I was on a phone call the other day with the VP of HR at a fairly large company. This call we were discussing the HR Analytics setup at his previous company and the pros and cons of it. He’s a great guy that always has a lot of insights and I learn something every time we talk. This call was no different and it got me thinking about Traditional HR Analytics and it’s promise and limitations.
During our call we were discussing the fact that at his previous company they were able to draw some correlation between the model of PC that an employee was issued and their likelihood of leaving the company. I thought that was sort of interesting and filed it away for later as I started to think about the mix of hardware in place at any given company and the complexity of that being a legitimate lever that HR and Management could pull to increase retention.
Tidbits of data insight such as that are interesting and can play a role as one piece of the puzzle that plays into the much larger pool of attributes that HR Analytics systems look at when determining At-Risk Employees. But it also is not a quick fix type metric as in, “Hey! Look what we found now let’s go issue 3,000 new PCs to all of our at-risk employees that are still using a Commodore 64!” That probably wouldn’t fly, but it got me thinking about At-Risk algorithms more and what they are missing in most traditional Analytics packages. Three of the top scenarios that came to my mind are below.
1. The Director of R&D That You Hired 6 Months Ago is a Real @$#@$
It’s not likely that this is going to pop on your traditional Analytics report within 6 months of hiring this guy. As a matter of fact, if it does pop on your Analytics report it likely means the damage has been done already (ie. a shared characteristic in x% of the people leaving the company is that they work for this guy). The early signs that you have a bad manager in an org are typically seen through attrition, exit interviews, HR complaints, and ultimately you will not see this guy’s name in any HR analytics report before the damage is done.
2. I Didn’t Think I Was At Risk But Now That You Mention It
Analytics people love data and they love algorithms to sort through that data. They may think of themselves as rockstars in the business world today but back in the 80s and 90s they were probably at home on prom night. With a book. About data. Alone. But, anywho… Those same algorithms which can unearth some interesting connections and tell us how likely you are to leave the company often miss something essential: employee sentiment. Your algorithm might tell you that I am at risk because I have a Commodore 64, I’ve been in my role for 25 years, I make 20% of the living wage of a person living in Ethniclashistan, and my last name starts with a G but you never really bothered to ask me if I thought I was at risk. I might be happy as a clam despite what your algorithms tell you. I’m a big advocate for looking for truths in between the algorithms and the voice of the employee — you’ll find some amazing insights here that help you to fine tune your workforce strategies as well as your algorithms.
3. Collaboration is Hard to Measure
Data that pieces together how someone actually works and interacts with others is actually really hard to come by. It calls for different methods of data collection that don’t exist in typical HRIS’ and it calls for some outside the box thinking. Traditional HR Analytics systems don’t handle this metric well, if at all, today.
If you want to chat about Predictive HR Analytics and how to tackle some of the above issues shoot me an email at philg [at] peakemployee.com and let’s talk!
Thanks for reading,