Know-How Exchange : Research/Metrics

Hi,

I find nothing wrong with the math of the NPS calculation: you can substract any two numbers from each other even if 'the subtrahend is not part of the minuend'. You are right though that it is a bit confusing to express all numbers in percentages; for instance, when there are more detractors than promotors, the NPS% will be negative. That is why I would prefer numbers.

Question is whether this difference is meaningful, and many believe it is. In this case, the end result is simply telling you whether you have more promotors than detractors. That would be similar to counting who voted for Obama and who voted for Bush, and then subtracting them to come up with one number that tells you who got the most votes.

It really isn't an equation relevant to math logic - more an assessment (somewhat controversial) at any given point in time that reflects a measurement of your standing in the marketplace. As indicated, there is controversy behind it, and at the very least, it should be a measurement tracked internally over time.

The logic is sound. Minuend and Subtrahend are each groups that represent sub-sets of the same set which we could loosely call "the market".

They are different parts, but they belong to the same overall set (market, survey audience, respondents, call it what you will).

所以如果我有十个πeces of fruit, five of which are apples and three of which are oranges, the difference between the numbers of apples and oranges is 2 (5-3=2).

Two what?

Two fruit, that's what the set was made up of.

Hope that helps.

Sometimes it's best to accept the logic and just go with the flow!