A Number To Help You Grow
The Net Delight Index (NDI) as a meaningful alternative to the Net Promoter Score (NPS)
If you have ever stumbled across a customer survey you have probably been asked the following question: “On a scale from 0 to 10, how likely is it that you will recommend a company to your family or friends?”. This measure is called the Net Promoter Score (NPS), and to say it is popular in managerial circles would be an understatement: at D-Sense we often find managers convinced that they just need the NPS, and not much else, to assess their company’s growth potential. This belief can be traced back to the intellectual father of the NPS himself: Frank Reichheld, who introduced the measure in a 2004 issue of Harvard Business Review, declared it The One Number You Need To Grow (Reichheld, 2004). And he meant that literally: as a one-item measure, the NPS is a lot cheaper to implement into a measurement routine than lengthy satisfaction surveys and, as opposed to those surveys, the NPS does not leave decision makers with a puddle of ambiguous results. Instead, it immediately shows whether or not the company is performing well: just ask the net promoter question, subtract the percentage of responses from 0 to 6 from the percentage of 9’s and 10’s, and there it will emerge. The God Particle of customer research; the only diagnostic you’ll ever need.
If we take Reichheld’s claims seriously, using the NPS sounds like a no-brainer: it is cheap, easy to implement, and much better at what it does than any of its competitors – so why doubt its use? One good reason is that the claims surrounding the NPS are a little too good to be true. Let me put it metaphorically: to a scientist’s ear, claiming you have found a single index that fully captures a social phenomenon sounds equivalent to saying you have proven the existence of a live unicorn. If it were true it would be absolutely extraordinary but, to paraphrase French mathematician Pierre-Simon Laplace (1878), we cannot believe anything extraordinary unless there also is extraordinary evidence for it. And Reichheld never came up with extraordinary evidence for the NPS. Various articles have even shown the NPS to lack in fundamental aspects of proper measurement (e.g., Keiningham, Aksoy, Cooil, Andreassen, & Williams, 2006). So however unfortunate, Reichheld never found himself a metaphorical unicorn. He found a regular horse, glued a cone to its head, and called it a unicorn. This has been sufficient to cultivate the illusion that the NPS offers something special but, arguably, illusions are not what decision makers are after. And this is exactly why D-Sense has been developing a more valid alternative measure – the Net Delight Index (NDI) – to replace the NPS in its surveys.
From NPS to NDI
Anyone with a background in scientific research knows how hard it is to actually develop a valid and reliable measurement scale. You start off by providing a clear and unambiguous definition of the concept you try to measure, develop dozens of separate indicators reflecting the concept, and conduct various studies to assess the mathematical properties of those indicators: you test whether they actually assess the same concept and investigate if the long list of original items can be reduced to a couple of key indicators. This whole process takes up a lot of time and effort, and there is no guarantee that the measure will turn out to be useful in the end. This is also why scientific standards are often frowned upon as unrealistic ideals in fast-paced commercial environments: to managers, measures just need to be SMART, useful, controllable, and understandable (cf., Merchant, 2006).
At D-Sense, we also see value in a pragmatic attitude toward measurement but we believe this should never threaten the conceptual validity – the meaningfulness – of a measure. The NPS, for one, was simply advocated by Reichheld because he found it to be related to business growth in ‘some’ areas. There were no conceptual reasons for its development, and Reichheld also never managed to clarify what it is that net promotor scores should actually be measuring. As Grisaffe (2007) observes, he makes several contradictory statements about it in his own writings: he simultaneously proposes that the NPS isn’t about customer satisfaction or even loyalty, that it serves as indicator of customer loyalty, that it is an outcome of customer loyalty, that it is a cause of customer loyalty, and that it measures both customer loyalty and satisfaction. Hence, it is clear that the NPS never went through the first fundamental step of measurement: it lacks reference to an underlying concept. It lacks a definition. It lacks meaning.
In contrast to the NPS, the Net Delight Index we use at D-Sense is clearly conceptualized. In fact, the NDI directly taps into the heart of D-Sense’s consultancy schemes: if companies want to obtain a commercial advantage they should never settle for customer satisfaction. Their actual goal should be to achieve a level of Customer Delight. This premise is based on findings in the literature showing that satisfaction, in itself, does not reliably predict customer loyalty and repurchase behaviors (Kim, Vogt, & Knudson, 2013); only at the extreme positive end of satisfaction scales effects are typically observed. Extremely satisfied customers are more likely to become emotionally attached to a company, and this emotional attachment provides them with a personal motive to spread positive word-of-mouth (Kwong & Yau, 2002). At D-Sense we have translated these characteristics into an explicit, one-line definition of Customer Delight: in our terminology, Delight is defined as the identified motivation to recommend a company to peers based on an extreme sense of satisfaction with the company’s products and/or services. To some readers this may sound overly technical, but understanding it is crucial as it uncovers some fundamental advantages of the NDI compared to the NPS.
