Net promoter score, growth, and profitability of transportation companies

Strategies of many contemporary companies are almost entirely focused on consumers. These companies, with the use of, inter alia, surveys or focus groups, constantly put their consumers under a microscope, aiming to understand their behavior. This is, however, not a straightforward task. Typically, companies tend to look at satisfaction measures, market shares, or sales figures. According to research undertaken by Bain Company, there exists a certain body of consumers, who, despite providing scores indicating “satisfied” in surveys, soon move to competitors. Analysis of the information stemming from market share, although realistic, can temporarily mask significant trends due to heavy spending on promotions or advertisements. Additionally, both market shares and customer satisfaction surveys, according to Blasberg et al. [2008], look backward – not forward; accordingly, these metrics provide little or no information on future consumer behavior. The authors [2008, p. 16] further recommend the companies to look for more advanced tools for examination of consumers’ contentment, in particular, in the loyalty area. According to Anderson et al. [2014], due to the commercial omnipresence of Internet customers, loyalty is harder than ever to achieve. Although customers’ loyalty has been a subject of many studies, there is no consensus on the perfect loyalty measure. One of the recent and modern customer loyalty measures is the net promoter score (NPS). This score is significantly reliant on promoters (i.e., people willing to recommend a product or a company to a friend or a colleague) and detractors, being the opposite [Reichheld, 2003]. Opinions, i.e., word of mouth, of clients are widely acknowledged in the literature as a mean of the client’s influence (positive or negative) on organizations [Lotko, 2012]. The value of word of mouth is generally considered in the literature to be significant, because it costs the retailer nothing, it can be immediately communicated through modern channels, such as the Internet or mobile phones, and it has an immediate sense of credibility [Grewal et al., 2003]. Moreover, it should be noted that pursuant to considerable online dialogue during social media development, word-of-mouth marketing has become increasingly important [Deszczyński, 2017].

Considering the significance, complexity, and difficulties of loyalty measurement, followed by the recent [Reichheld, 2003] proposal of the NPS index as an efficient tool for loyalty measurement and management, the objective of this paper is to verify whether the stated index has an impact on a company’s growth and profitability. The relevance of the subject of the study is further supported by the high popularity of NPS among companies and the split opinions presented in the literature, i.e., some researchers endorse the NPS, whereas others have criticized it. In this research, the impact of NPS on the growth and profitability of Polish transportation companies has been studied.

The paper is organized as follows. In the first section, the characteristics of customer loyalty are provided. In the second section, the NPS index is described and discussed. In the third section, the criticism of the considered index, based on literature review, is presented. Based on the work done in the aforementioned sections, hypotheses are developed. In the following research section, key figures concerning the selected study transportation companies are provided. The research follows with descriptive statistics of the studied variables. In the next step, normality tests are performed, based on which, the relevant correlation tests of significance are conducted to verify the hypotheses developed. The hypotheses are also verified with the use of linear regression. Finally, the conclusions of the study, practical implications, and directions for further research are presented.

Customer loyalty has been the subject of many studies in the past decades, e.g., Tucker [1964], Jacoby and Chestnut [1978], Pearson [1996], Griffin [1997], Oliver [1999], Dziewanowska [2007], and Rahim [2016]. According to Pearson [1996], customer loyalty is defined as the customer’s mind-set concerning the company, its products, or services, which provides a favorable attitude toward an organization, a commitment to repurchase the company’s products, and a tendency to recommend the product or a company to other clients.

Loyalty is closely related to satisfaction, however, as it requires an action; it is, therefore, a larger term than satisfaction. According to Jones and Sasser [1995], loyalty increases together with satisfaction. The relationship between these variables is neither simple nor linear, and it depends on the industry. The higher the competition, the more important is the customer satisfaction. Authors claim that since customers are reasonable, if they are not completely satisfied and if they have a choice, they can be lured away easily. It is essential, therefore, to keep clients completely satisfied. Completely satisfied customers are, in turn, much more loyal than satisfied customers. They are, therefore, a key in generating sound long-term financial performance. The authors divide loyalty into true long-term loyalty and false loyalty. The latter results from several factors that make customers seem loyal, when indeed, they are not. These factors include high switching costs, strong loyalty-promotion programs, government regulations that limit competition, and others.

Reichheld [2003] defines loyalty as the willingness of someone to make an investment or personal sacrifice in order to strengthen a relationship. In the case of a customer, it can mean, e.g., sticking with a supplier even when a supplier does not offer the best price in a particular transaction, hoping for a good and fair value in the long term. Reichheld [2003] also notes that loyalty is not about repeating purchases, however. Someone might repeat purchases because of high exit barriers, indifference, or lack of alternatives. Conversely, a loyal client may not repeat purchases frequently because of the reduced need for a product. An important characteristic of a loyal customer, however, is that such a person talks up a company or a product to friends and family, thus promoting it.

The loyalty of clients tends to be very important through recessions. As indicated by Anderson et al. [2014], the strategy to invest in customer loyalty, as opposed to cost-cutting strategies, during significant economic downturns, although counterintuitive, was found to be superior in terms of the companies’ survival, while simultaneously gaining a long-term competitive advantage.

A significant portion of research has indicated the dependence of NPS or customer loyalty on the company’s growth or profitability [Anderson et al., 1994; Wansink, 2003; Blasberg et al., 2008]. Reichheld et al. [2000] state that a small increase in customer loyalty contributes to a high growth of profits.

The NPS was introduced by Frederick F. Reichheld in a 2003 article in the Harvard Business Review. The author initially criticized the use of customer retention rates, which despite substantial usage in business, were a poor indicator of customer loyalty and growth. Other conventional customer satisfaction measures, including the American Consumer Satisfaction Index (ACSI), were also not reliable, mainly because of their complexity, difficulties in interpretation, high preparation costs, and no overall significant connection with the growth of sales.

Reichheld [2003] noticed that the key to measuring customer loyalty was to give priority to the most satisfied clients. Delighted customers promote the company’s products to their friends, becoming – in effect – an unpaid marketing department of the company. Reichheld [2003] also noted that it is not only enthusiasts who count: the level of frustration and disappointment among customers of the company is also of utmost importance, as the company is likely to lose these clients, unless corrective actions are taken. The critical question in loyalty measurement is neither about satisfaction nor about loyalty, but about the willingness to recommend a product or service to a friend, colleague, family, or just someone else.

The NPS is calculated based on the survey responses of a likelihood-to-recommend question on a 0–10 scale. This question is “How likely is it that you would recommend ‘a company’ to a friend or colleague?” The rank of ten means the customer is very likely to recommend a company to someone, while zero means that the client is unlikely to recommend a company to someone, and five is neutral. Typically, this question is followed by several auxiliary open-ended questions, in which the companies ask consumers to describe the reasons for their score. In order to calculate the NPS, the proportion of respondents who rated the company ≤6.0 (Detractors) is subtracted from the proportion of respondents who rated the company 9.0 or 10.0 (Promoters). The difference represents a company’s NPS. Respondents who rated the company 7.0 and 8.0 are called passively satisfied. The division of customers into these three stated clusters was a result of the examination of referral and repurchase behaviors [Reichheld, 2003; Blasberg et al., 2008].

According to Blasberg et al. [2008], measuring a brand’s NPS and subsequent comparison of the results with the NPS of competitive brands gives some insight on both the rational and the emotional sides of consumer behavior. This is because NPS is supposed to indicate future consumer behavior [2008].

Pursuant to an issue of the Harvard Business Review containing articles on NPS, many articles appeared confirming its significance and benefits in practical application [Korneta, 2014; McGregor, 2006; Reichheld, 2006a, 2006b; Lotko, 2015]. The possibilities for the practical application of NPS in business environments were confirmed by the use of that score in the following, inter alia, prominent companies: Allianz, American Express, Procter & Gamble, and Amazon [Gajewska and Szewczyk, 2012, s. 413; Pomaskow, 2014].

Nonetheless, despite the considerable benefits attached to the NPS, shortly after that score had been introduced, criticism from other researchers has begun to appear [Kristensen and Westlund, 2004]. Grisaffe [2007] stated that because the NPS is calculated based on only two out of the three categories of customers (promoters and detractors), a significant portion of information is being lost. The author also raised considerable doubts regarding the cutoff points in the recategorization of the initial 0–10 scale. For instance, a response of 6.0, despite being above the midpoint of the scale, was placed among the detractors and, therefore, interpreted as negative. Grisaffe further questioned the sensitivity of choice of all the cutoff points. Finally, the author criticized the low precision of NPS, claiming that better, i.e., more precise, predictive power would be obtained with the application of the original 0–10 scale. The issue of NPS precision was also touched on by Seal and Moody [2008].

A significant portion of the criticism referred to the simplicity of the measure, stating that NPS is overly simplified and therefore cannot measure complicated sets of relationships [Brandt, 2007; Pingitore et al., 2007].

Keiningham et al. [2008] claimed that NPS is not a reliable indicator of growth and that it cannot be considered superior to other loyalty measures. Similar claims were also raised by Morgan and Rego [2008] and Kristensen and Eskildsen [2014].

