WTF??!!! POLLSTERS WITH THEIR “PREDICTIVE ANALYTICS” GOT THE ELECTION ALL WRONG – WHAT IT MIGHT MEAN FOR HR AND HR TECHNOLOGY.

        As clearly evidenced by the Trump election and the fact that a vast majority (if not all!) of the pollsters who predicted a different outcome were wrong. With all their algorithms and underlying BIG DATA points – they still got the outcome wrong. The issue, in my humble opinion, is simple – nothing is easy or straightforward when trying to predict human behavior. In my sphere of influence, that of the HRMS and HR’s use of technology, we seem to be (up to Nov. 8 at least) enamored with the concept of “predictive analytics.” So, one issue that deserves some thought – does HR’s utilization of HRMS delivered predictive analytics to bolster and support actionable strategies regarding decisions affecting the workforce – globally or even individually – have a strong basis for validity? No less reliability? Certainly, many HR technology software providers have, in the last few years created and delivered thoughtful and reasonable workforce metrics. Those almost “out of the box” reports have added tremendous “Clout” to the role of HR in an organization. The software puts in the hands of HR Technology executives and HR C-level executives the opportunity to move from “Data Management to Information Craftsmanship” – providing future projections of workforce staffing levels and strong indications of the pipeline of talent, among other types of forecasts, in keeping with many well known lists of HR metrics. Models and forecasts for staffing and performance outcomes based on incumbent performance have proven to have a basis of validity and reliability, as much research has already shown. Executive and Management dependence on an effective HRMS delivering metrics and dashboards is certainly is needed and justified and would prove effective in a majority of cases calling for deep dive and a view of the underlying raw data that can enable cause and effect analysis. More recently, a few of the top tier comprehensive HRMS providers and the “best of breed” niche functionality providers in talent management suites have begun to emphasize their concept and definition of “Predictive Analytics” which seemingly attempts to incorporate behavioral analysis added to the standby standard metrics related to HR management. However, the approach must meet the challenges of quantifying human behavior, and must incorporate additional aspects and applications of social media, beyond the input of baseline data elements. I think the HRMS provider community will endeavor to do this, but there are some caveats. Specifically, with regard to the behavior of people in any company’s workforce, as we have seen, human beings are under many influences – some stated and some left or kept private by any one individual. So predictions – for example – the “9 box model” of predicting which employees are prone to leave (especially those that are considered “key”) – a popular and often “out of the box” delivered metric by a good many leading HRMS vendors, must, or might, or should be viewed (and acted upon) with some skepticism. Is the report truly reliable and useful? Are we relying too much on statistics based on data, instead of listening to our “gut” and observing behavior about specific individuals?. It could be so. Could taking action based on such predictive “models” end up overtaking other important inputs? Again, it could be so. Larry Acton, in his “Under 30” blog (Forbes – 11/18/16) “Can you quantify your Human Resources Department?” points out that HR poses a unique problem in the field of business analytics because its bottom-line goals involve a degree of subjectivity and because not all employee actions and behaviors can be easily quantified and humans behave, well, like humans, even lying when asked their opinion (or, how they voted in exit polls – as we found out,) not to generalize – but still. HR executives and managers must consider, and take advantage of all points of intersections between any employee and his/her manager. That would include such standard input resulting from frequent communications and exchanges. Exchanges begins with some formalized interactions – usually a performance review process. But by no means should that be the sole dialogue, and it would be a mistake to think that any exchange of future “engagement” on the part of the employee is fully candid or even truthful. How HR Can Currently Be Quantified Moving into 2017, HRMS providers will undoubtedly seek to integrate even more Organizational Behavior and Industrial Psychological aspects into their efforts in delivering more meaningful, and selective Predictive Analytics for which they have the underlying data. Here are 3 areas of HR measurement that may become potential breakthrough metrics to be seen shortly – integrated within Talent Management and general HRMS...