Net Promoter Score (NPS) — Almohannad Alsbeai

Historical Background

The first peer-reviewed scientific paper on this metric was published in the Harvard Business Review in 2003, titled The One Number You Need to Grow, by Frederick Reichheld. After that article, many related studies and even several books appeared, and today you’ll find numerous companies offering consulting services to help organizations build a complete program around using this metric.

The metric is a registered trademark of Bain & Company Inc., Satmetrix Systems Inc., and Fred Reichheld. No fees or licenses are required to use it, but academic integrity requires that any page or slide on which the metric is used carry the following note:

NPS®, Net Promoter® & Net Promoter® Score are registered trademarks of Satmetrix Systems, Inc., Bain & Company, and Fred Reichheld.

Introduction

NPS is the most widely used metric among customer loyalty measures. It has proven its merit in multiple empirical tests and has been adopted by many companies as an approved indicator for measuring customer loyalty.

How the question is asked: “How likely is it that you would recommend dealing with company X / buying the products of company X? Please express your opinion using a scale from 0 to 10, where 0 means ‘not at all likely’ and 10 means ‘extremely likely.’”

The scale consists of 11 points, from 0 to 10, and is divided — when analyzed — into three main groups representing customer loyalty states:

Group 1 (Detractors): ratings from 0 to 6 — customers who speak negatively about the company and recommend against dealing with it.
Group 2 (Passives): ratings of 7 to 8 — neutral customers.
Group 3 (Promoters): ratings of 9 to 10 — customers who speak positively about the company and recommend dealing with it.

An additional option — “Don’t know / Can’t say” — is added to the scale to avoid getting distorted data from customers whose experience has not yet matured enough for them to pass judgment. When applying the formula, all responses under this option must be excluded so they do not skew the calculation.

The Formula for Calculating NPS

There are two ways to apply this formula:

Subtract the number of Detractors (those who answered 0–6) from the number of Promoters (those who answered 9–10), then divide the result by the total number of respondents — excluding the “Don’t know” responses if they were included in the scale.
Subtract the percentage of Detractors from the percentage of Promoters, again excluding “Don’t know” responses if included.

Example: Suppose we have the following distribution of answers:

0 to 6: 10 responses
7 to 8: 9 responses
9 to 10: 20 responses
Don’t know / Can’t say: 5 responses
Total respondents: 44
Total respondents after excluding “Don’t know”: 39

Method 1: (20 − 10) / 39 = +25.6% Method 2: 51.28% − 25.64% = +25.6%

Reading the Results

Many people misread the result of the formula, which comes out as a positive or negative percentage. Here’s what it means: if the result is +20%, the percentage of Promoters who recommend the company exceeds the percentage of Detractors by 20 points. If it’s −20%, the percentage of Detractors exceeds the percentage of Promoters by 20 points.

Logically, the higher the positive score, the better; the more negative the score, the worse. You can also analyze the metric outside the formula by examining the customer groups themselves. A large concentration of customers in the Detractor group is a danger signal and an indication that more customers will defect to competitors in the near future. On the flip side, focus is placed on the Promoter group who will recommend your products or services — offering them exclusive incentives such as promotions, discounts, gifts, seminar invitations, and so on.

NPS vs. Customer Satisfaction Metrics

Some people confuse these two metrics and assume there is a direct positive relationship between them — but this is wrong. Each metric has a different purpose. Customer loyalty cannot, in any way, be used as a proxy for customer satisfaction, and vice versa. You may notice that you lose some of the customers classified as the most satisfied, while some of the customers most dissatisfied with your products or services keep dealing with you and continue buying. NPS does not replace measuring customer satisfaction — satisfaction has its own correlations with other internal variables. Neither metric substitutes for the other.

