The NPS Metric Shifts Downward with a 5-Point Scale
In my last post, we looked at the correlation between 5pt and 11pt scales over time. We saw that they were very well aligned in how they changed together. However, an average score is not a complete picture. A casual observation of the frequency chart suggests a non-normal (non-bell-shaped) underlying distribution. As a reminder, these were responses to the questions:
- How likely are you to recommend this company to a friend or colleague
(0 = Not at all likely . . . 10=Extremely likely)
- I would recommend this company to a friend or colleague
(Strongly Disagree . . . Strongly Agree)
Between the two scales, you might expect 9’s and 10’s to match well with ‘Strongly Agree.’ However, with this data set, the percentages are a lot different; 43.8% responded 9/10 and 35.2% responded ‘Strongly Agree.’ This results in a significant reduction in the overall NPS score when using a 5pt scale.
For many companies, the Likelihood to Recommend question is used to calculate an NPS. For some companies, the NPS is the one number they need to grow, and other statistics are secondary. A standard mapping of scales onto NPS categories is below.
An NPS Score is calculated by the percentage of Promoters minus the percentage of Detractors. One would expect the NPS score to be the same. In our data, the NPS from the 11pt question is 10.2 points higher than the NPS from the 5pt question. The 5pt scale appears to have a downward bias on the NPS scores. In other words, Promoters decreased from the 11pt scale to the 5pt scale while Passives and Detractors increased.
11Pt NPS Score: (43.8%-22.7%)*100 = 21.1
5Pt NPS Score: (35.2%-24.3%)*100 = 10.9
Practitioners should be aware, when benchmarking or when simplifying a scale, that there may be a systemic reduction in scores which may be solely attributable to the scale.
Kyle LaMalfa is Senior Business Insight Analyst at MaritzCX.