A practical example for HR analytics

A former student of mine from IIM Ranchi called me a few days ago with a challenging problem. Her  consulting assignment was to figure out metrics to identify the value that people were bringing to the organisation with an eye on rationalising (don’t you just love that innocuous word?) the workforce.  She wanted to know how she would need to start thinking about it.

So, how does one think about this? I have found that the best way to start is to work backwards from the overall strategic objectives of the business. Why was the business looking at rationalising their workforce? Was it because they were changing directions and needed to bring in fresh blood with alternate skill sets? Were they trying to infuse some fresh thinking in the organisation? Or were they trying to just cut costs and the flab that inevitably accrues in any large organisation? Turns out it was the last of these, with the organisation feeling that while the salaries were among the lowest in the industry, their overall wage costs were considerably high and were impacting their bottom line.

Once this was clarified, the next thing to do was to go out and get some data on revenue/employee and cost/employee for the organisation vis-a-vis its competition. This would then allow us to see whether what the management felt about the excess flab was actually based on data or just came out of someone’s gut instinct. As a data scientist, gut instincts are fine for enabling quick decision-making but data sure helps in making more considered decisions and it is always good to check assumptions rather than taking them for granted.

We then got down to figuring out numerous ways on pulling together the data around people to figure out the value they brought to the organisation. The utopian way (providing we all lived in a perfect world) would be to identify the varying skillsets of each and every member of the organisation and give them weights according to the value they hold for the organisation going forward. A further refinement to this is to identify the degree of automation that each of these skillsets could undergo and give those weightages as well. A combination of these two could then potentially identify individuals who would add the most value to the organisation. Of course, one could get even more creative and start adding performance data to this to enable a further ranking of candidates or build a model that also factors in the propensity of the employee to stick with the company as well as include the employee’s compensation as another variable into the mix.

While all this sounds great from a data scientist’s perspective, an HR practitioner must use this in conjunction with their own perceptions of the employees. Above all however, they must implement any decision with a lot of empathy.

The writer is founder and CEO of HR analytics start-up Factorial Analytical Sciences