NPS and Revenue

NPS is primarily linked to revenue. It has been tested by several companies and a positive relationship between this metric and income has been demonstrated in many experiments — the higher the score, the higher the revenue, and vice versa. That said, an integrated program must monitor this metric, study the factors that drive it up or down, and track the internal variables tied to it. Simply asking the question and then waiting for magic to happen is, as Einstein defined insanity, “doing the same thing over and over and expecting different results.” If you ask the question but make no decisions based on the answers, you have effectively changed nothing — and so the result will not change either.

Implementation Approach

Asking the question and getting an answer is not enough; there must be an integrated program adopted from the top of the pyramid down. I may devote a separate article to this given how many branches it has, but to give a non-exhaustive sketch: such a program involves dedicated budgets for the changes that program owners will impose across several units and departments; standard procedures and policies that everyone must follow, including the channels through which loyalty will be measured; the internal parties authorized to perform the measurement (or whether it should be centralized); the measurement cadence; the targeting strategy (we don’t necessarily care about measuring every customer en masse — if we did, we’d still need to analyze the small group that drives the largest share of revenue, per the 80/20 rule); identifying the loyalty drivers that differentiate the company from competitors and monitoring the impact of specific loyalty programs on the metric; procedures for circulating results among stakeholders and making decisions based on them; and finally appointing an oversight team to supervise all decisions and continuously study their impact throughout the year.

Challenges to Expect During Implementation

Expect many challenges when you try to implement something similar. The biggest is internal resistance and fear of change — which is why such a program must be sponsored from the top of the organization, otherwise it will be killed before it is born.

There are also problems related to immature customer relationship management systems in the company — including a complete absence of a CRM system, or a system that exists but isn’t activated, with data that is outdated or inaccurate.

Awareness is another challenge: before launch, make sure all stakeholders fully understand the program’s importance, its objectives, what its results will be, and how it will strengthen the company’s market share if the results are properly adopted.

Setting targets to achieve is also a challenge. Once you’ve hit a certain number on the metric, how do you decide what target to set for the next measurement? Any planned increase must be carefully studied and realistic; otherwise it may negatively affect employee performance and produce counterproductive results. Best practices and benchmarks can help here.

I hope this brief introduction to the metric has been useful. If you have any questions, don’t hesitate to message me or post them in the comments.

You may also want to read my companion article: Does NPS Really Deserve All This Hype?

Introduction (Customer Effort Score — companion piece)

In this era, the customer has become far more aware than before, thanks to their varied experiences with many service providers and product vendors. Their expectations are higher, and their tendency to file complaints has grown. Alongside this evolution in customer behavior, the metrics used to predict loyalty and future behavior have evolved too. One of the newest directions in earning customer loyalty — based on recent research — has been to reduce the effort customers have to expend along the path of a given process in order to achieve a specific goal, such as getting a product or service.

Today’s article is about a loyalty metric designed to measure the effort a customer puts in to achieve a particular goal across multiple touchpoints in the customer journey and through various service channels. We cannot reduce effort unless we first measure it. The question is concerned with how the customer feels about the effort they expended to do something. The inventors of the metric found a strong statistical correlation between low-effort experiences and customers who would: deal with the company again, increase the volume or value of their dealings, and speak positively about the company. Because all of these behaviors are linked to loyalty, this metric was classified as one of the strongest predictors of customer loyalty — as illustrated on the cover image — outperforming both customer satisfaction and the famous Net Promoter Score.

A Short Story

A few days before writing this article, I had a very bad experience renewing my internet subscription with one of the service providers. The first thing that irritated me was that they cut off the service without any prior notice. Where did I get the assumption that the provider should warn its customer that the subscription is about to expire? From other experiences, of course. In short: I was supposed to pay the renewal cost via the SADAD payment service and the service would then resume. What actually happened was that I paid the amount via electronic SADAD, but the amount wasn’t credited to my account with the provider. I had to obtain an official payment voucher from the bank, deliver it to one of the provider’s branches, where a branch employee would file an objection. The amount was then credited — but the service still didn’t resume. So I had to go back to the provider’s page and request the renewal again as instructed, then remove the SIM from the modem, put it in a phone, make any call, then put the SIM back in the modem … all this hassle just to renew a subscription! I filed a complaint with the Communications Authority and resolved that this would be my last dealing with this provider. The journey should have been, in short: pay the fee and the service is activated. Why all that hassle?

