As many people know Net Promoter Score is a customer feedback measure calculated from the response to the question: "How likely would you be to recommend [us] to a friend or colleague". The response is a discrete value from 0 [very unlikely] to 10 [very likely].
The responses are then used to calculate NPS:
NPS = % of 0-6 responses less % of 9-10 responses.
In an effort to understand what drives NPS there is a desire to identify high correlations and models that have NPS as the dependent variable and other business variables as independent variables.
Now, although NPS looks like a continuous variable it is calculated from Ordinal values (ranges of discrete values). So can it be treated as continuous for the purposes of these calculations? If not, why not and what are the alternatives.
Best Answer
In my experience the rate of change is different between promoters and detractors depending on the level of the independent variables. Basically, customers tend detractors when they have a really bad experience on an the independent variable where as promoting a brand is accumulation of good experiences.
Basically the whole scale is not always linearly affected by independent variables.