Pollack and Alexandrov [2013], following a study based on 159 surveys, research, and an analysis of a set of pros and cons, recommended to include the NPS in a portfolio of customer metrics, however, not as a stand-alone diagnostic tool, as proclaimed by Reichheld [2003].

Given the considerable benefits of NPS presented in the literature, its simplified and low-cost approach toward description and interpretation of complex customer loyalty matters, set against significant criticism received from some researchers, the objective of the study is to verify whether NPS is, in fact, an effective loyalty measure. The study has been undertaken in the Polish transportation industry. The limitation in terms of studying only one country and one industry is introduced deliberately. This is primarily because the findings presented so far in the literature, from different countries and different industries, are often contradictory, i.e., some support NPS, while the others criticize it. The papers supporting NPS or loyalty often link high loyalty or NPS with growth or profitability [Jones and Sasser, 1995; Reichheld, 2003; Blasberg et al., 2008]. The paper aims therefore to verify whether there is a relationship between NPS, growth, and profitability of Polish transportation companies. In order to achieve this objective, it was hypothesized as follows:

The sample for this study comprises figures relating to 34 Polish transportation companies in three consecutive periods 2014–2016. The data used in the study were obtained from the following two data sources:

Because either not all the companies participated in the logistic reports in each of the considered years or the EMIS database did not contain the financial statements of the analyzed companies in each of the three studied periods, only 76 observations, instead of 102, were subject to analysis. Accordingly, as shown in Appendixes 1 and 2, the observations of 18 companies appeared in all three periods, the observations of six companies occurred only twice in all figures and so were used only twice, whereas the observations from the remaining 10 companies were used only once. It should be noted that no company, nor any observation, was intentionally removed from the studied sample, i.e., the data set is unbiased, while the removal of the most atypical observation would enhance the result of the tests undertaken in the study.

Growth is primarily considered in the literature as the growth of sales [Glancey, 1998]. In this study, however, the growth is also measured as the growth of earnings before interests, tax, depreciation, and amortization (EBITDA) and the growth of net results.

In the case of H2 verification, return on assets (ROA), return on equity (ROE), and return on sales profitability (ROS) are selected as the profitability measures. Each of the stated ratios is of particular interest to various stakeholders and is frequently used in studies in literature [Dawkins et al., 2007; Lyngstadaas and Berg 2016; Nauriyal, 2017]. ROE, however, is considered as the closest ratio to the shareholder value and, in that respect, it is considered superior to other profitability measures [Wright, 1975]. All the three above-described ratios are calculated using both the net result and the EBITDA; again, this is a common practice already seen in the literature [Dawkins et al., 2007]. The variables used in the study are described in Table 1.

As indicated above, the studied sample consisted of 34 companies. The studied sample comprised the most prominent logistic companies, e.g., three DHL companies, DPD or FM Logistic; principal national mail operator Poczta Polska, as well as several minor companies. As a consequence, the mean sales revenue of the studied companies varied highly from several million PLN to > 5 billion PLN. More details on the mean values of the selected variables of the studied companies are shown in Appendixes 1 and 2.

Table 2 contains the descriptive statistics of the variables used in the study. Mean of 54.4, standard deviation (SD) of 24.8, minimum of –0.20 (negative), and a maximum of 100 of NPS suggest great differences among the companies within the studied period. Negative minimum values of variables, resulting from negative results in terms of both net result and EBITDA, or a higher proportion of detractors than promoters in the case of NPS appeared within the studied sample of observations. Mean growth of sales revenues of 0.18 can be considered high; however, a median of 0.09 points out that half of the companies grew slower than 0.09 per year. The highest mean profitability of shareholders’ funds (ROEE) measured with EBITDA amounted to 0.73, while when measured with net result, it totaled 0.3, confirming that the industry is, in general, highly profitable. A wide range of variations among the profitability variables is apparent in Table 2, as represented by their minimum and maximum values.

Since a normality assumption is a key assumption in the majority of statistical tests of significance in the very beginning of an empirical study, all variables were subject to verification that they follow a normal distribution pattern. This is mainly because almost all the statistical tests are developed by keeping a normal distribution in mind. And so, e.g., the frequently used Pearson correlation coefficient requires that a set of data upon which a statistical test of significance is being undertaken must either exactly, or at least approximately, follow a normal distribution. Since not all the normality tests are equal nor are suitable for all the situations [Siddiqi, 2014, pp. 290–291]; in Table 3, the results of four different normality tests are provided. All the four statistical test have the same null hypothesis, i.e., the data are normally distributed.

