Predicting the Impact of Change
Written by Sian
Our clients often find themselves sitting on a lot of powerful insight, with a clear view of where efforts need to focus to drive CX transformation. The insight is often prioritised by the importance of different service elements in driving a key metric i.e. NPS. Another key evidence point in the CX transformation journey is sharing what impact the recommended actions will have on experience, specifically, by how much will they move a key metric.
Once a view of the customer journey has been established the key next steps are:
What next? – the ‘now what’
What impact will improvement have?
We’ve got a tool to help you predict how a change in an element of service will move the main metric – The Impact Predictor.
It takes your insight and shows you where to focus your efforts.
The mechanics
At the foundation of The Impact Predictor, is Key Driver Analysis and specifically, regression modelling. We use regression as it allows us to consider the fact that consumers do not experience one element of service in isolation, it is a mixture of several elements. Regression understands the interrelationship between all the service elements measured – the multi-collinearity.
Regression means we can identify:
The most important drivers of a key metric
The relative importance of each service element
How effective the survey content is – are you asking the right questions.
We then use the modelling to predict the impact of change, by:
Regression provides a number that shows you how much of the key metric is explained by the service elements measured – the R2 figure. We use this to create a proportion for each service drivers and how much they explain the key metric.
Then we take the key metric score and chop it up, giving each driver a part it is responsible for.
We then take the performance scores for each driver and multiply it by the amount of the key metric scale it is responsible for, and multiply the amount of the scale it is responsible for by it’s score.
Finally, we sum up the scores, and subtract from the key metric score, to find out how much of the key metric score is down to the things we do not know about.
What we get at the end
The Impact Predictor allows all to know how much each driver contributes to a key metric, we can then change the score for a service driver and get the new key prediction. It is easy and quick to use and can be used offline or built into a dashboard.