Predictive Analytics - what are they?

The simplest answer is a prediction of what might happen based on what has already happened.

To support the predictions made on this site, we've analyzed hundreds of decisions where Canadian courts have expressly considered the Bardal Factors in determining what period of notice an employee should reasonably be given on termination of their employment. This analysis made clear that despite there being only 4 Bardal Factors (age, length of employment, character of employment and availability of similar employment), courts routinely consider additional factors and quite frequently disregard, or at least don't report on, one of the factors. Often influential are things like inducements that led the soon-to-be-unemployed person to leave secure employment to take the job they are now losing. Often ignored is the question of availability of similar employment. There is an art to this, but also a science. A simple severance pay calculator doesn't do justice to either.

We looked at the results of all cases, and the inclusion or exclusion of factors to develop predictive models that would allow us to anticipate how a future court might deal with a particular fact scenario.

It's pretty good, definitely useful, clearly superior to simple severance calculators, but still far from perfect. For example, predictions are constrained by the available data. So predictions about particular employment groups in specific provinces, might be stronger than others where the available data is lighter. The results are reflected in the breadth of the predicted range and the confidence expressed in the prediction.

This site offers the visitor an opportunity to discover the influence of hundreds of cases on the development of notice periods. That smooths out the human factor...but only to an extent. After all no human involved in settling, litigating or adjudicating a wrongful dismissal dispute is bound by algorithm.  Take heed of our favourite quote on this topic: