Many companies like the idea of using predictive analytics to assess flight risk (the probability that an employee will quit). The problem is that most organizations don’t have the skills to do this kind of analysis, and even where they do it can be expensive. The solution: tech vendors.
An HR analytics vendor like Visier spends all it’s time working on analytics so it’s no surprise they’ve built an engine to predict flight risk. They can use this standard engine, look at an individual organization’s data, and tune the algorithm for the specific case.
The reason I highlight this is that the big tech vendors are in a much better position to do advanced analytics than all but the most sophisticated companies. Visier has data aggregation rights (with suitable privacy protections) for 3.5 million employees. No individual company has that much data.
If we look at a topic like career planning, big vendors like Workday and SAP SuccessFactors can potentially look across vast numbers of companies to see how people move through various career paths. It makes more sense for the big tech vendors to “solve the problem once” (and solve it better due to access to a big data set) than have individual companies tackle the problem.
It’s also the case that home-grown analytics solutions may get most of way there but struggle to keep their solution up-to-date. For flight risk, as a matter of course, Visier’s software frequently looks back and compares its predictions to what actually transpired; so there is continually validation of the algorithm. That kind of polish can be hard for in-house analytics teams to match.
What is interesting
- Visier uses a machine learning technique called “Random Forest.” The technique comes up with large numbers of models for making predictions, runs simulations of those models to get predictions, then compares those predictions to the actual outcomes.
What is really important
- We are going to see more and more great analytics solutions coming from tech vendors, they have the focus, the specialized resources, and the data to do a better job than most individual companies.
- While the vendors will do the heavy lifting, companies still need to understand what they are buying. You can’t take a vendor’s word for it that their making appropriate use of AI/machine learning or whatever the latest claim is.