This technique unpicks the relationship between each one of a set of variables (independent variables) and an overarching measure (the outcome or dependent variable), such as customer satisfaction or brand usage.
What are the aims of Driver Analysis?
The outputs of driver analysis establish the degree to which the headline measure is driven by each contributing factor, allowing businesses to prioritise efforts on the things which matter most. Therefore, whilst driver analysis is not predictive analysis it can be used in making strategic predictions.
How is Driver Analysis undertaken?
There are many techniques for undertaking Driver analysis – here are details of the 2 we use most frequently:
Multiple Linear Regression is one of the most commonly used techniques. Here, drivers are calculated by looking at the correlation between independent variables, and attempts to find the optimum combination to predict the dependent variable. The degree to which the model describes variance in the dependent variable is given as an R-squared value.
Shapley Value Regression is becoming more popular as an approach. It is based on Game Theory, which was developed by Lloyd Shapley in the 1950s. He developed this as a way of allocating profits to actors who had all contributed in varying proportions to a particular event. It has now been adopted in market research as an effective way of calculating relative importance when there is a lot of collinearity between attributes (this very often being the case with attributes on a research survey). In simple terms, it identifies what proportion of R-square from a linear regression model can be attributed to each independent variable. The benefit of this method is that outputs are stable and can be expressed in a way that sees them add up to 100%; making them easy to communicate/understand.
What's an example of Driver Analysis?
Creating a 2x2 matrix of importance vs. current delivery is a useful way of considering driver analysis results. This allows you to see which aspects to grow, protect, soft peddle or ignore.
What is Driver Analysis?