Is the Net Promoter Score a Reliable Performance Measure?
Kristensen, K. & Eskildsen, J. (2011)
This paper examines the relationship between the NPS and customer retention in the insurance industry in Denmark.
The authors don’t look at revenue or sales growth, but rather correlate the 11-point NPS to 3,400 respondents’ stated intentions to continue using their current insurance provider, from 1 = No, Absolutely Not to 10 = Yes, Absolutely.
They use five criteria defined by Kristensen and Westlund (2003) and claim that “a 10-point scale is by far the most efficient” compared to an 11-point scale (no citation was provided to support the claim).
Validity: They compare responses to the LTR item to “Will the company still be your insurance company next year?” 1 = No, Absolutely Not to 10 = Yes, Absolutely.
While there is a strong correlation with stated intent, the authors found their clusters didn’t align with Reichheld’s designations of promoter (9–10), passive (7–8), or detractor (0–6). They found three clusters—0–4 for detractors, 5–7 for passives, and 8–10 for promoters. Curiously they use this as proof that NPS “fails the validity test.”
Reliability: Authors report on the commonly cited issues of transforming the 11 points to 3, increasing the margin of error and reducing the signal in the noise. While true, as we’ve reported, the loss might be worth it( see also de Haan et al., 2015).
Robustness: The authors through a simulation study argue the S/N ratio of the NPS is lowest for a mean value of the underlying distribution between 65 and 75, which in our data are relatively high scores.
The authors conclude dismissively, “The best we can say about NPS is that it is a mistake! We hope that companies will realize this before too much harm is done” (p. 253).
Takeaway: The authors found different clusters for promoters, passives, and detractors when correlating the NPS to stated intent to continue as an insurance customer. Even though this deviates from Reichheld’s correlation to revenue growth, they dismiss the NPS as a mistake. The authors use very strong language but did not make a very compelling case with their data.