Net Promoter Score

The Net Promoter Score (NPS) is an index intended to measure the extent to which customers will recommend an organization.^[1] It is used by service organizations as a way of monitoring their performance between different employees, different departments and over time.

How the data is collected

The data required to compute the NPS is obtained from the following survey question (which has lots of variations):


Q1.	Based on your experiences for this INSERT CATEGORY, how likely are you to recommend INSERT COMPANY NAME to a friend or colleague?
	
0 Not at all likely
1
2
3
4
5
6
7
8
9
10 Extremely likely

Computing the NPS

Computing from tables

The categories of the question are collapsed into three groups:
- Detractors: ratings of 0 to 6.
- Neutrals (or passives): ratings of 7 or 8.
- Promoters: ratings of 9 or 10.
The NPS is computed by calculating the proportion of Promoters and subtracting the proportion of Detractors. In the case of Microsoft, using the data below, the NPS is:

[math]\displaystyle{ NPS = (7.1 + 8.7) - (7.1 + 3.8 + 4.8 + 3.5 + 4.8 + 16.0 + 9.3) = 15.8 - 49.4 = -33.6 }[/math]

File:LikelihoodToRecommend.PNG

Computing by recoding

An equivalent way of computing the NPS is:

Recode the data as follows:
- Values of 0 to 6 are recoded as -100
- Values of 7 or 8 are recoded as 0
- Values of 9 or 10 are recoded as 100
Compute the average of the resulting variable.

Interpretation

Lots of different things are known to influence an NPS score. In particular, although it is a measure of a firm's performance, the NPS score is determined by numerous factors relating to how the data is collected, including things such as: instructions given in the NPS question (e.g., whether it asks about the most recent experience, general experiences or a particular aspect of a firm's performance), the expertise of customers, how the survey was administered (e.g., phone, face-to-face, email), the order of questions in a questionnaire.
Although it is called the 'Net Promoter Score' the actual NPS number that is computed does not directly relate to the proportion of people that promote versus detract. That is, while a firm with a good NPS score should expect to have more people saying good things about it than a firm with a poor NPS score, it is certainly not the case that a specific score indicates a specific level of promotion of a brand by users. Similarly, there is no straightforward relationship between the NPS score and other behaviors of interest (e.g., churn, satisfaction, profitability).
The three categories that are used to compute the NPS – detractor, passive or neutral and promoter – are arbitrary. And, for the same reasons described in the previous two points, you should not expect that everybody in the promoter category will promote the brand or that all the detractors will detract from the brand.
The NPS is a relative score. It is useful for comparing performance over time. It is useful for comparing different geographical territories (e.g., performance in one state versus another. The key is always to look at an NPS score and assess it relative to another comparable NPS score
Although the categories are arbitrary and the NPS score itself is consequently also arbitrary, there is no evidence that this arbitrariness reduces the validity of the measure. For example, if the detractors were classified as people with scores of 0 to 5 and passives 6 to 8 you would still anticipate getting the same basic conclusions regarding relative performance.

References

Template:Reflist

Articles on the Displayr blog

The Basics

Analysis and Visualization

More Information

↑ Reichheld, F. F. (2003). "The One Number You Need to Grow." Harvard Business Review(December).

[1] Reichheld, F. F. (2003). "The One Number You Need to Grow." Harvard Business Review(December).

[1]