of sorts).  But, does a programmer’s reputation among other programmers really matter?  Big data approaches are very good descriptors of what’s out there.  In this application, it needs to be a good predictor as well.

Any approach that brings data to the recruiting process is a good one.  If I hear another recruiter say, “I have a good feeling about this person” I’ll be ill.  If we can demonstrate the relationship between tests and interviews and performance, we should be able to do the same for the background and resume screening tools that recruiters use.  The only problem is that the history of research on these past experiences does not provide one a lot of optimism.  Perhaps we haven’t been looking at the right experiences, or enough of them.  Or, maybe it’s only now that the digital footprints of our candidates provide enough data for meaningful analyses.

Regardless, the amount of data publicly available on all of should be a treasure trove for recruiters.  While the information that big data will bring will have valuable insights about candidates, it is still incumbent on HR to ensure that this information helps us make more accurate and informed choices about candidates.  Just because some data seems like it is valid doesn’t mean that it is.

For more information on skills assessment and talent management, please contact Warren at 310 670-4175 or [email protected]

Thanks for coming by!

Please provide this information so we can stay in touch.

CLOSE