What is NPS: Net Promoter Score, the Customer Loyalty indicator

What is NPS
28 min

NPS – Understand about the methodology that builds customer loyalty

THE NPS, acronym for Net Promoter Score, is a customer loyalty metric created by Fred Heichheld in 2003 through the publication of an article called The One Number You Need to Grow in Harvard University magazine.

The objective of the metric created was to measure the degree of customer loyalty of companies in any segment, reflecting the experience and satisfaction of these.

Since then, the key performance indicator has been widely used due to its simplicity, reliability and flexibility. Its literal translation means “Balance of Promoting Clients” and if analyzed in a simplistic way it portrays an old indicator used in all segments - the famous word of mouth.

This indicator is nothing more than the calculation resulting from the balance of promoting customers less the detracting customers of a brand., using a scale of 0 to 1 10, starting with a single question:

How likely are you to recommend [company X] to a friend or colleague?

What is NPS?

0 to 1 6 are considered Detractor customers;
7 and 8 are considered Neutral customers;
9 and 10 are considered Promoter customers.

NPS =% Promoter Clients – %Detractor Customers

Why NPS was created?

Aware of the power of customer loyalty, company executives and market leaders exchanged experiences and debated on how to enhance their loyalty efforts. On this occasion, the great demand was for a simple solution, where it allowed them not only agility in the operation but above all high response rates, so that they could have a high volume of data to categorize and share through a ranking of their stores (at the time there were more than 5000 units) all over the US. These market leaders believed that through this methodology it would be possible to provide feedback on the status of each unit, generating networking opportunities among their peers.

One of the executives present shared that he had been working on building a satisfaction metric that did not use traditional and complex satisfaction surveys, and this debate was the framework for a study of 2 years then carried out by Frederick F. Reichheld, where the synthesis was the use of this single question that identified not only customer satisfaction or loyalty, but also your willingness to recommend the company to someone else.

The date of the creation of the NPS as a metric was recognized in the year 2003, date it was mentioned in Harvard Business Review through the article The One Number You Need to Grow (The number you need to grow). However, the official publication of the material was dated 2006 through the launch of the Book The Ultimate Question (The definitive question)

The Definitive Question 2.0

NPS 2.0 (Net Promoter System)

After 7 years of its publication in HBR (Harvard Business Review), and 5 years after his first book on the topic, Fred realized that the market had already adopted the indicator as an important metric for his business. It was then necessary to bring new approaches. O NPS should no longer be seen as an indicator, but rather as a system, moving from Net Promoter Score to Net Promoter System.

Being the NPS a simple and low cost metric for collection, Reichheld then proposed that the research should go beyond sampling (semester, quarterly or monthly) for transactional searches (every transaction carried out), because in this way it would be possible to categorize the process as a whole more accurately, being possible to adjust it in real time.

In addition to this approach, the possibility of categorizing clients into Promoters, Neutrals and Detractors would facilitate loyalty by allowing those customers who at first did not adhere to the brand to be impacted, worked and converted into promoting clients.

Thus, Reichheld proposed in his last book, The Ultimate Question 2.0 (The Definitive Question 2.0), categorize as many customers as possible and close the loop by treating detractors.

NPS timeline

  • 2001 – Fred participates in a forum with market leaders who understand the power of customer loyalty and begins his study;
  • 2003 – Fred publishes in HBR (Harvard Business Review) an article about the Definitive Question
  • 2006 – Fred publishes The Ultimate Question presenting the NPS as an indicator (SCORE)
  • 2011 – Fred reviews the book The Ultimate Question, and renames it to The Ultimate Question 2.0 where he introduces the concept of NPS as a SYSTEM and not a Score.
A NPS history
NPS history 2001-2011

Diferencial NPS

The great metric differential NPS is that it is not just a simple indicator of customer satisfaction or loyalty. This metric goes beyond, as it reflects the customer's level of desire to recommend your brand!

In addition to the NPS, other indicators are also used and will be addressed in more depth in the course of this material., are they: CSAT, CES, ACSI e Escala Likert.

NPS Differentials

Benefits of NPS

Simplicity in methodology
With a single question and very simple calculation, the NPS concept is accessible to all levels of the organization.

Application simplicity
Technology can be used to facilitate data consolidation, However, NPS can be applied even in print.

High response rates
The agility of the research generates a high response rate, a fact that helps to obtain great statistical relevance not only for the brand, districts and regionals, but also at the store level.

Main indicator used by the market
The high adherence of NPS in different segments even allows Benchmarking (comparison with competitors).

Search Form – Design

Focus on the NPS question:

It is quite common in the implementation of the NPS to involve the marketing team that wants to make the image of the email look attractive and clearly stamp the visual identity of the brand. The problem is that this desire often overcomes the need to emphasize the NPS question.

