As well as providing everything you need for a Key Drivers analysis, our CCR methodology enables us to analyse the underlying components of a key drivers problem.
The components are composite factors produced using the on-line predictor scores and tend to be in low dimensions with typical problems producing 2-4 factors, making these ideal for plotting on scatter or bubble charts.
Similarly to factor analysis, these can be rotated to enable interpretation. Understanding how to describe these factors in terms of the underlying predictors gives insight into:
- The main dimensions driving the Dependent Variable (e.g. Likelihood to recommend brand)
- Differences on these dimensions across key subgroups in the sample. For example, if observations are based on multiple brands we can plot these brands by their factor scores.
- How you might create a Key Performance Indicator, based on these dimensions.
As CCR is applicable to any family of modelling, we can apply this in logistic and ordinal settings as well as for standard linear models.