Net Promoter Score (NPS) is a popular measure of customer loyalty that many companies like EBay, PayPal and Amazon swear by to measure customer loyalty. Bonuses in many companies for top executives are also tied to this single metric, making it even more important come bonus season. Here is an explanation of Net Promoter score by Fred Reichheld from Bain & company who made a business case for instituting NPS for measuring customer loyalty in his book, The Ultimate Question.
Net Promoter Score (NPS) is based on the fundamental perspective that every company’s customers can be divided into three categories. Promoters are loyal enthusiasts who keep buying from a company and urge their friends to do the same. Passives are satisfied but unenthusiastic customers who can be easily wooed by the competition. And detractors are unhappy customers trapped in a bad relationship.
NPS is measured on a scale of 0-10, based on likelihood of recommendation: “How likely is that you would recommend company X to a friend or colleague?” Customers who provide a score of 9-10 are categorized as promoters, those who give a score of 8-7 are dubbed passive and those who give a score of six and below (6-0) are classified as detractors. Any company’s NPS score is equal to the percentage of customers who are promoters subtracted by percent of customers who are detractors. NPS=P-D
- Statistically speaking, it is highly biased towards detractors since anyone who scores between 0 and 6 is a detractor. Among the three categories, detractors have the widest range (0-6), while both promoters and passive have a range of two with promoters ranging from 9-10 and passive 8-7. In any introductory statistics or survey research class, one of the first things that they teach you is to have equal ranges while designing choices to a survey question, and to make room for a neutral response also. This ensures that respondents to the survey have an equal chance of picking each point of view (very positive, positive, neutral, negative, very negative). In other words, the probability of someone choosing a neutral rating is equal to someone choosing a very negative rating. This is not rocket science; it is just survey design 101. The problem with NPS is that it increases the probability of someone being classified as a passive compared to a promoter & even a detractor. I am sure proponents of NPS might argue that the bias is by design, but it compounds the second issue that I have with NPS discussed below.
- Different societies have different ways of evaluating quantitative metrics. When I did a cross-national studying looking at youth in India, China and the US, I found that young people in India and China are more stingy with their scoring: they are less likely to give higher ratings, and this extends to their own self-assessment. So someone in China is less likely to give a score of 10/10 then someone in US. Modesty is considered a great virtue in these societies and it extends to answering questions on surveys. There is the possibility that if someone in India gives a score of 8, according to him/her the service is great, but NPS will categorize them as passives.
- This brings me to my third point. It is extremely important to label scales. I have seen most quantitative market researchers ask respondents to choose one option on 0-10 scale, and when they want to look at top responses, they just look at top two boxes and classify that as extremely/very important. Honestly, in my mind 9 classifies as extremely good (yes, I did grow up in India and am less generous while providing a rating), whereas for someone else 10 will be extremely good. I strongly believe that it is important to label scales to eliminate response bias and ensure that respondents are interpreting the question in the same way.
What is my advice to companies?
If you want to continue using NPS in your organization, rescale the metric to ensure that it is not negatively biased, provide labels to ensure that all the respondents are on the same footing and lastly segment your customers to account for regional and other disparities like customer engagement.