From the data shown in Table 3, it is apparent that the majority of variables have no normal distributions, assuming a p-level of 0.,05, i.e., none of the variables have a p-value >0.05. Therefore, the null hypothesis that the data are normally distributed must be rejected. It should be noted, however, that the null hypothesis could be accepted with a p-value of 1% in the case of two variables: NPS and ROAE.

With respect to the results of the normality test presented in Table 3, the decision was made that Spearman’s coefficient of rank correlation is the most appropriate choice for further research. This statistics is a nonparametric correlation based on ranked data and not on the actual data. This test is frequently used in the literature for ordinal or interval data and when normality assumption is not being met [Brau et al., 2009, Chan et al. 2009].

Table 4 reports the Spearman and Pearson correlation coefficients; however, as the normal distribution assumption is not met in the studied variables, the most appropriate is the Spearman correlation coefficient; hence, the Pearson correlation coefficients are shown only for indicative and comparison purposes. The results of the estimations confirm the statistically significant and positive relationship between NPS and the three profitability variables ROA, ROS, and ROE – measured with net result – and one profitability variable ROEE – measured with EBITDA. The highest coefficient of 0.3646 referred to ROE, indicating the most significant relationship between NPS and that particular variable.

Next, the developed hypotheses were verified with the application of linear regression methodology. Table 5 provides the results of the regression coefficient calculations. Since none of the NPS coefficients have a p-value <0.1, the results are aligned to those obtained through analysis of Spearman correlation ranks, i.e., the relationship between growth of sales, net result, or EBITDA and NPS is econometrically insignificant.

The results of the regression calculations between the profitability measures and NPS are shown in Table 6. The results of regression analysis endorse the findings obtained through Spearman correlation ranks, i.e., confirmed the positive and econometrically significant relationship between NPS and profitability measures. The results are, however, slightly different in case of selected profitability measures. Regression analysis confirmed the positive relationship between NPS and ROA and between ROAE and ROEE.

Accordingly, in due course of the research part of this study, the positive and statistically significant relationship between profitability variables and NPS was confirmed. Consequently, H2 was accepted. The empirical study has not, however, confirmed the impact of NPS on growth of sales revenues. Therefore, H1, stating there is a positive and statistically significant relationship between NPS and growth, was rejected.

Introduced by Reichheld in 2003, the NPS, due to its simplicity, the ease of interpretation of its results, and the assumed impact on future growth and profitability very quickly gained popularity among companies, including the most prominent ones. Although considerable research has endorsed the NPS concept, a significant proportion of researchers have criticized it. Due to the ambiguity of the indications presented in literature and the practical application of NPS by many companies, the objective of the paper was to study the impact of NPS on the growth and profitability of Polish transportation companies. In the due course of study, the Spearman correlation rank and regression analysis were found to be the most appropriate influence measures. Neither regression, nor Spearman correlation rank, confirmed the impact of NPS on the growth of the analyzed companies. This study confirmed, however, that NPS has a statistically significant and positive impact on ROA, ROS, and ROE profitability measures. This was confirmed by both regression analysis and Spearman correlation rank, which cumulatively confirmed H2. Therefore, the findings of the study can be considered material. The above-presented findings are aligned to those presented by Pollack and Alexandrov [2013], who recommended the inclusion of NPS in a portfolio of customer metrics, however, not as a stand-alone diagnostic tool, as proposed by Reichheld [2003].

The study, however, has several limitations, mostly concerning the studied sample. First, and overall, the size of the studied sample was relatively small. Although the findings are statistically significant with respect to the sample size, the number of observations could have been higher, so as to generalize the findings identified. Second, the studied sample comprised only Polish companies; therefore, the findings should not be considered applicable to other countries. Third, the study focused solely on transportation companies; as a consequence, companies from other industries should not rely on these findings. The second and third limitations have been embraced deliberately; given the dichotomy of the indications presented by various researchers, it was more convenient to focus on one country and on one industry, so the findings cannot be questioned in that matter. With respect to the first limitation, the key problem was lack of relevant bigger data. It should be understood, however, that NPS was introduced as recently as 2003. Accordingly, NPS is less mature as compared to other ratios; consequently, less information is available. On the one hand, from a scientific point of view, it would be better to wait with research until a bigger sample of data is being gathered; but, on the other hand, given that NPS is already working in businesses with full speed, proceeding with research comprising smaller samples of observations seems to be justified.

The aforementioned limitations can be, however, seen as an indication for further research since the question regarding the effectiveness of NPS as a loyalty measure in other industries, countries, and – overall – its presumed relationship with growth remains unsettled.