What is most important in the journey of an NPS survey is the question, and therefore no image or text is competing with the question.

Design Mobile SoluCX
SoluCX NPS Form

Scale coloring

Some models like monochrome, colored or gradient can be used. To the SoluCX the search result is independent of the model (data collection), because the great differential for this process is the constant measurement and its systemic comparison of the results (always using the same methods and channels). It is this systemic constancy that guarantees the fidelity of the process without any bias.. That is, regardless of the scale used, the important thing is to compare the results of the same scale in different periods and find progress in the indicator.

Note: At SoluCX - Customer Experience Management Software the customer is offered the possibility to choose the type of scale to be used.

NPS Scales
NPS scale

Start with 0 or with 1? Scale up 5 Or until 10?

As for the scale, many question whether it is possible to use the 1 to 1 10, excluding the 0, or also use the scale until 5.
Sim, it's possible, however Fred Reichheld remembers that the 0 was inserted in the scale to eliminate the respondent's doubt about the scale order, after all nobody considers that the 0 can be a good grade.
In the case of the scale up to 5, the problem encountered is that the scale loses its accuracy.
For example: On a scale of 0 to 1 10 many customers who would rate 9 (District Attorney), if the scale was 0 to 1 5 would give the grade 4 (Neutral).

NPS SoluCX Evaluation
SoluCX NPS Forms

Additional questions

As mentioned above, in The Ultimate Question 2.0, a chapter was dedicated exclusively to address the best practices of data measurement.
The need for objectivity in research design is so great that the most relevant principle addressed by the author is about which questions need to be really considered in the form.

As already discussed, the NPS question (on a scale of 0 to 1 10, how much would you recommend our company), whose focus is to categorize customers into Promoters, Neutrals or Detractors, there is also the possibility of asking a second question in order to diagnose the root cause and even direct the deal to the responsible department.
Some suggestions pointed out by Reichheld can be considered for the elaboration of the form:

Desgn mobile SoluCX

1. “What is the main reason for the grade you gave us?”
2. “What is the main improvement we should make to increase the chances of you recommending us?

Tip SoluCX:

Avoid adding more than 1 or 2 questions on the same form. That's because they will yield few good insights, wasting customers' time and consequently lowering engagement rates.

Ideal time for collection?

It is important to remember that the NPS methodology is an indicator to assess the likelihood that the customer will indicate the brand to a friend or relative. Thus, it is not appropriate to use this method to measure satisfaction at a specific point on the Customer Journey, such as the SAC.

In this case, to 1 SoluCX recommends using the indicator CSAT (read here about CSAT).

Tip SoluCX:

The ideal time to collect NPS should always be ASAP (after the Customer Purchase Journey is completed).
The close collection of the shopping experience is important to obtain the details of this customer's experience, especially when this is detractor, often unhappy with the situation and totally willing to detail their problems.

Important: If we are talking about a recurring service, the NPS can be collected with a certain frequency, monthly for example.

Handling customer comments?

There is no doubt that open-ended questions are rich in insights, However, due to the high volume of responses, non-sampling analysis becomes impossible without the use of any specialized tool in the classification of texts.

Among the text analysis options for open-ended questions, we have:

Word cloud:

It's a photograph of the moment. A visual resource, but subjective and without clear metrics. In this method it is not possible for example to compare periods. It is a good tool for gaining insights, but shallow in deepening the metrics.

Text Match:

It is a simple keyword labeling system that reports which ones are more repeated. These are not classified as positive or negative, therefore, it is not a reasonable indicator for the day-to-day operation.

For example:

Label – Attendance

Key words or sentences: Attendance, Employee, Agility, It takes at the cashier, Grumpy, Grumpy, no manners.

The biggest problems related to the type of classification above are found in the fact that the isolated word gains incorrect classifications.
For example, “Grumpy”, could be associated with an assembly worker, however it will be classified as Customer Service. Another problem is the need to write numerous variations of the word within each label, that is, it will be necessary to register moody and also moody, or bad mood (without hyphen), and so on.

Natural Language Processing:

Through artificial intelligence, a pre-trained algorithm with a minimum amount of comments (approximately 200 texts by tag), automatically classifies all received texts.

The accuracy of the classification is improved with the amount of texts received, allowing the training of machines for different segments.
The biggest benefit of this type of treatment for comments on open questions, is that this classification allows the creation of indicators for each tag, for example:

“Although the product is excellent, exactly what I was looking for, the boy who attended me was very rude and slow! Besides that, the assembly was not very good either, the doors were not well aligned as I wanted, but otherwise it was all very good!

Machine Learning no NPS
Likes and Dislikes Report

Collection Channels


Without a doubt, this is the best and most suitable channel to be used in triggering Satisfaction Surveys. Both for the cost benefit and the volume of data absorbed.

