Net Promoter Score (NPS) is one of the many attempts to make customer satisfaction scores more understandable. It is a single statistic like the mean or the percentage of happy customers. It is calculated by subtracting the percentage of unhappy customers (0-6) from the ”promoters” (10-9) ignoring the neutral (7-8).
This method has several flaws. It assumes that scores have a universally accepted meaning. It also ignores a lot of information in the answers. This is an important difference which the users of NPS do not seem to grasp. The original answers contain more information than the NPS which ignores the actual score and reduces it just to the range.
Why not use the simple mean? It is based on all the information in the answers, not just some percentages. The reason is that the values are not very informative. What does customer satisfaction of 8.0 mean, is it good or bad? Getting a NPS of 36 sounds better, or does it. If you think it does, you are mistaken.
Another problem with NPS is that it is unstable. Here are two cases with the same average of 8.0 but the NPS values are 18 and 36. It is far from clear which would be the more positive case. It could be that the populations just have a slightly different way of expressing their opinions.
It is also possible to have the same NPS but wildly different averages. A NPS of 50% can have an average between 7.5 and 9.0.