Is the NPS a trustworthy performance measure?

Abstract

Purpose – In 2003 Reichheld published an article in HBR, in which he claims that the net promoter score (NPS), is the only number you need to grow, and the only number you need to manage customer loyalty. The purpose of this paper is to demonstrate that the NPS is inferior to the standard measures of loyalty used by the American Customer Satisfaction Index (ACSI) and EPSI rating. Design/methodology/approach – In 2006 a customer satisfaction survey of the entire insurance sector in Denmark was conducted. The survey design was based on the questionnaires from EPSI rating and ACSI supplemented with insurance-specific questions, consumer sentiment questions and the basic Net Promoter Question. The sample consists of approximately 2,000 observations. Findings – The analyses presented in this paper show that the NPS it not what it claims to be: the one number you need to grow. The NPS is found to be a very poor predictor of both customer loyalty and customer satisfaction. The measure is very sensitive to changes in the underlying distribution, and finally the precision of the NPS was found to be low compared to other measures of loyalty, and it is not possible to predict the NPS categorization and hence it is hard to say precisely, how organizations can influence corporate growth based on the NPS. Research limitations/implications – The analysis is only conducted on data collected in a Danish business-to-consumer setting. More research is needed to shed light on the performance of the NPS across cultures as well as in a business-to-business setting. Practical implications – The paper demonstrates the dangers of using the NPS as an input to managerial decision making. Organizations are far better off using a standard customer loyalty measure such as those employed by the ACSI or EPSI instead of the NPS. Originality/value – Previous studies of the NPS have not replicated the methodology directly. Either there have been differences in scale length or in wording. The authors have constructed an experiment in the Danish insurance industry that answers some of the questions concerning the NPS without the shortcomings that most of the previous studies have suffered from.