First, the definition requires the NDI, as a measure of Delight, to always assess the existence of a motive to recommend. This means that there are two separate situations that can be counted toward the NDI score: the current intention to actively recommend, and past recommendatory behavior. The NPS, in contrast, only considers people’s intentions to recommend and ignores the value of actual behaviors. As studies have often reported large gaps between people’s intentions and their subsequent actions this approach seems rather naïve and could introduce a considerable bias (Ajzen, 1991).
Second, the D-Sense definition of Delight implies that a motive to recommend can only ever count as an indicator of Delight if it is relatively ‘identified’ or ‘intrinsic’ to the self (cf., Ryan & Deci, 2000). In other words, a recommendation needs to have its roots inside the individual; external causal factors do not qualify as part of a Delight experience. This is very different from the logic behind net promoter scoring. For example, assume that you do not actively go around and promote your internet provider but would certainly recommend it if someone asked you to. If you’d have to respond to an NPS measure, this willingness would be sufficient to be considered an active promotor – the top category in the NPS. The NDI, in contrast, would not count you as part of its top Delighted category as the recommendation was provoked instead of spontaneous. As a result, the NDI is a much more stringent metric than the NPS! This is also illustrated by the fact that even spontaneous recommendations will not suffice to speak of delight. After all, there always needs to be an extreme sense of satisfaction involved. This means that if you spontaneously recommend an, in your view, mediocre internet provider to a student with a heavy loan ‘because it is cheap’ this cannot count as an instance of Delight. Discarding these types of recommendatory behaviors has a considerable advantage: it filters out any recommendation that is not intrinsic to the self and may also contain some negative or ambivalent information (e.g., ‘it’s not a good provider but having bad internet is better than having no internet’). What’s left, then, is a recommendation rooted in purely positive sentiment. An instance of delight.
The NDI scoring system
In order to take all aspects of our delight definition into account the NDI measure is composed out of two distinct items: an overall, 5-category Customer Satisfaction item (CSAT) and a modified version of the net promotorship measure. Across our studies, we typically find the two to be moderately correlated and mathematically reducible into single index (a so-called ‘principal component’ explaining about 60% of the spread in its constituent parts). Table 1 visualizes what this combination looks like: colored in green are the categories fitting the definition of Delight – they contain ‘Delighted Customers’. The red categories contain those customers that are far from achieving the characteristics of Delighted customers on either or both indicators (i.e., comparable to what the NPS calls detractors). The yellow categories fall somewhere in between. Customer in this category do not fully fit the definition but may grow to become delighted if properly managed.
Table 1. A representation of the NDI scoring grid.
The NDI, then, quite simply reflects the difference between the percentage of customers in the green and red categories centered around a value of 100. Any score above 100 shows that the customer base consists of more delighters than detractors. Any score below 100 implies the opposite. While intuitive, this scoring system shares some downsides with all one-size-fits-all-metrics: just like the NPS the NDI ignores the bulk of the customer base (the yellow part) at the center of the distribution, and by combining everyone into one score it obfuscates important differences between individual customers, customer segments, and performance indicators. At D-Sense we have no issue acknowledging this: the NDI is developed as a meaningful, firmly conceptualized alternative to the NPS, but it’s not the only thing you’ll need to grow. Our studies will always accompany the NDI with fine-grained insights about specific customer segments and performance indicators. Because we don’t believe in Reichheldian unicorns.
References
Ajzen, I. (1991). The Theory of Planned Behavior. Organizational Behavior and Human Decision Processes, 50(2), 179-211.
Grisaffe, D.B. (2007). Questions about the ultimate question: Conceptual considerations in evaluating reichheld’s net promoter score (NPS). Journal of Consumer Satisfaction, Dissatisfaction and Complaining Behavior, 20, 36-53.
Keiningham, T.L., Aksoy, L., Cooil, B., Andreassen, T.W., & Williams, L. (2006). A holistic examination of Net Promotor. Database Marketing & Customer Strategy Management, 15(2), 79-90.
Kim, M., Vogt, C.A., & Knutson, B.J. (2013). Relationships Among Customer Satisfaction, Delight, and Loyalty in the Hospitality Industry. Journal of Hospitality & Tourism Research, 39(2), 170-197.
About the Author:
Peter Van Welden is head of D-Sense in Antwerp, Belgium, and is the Belgian representative of Global Growth Group