I mention my story to illustrate that many companies don’t examine their customers’ experiences carefully. If you provide an electronic payment service, that has to be paired with an experience review and verification that it actually works — or don’t provide it at all.

Explaining the Metric

This metric consists of a single main question asked of the customer after a service interaction is completed. The result helps decision-makers understand whether their company is delivering a smooth, easy experience. Because the metric targets specific processes, it makes identifying the pain points in the customer journey straightforward — whereas other metrics gave a general assessment without clarifying where things were going wrong.

The metric is called the Customer Effort Score (CES).

The first discussion of this metric appeared in an article titled Stop Delighting Your Customers, published in Harvard Business Review in 2010. The article explored benchmarks related to evaluating customer interactions with contact centers, and argued that customer effort is the key predictor of loyalty — better than customer satisfaction and Net Promoter Score. The article also noted that customers’ desire for easier, faster service is the cornerstone of their decision to keep dealing (or stop dealing) with a given provider. It emphasizes the importance of this metric for organizations offering electronic self-service channels, which have proliferated in recent years to cut costs.

How the Question Is Asked

This metric has had two versions. Although the old version is no longer used, it’s worth understanding both — and why the inventors moved to a new wording.

The old version (image omitted): If you’re well-versed in survey design, you’ll spot a problem with the scale above, specifically at its endpoints. Usually, point 1 on a scale carries the negative label and point 5 the positive — but in this scale it was reversed, which created comprehension problems for respondents whose habits pulled them the other way. You might say: simple, just flip the endpoint labels. True — but that wasn’t the only problem. The word “effort” itself was hard to convey in some cultures when translated into the local language. Rather than dwell on this, let me show you the improved version.

The new version (image omitted): From a survey-design perspective, the new scale is much better than its predecessor — both in its endpoint labels and in being a 7-point scale, which ensures a more varied distribution of responses.

It’s also important to add an open-ended follow-up question asking why the respondent chose that point on the scale. From the 7-point scale above, what’s called the Net Easy Score is calculated: subtract those who found the experience difficult (those who answered 1, 2, or 3) from those who found the experience easy (those who answered 6 or 7).

Is the Metric Effective?

If your company wants to use it, the culture must be customer-centric. Using the metric and then doing nothing with the results is a waste of time, money, and effort.

I’ve read many articles, some endorsing the metric’s effectiveness and others criticizing the hype around it — as with any metric. There is consensus, however, that CES is the best metric for measuring contact-center and phone-based technical-support performance, and that it is better than NPS for getting into service-level detail. NPS remains the preferred metric at the overall level.

I previously criticized NPS in an earlier article, arguing that being published in Harvard Business Review is not enough to legitimize it — the same caveat applies here. The burden of proof is on the claimant, and experience is the best evidence. My professional view is that no company should rely on a single metric to study customer loyalty or to predict expected income, churn rates, or market share gains and losses. What made NPS famous was its simplicity, and this metric we’re discussing today is also simple and is starting to spread. Most companies want something easier. Decision-makers want the simplest, shortest path to their decisions. In my personal view, having a dedicated customer-experience analysis function is just as important as adopting and applying methodologies to measure those experiences. This function will lead change, shift the culture, and analyze every experience in exhausting detail.

Closing

Claims and debates will continue around these various indicators and the specialized ones offered by global firms. Some claims are built on commercial ambitions; others are well documented and recommended by academic bodies. The takeaway from this article and several prior ones is: there is no single magic indicator you can rely on entirely, and many more indicators will emerge in the coming years, with competition continuing between them.

See also: NPS, CSAT, CES, Voice of Customer (VoC).

← Articles MOC