1Â°) HR Big Data means nothing. Big Data is only the engine while the value depends on the output. What matters in these case is analytics and how they capture patterns to predict the impact of a decision in order to make better ones.
2Â°) It’s far from being widely adopted. It’s already complicated for many business lines and functions because it questions lots of assumptions so you can imagine what is likely to happen with more conservative functions.
Example. Earlier this year I was talking of a large change management plan for a very large business. An infinite number of HR data made it possible to finely segment the whole population to design an action plan. The point was to choose the criteria to use. Which were relevant regarding to how a given population will react to the change in question. We could think that people will react differently according to their age, their role, job, are using a computer at work or not. And that’s the problem. “We can think that”.Â But if a young person may react in a way and someone not equipped with a PC in another, what will happen to not equipped young ? It’s also possible to start with things observed in other companies but what are their value in a given organization and culture ? Nothing.
A decision that’s not backed with data is nothing but approximation.
Bottom line : we could have spend weeks working on segmentation critera, segmenting in many ways, the lack of data and correlation made that, in the end, all the approach would have lost its rigour because of one and only step.
Truth is that most of times we make decision based on our judgement and not on tangible data, even when they exist, and the result is an infinite number of bias impacting the quality of our decisions. When confronted by data analysis it appears that many assumptions we used to take for granted and back our decisions with were wrong.
If more and more functions are starting to use data for decision making, mostly by using analytics, HR seem to be lagging. That’s what Applying Advanced Analytics to HR Management Decisions: Methods for Selection, Developing Incentives, and Improving Collaboration by James C. Sesil is about. It shows us why and how HR should use this kind of approach and technology
The author begins with explaining the different thinking and decision making processes that exist and what analytics can improve. Moreover, let me precise the different kind of analytics that exist :
– descriptive analytics : what happened, what’s happening
– predictive analytics : what will happen or probably happen
– prescriptive analytics : what we should want to happen.
But be careful. As I always repeat when discussions come to this matter, the point is not to replace humans with machines but to provide humans with the needed information to make fact based informed information, backed with tangible things.
Then follows an unexpected but very interesting part on collaboration and collaborative decision making that highlights the need for more information sharing so decisions won’t be made on a fragmented view of the matter. That’s something technology can fix…provided data are accessible.
Then comes the much expected part : in which fields can analytics help to make better HR decisions.
In hiring 87% of decisions are biased
Of course it starts with hiring and talent management. There’s not a field with more biases. In the end, the decision made after a job interview is everything but rational even if the recruiter is not conscious of the biases he brings into the process. As shown in this Harvard Business Review article, from 85 to 97% of the decision made by professionals rely on their intuition when it comes to assessing a person. Made blind by a quality that masks a fault, a real quality that is useless for the position in question, by the intuition this profile is not made for this job….
I paid lot of attention to many things that were said in the Kenexa sessions during the last IBM Connect conference. Nothing is more difficult that managing atypical profiles and make the most of them profiles because it’s so easy to decide that a given position needs a given profile and no one else can be successful. At the scale of a company one can identify that a given kind of profile is often successful in positions they were not made for but these cases are so few that there is not enough volume to draw a conclusion. On the other hand, when a system is used by thousands of companies for hundreds of thousands cumulative employees, it’s easy to benchmark how likely a given profile is to succeed in a given position. Volume brings relevance.
Considering this point I can already hear those who will fear to rely on analytics. My answer is :
– data are data and they don’t lie. The fact you don’t loke what they say is another point.
– as I wrote above, technology is not making decision. Humans helped by technology do.
– regarding to hiring and talent management, this is a big anti-discrimination weapon. Assessment based on profiles and results, no halo effect or any other bias that prevent talented people to get the position where they would be successful and happy.
Analytics Vs. Received ideas
Analytics also help to assess the impact of a compensation package on employees performance in order to avoid counterproductive policies. Do you think that stock options boost the performance of executives ? Data show you’re wrong. Analytics also help to identify the jobs that contribute the most to customer experience and profitability. As I learned by AMC Theatres during a speech they gave, they realized that the most important people in a theatre regarding to revenue are…those who sell popcorn and other kinds of foods we eat while watching a movie. (Here’s a short case study and there’s also a good article on Forbes). Analytics also help to measure engagement and identify what drives employee engagement.
In short that’s a book that is sometimes a little bit theoretical (because it introduces many concepts that are new to many people) but makes us aware that HR won’t avoid the data revolution that is already at work in fields like marketing and customer relationship. Hoping they won’t miss the window as they did many times.
Another solution can be keeping on making decisions in the fog, haphazardly, based on biased reasonings and assumption.