However, some care needs to be considered for the success of this channel, are they:
• a clean cadastral base in order to avoid blacklists;
• good structuring of art in html to avoid distractions in research;
• good structuring of the title and key information of the customer and his purchase in html so that he identifies with the email and eliminates possible doubts about the day of the experience to be evaluated.

Response rate: In addition to the benefits listed above, in retail, it is common to find companies with 10% response rate, some cases without incentives with 20% and others associated with a loyalty program with up to 28%. In health, however, superior results are found 40%.

Email Search SoluCX


SMS is considered a necessary evil for many companies. For those who do not have a good registration base with email, SMS is often the only channel available. The problem with SMS is that it has a high cost and usually its engagement does not match.

Just like email, SMS has the benefit of not being intrusive, after all, the customer just has to ignore him if he doesn't want to answer the survey.

Answer rate: It is quite common to find companies with 2% on this channel, that is, 98% of the cost with SMS are lost. In rare cases, there are companies with engagement rates of 14% on this channel.


With operators:

Against the technology, but totally humanized, the satisfaction survey for this channel is often considered intrusive, especially if operators are not well educated in the approach.
In digital media, the customer can ignore the message and spend no time on research, already on the phone, even to ignore the approach, it is necessary for the interviewee to dedicate time for such. A point to be considered is that through this channel the possibility of generating detractors is greater than in the others listed above. Besides that, the cost for collecting NPS assessments through this channel is the highest.

No operators (HELLO):
Research carried out using telephony without an operator takes place through the IVR system (Audible Response Unit). A recorded script is played to the customer when he accepts the call. This script can give details such as date and unit of purchase and the customer is then asked for their NPS responses by voice (ASR – Automatic Speech Recognition) or typing (DTMF – Dual-Tone Multi-Frequency).

Telephony response rate: Between 22% and 35%.

Listen to a Research Model via IVR

Application Push-Notification

The integration of the brand's application with the NPS platform is a good channel in terms of cost and engagement, However, the presence of detractors in that channel is much less frequent, after all, most of the time they are very loyal customers, sometimes fans of the brand and who are more tolerant of problems.

Answer rate: Fees between 30% and 40% are easily found.

Push Notification

Totem on the counter

The totem or tablet is a channel that involves the entire shopping environment, including:
• salesperson requesting evaluation in a direct approach;
• pressure inherent in the environment to quickly assess;
• customer time that sometimes may be in a hurry to leave the establishment.
One of the major impediments to this channel is the high associated investment and the need for adequate infrastructure. There are those who do not like this channel due to its vulnerability to employee fraud that, charged for results, perform random assessments.
Totem response rate: Very high, between 30% and 60%.

Satisfaction Survey Toten

SoluCX Case Study

At the request of one of the main SoluCX customers, the telephony channel was used.
When implementing the solution, SoluCX did not maintain high expectations for this channel to bring the expected engagement, since the channel is increasingly in disuse.

However, as soon as the collections started, an engagement of 4%, higher than 10% that the customer had with the predecessor supplier SoluCX.

With script improvements and use of the ASR feature, it was possible to achieve 22% about valid contacts. The achieved indicator surprised everyone at SoluCX and after feedback meetings with the client it was found that the public was older and without much interaction with the technologies available in the market such as e-mail.. For this reason, even being intrusive, the phone still remained one of the best service channels.

Search bias due to method or channel

But after all what is the bias in a Satisfaction Survey?

Also known as differential error, bias produces deviations or distortions and currently still generates one of the biggest controversies in the world of research.

Where it can be found?

• Collection channels (e-mail, SMS and totem);
• Approach message (Email title, SMS, HTML to start the survey and collaborator requesting a totem response);
• Search form design (scale coloring, question asked, volume of questions in case of negative answers).

Tip SoluCX:

The satisfaction survey bias is practically impossible to eliminate.

Eliminating bias can lead to constant changes as approaches, research design and even interpretation of results. These changes preclude the most sacred aspects of a customer satisfaction indicator, which is the comparison of the indicator with previous periods.. How will you compare the results of the previous month with the current month if there is no collection pattern in your NPS system?

Who should be responsible for NPS in my company?

It is common to find companies that name the same employee to carry out the collection process and provide improvements in the NPS.
The big problem is that many professionals are “playing” with channels and methods instead of seeking improvements in their processes, that is, seek “improve ”the indicator through different strategies in the collection tool.

Thereby, the NPS loses its strength in the organization, after all, however much it improves, the company does not progress as expected.

The correct thing is that the company has:
• A partner to carry out the collection and thus the brand focuses on analyzes that add value;
• A professional responsible for managing the partner who collects the assessments, the implementation and internal evangelization of the NPS;
• Process engineers to analyze research data and improve NPS through changes in processes rather than collections.

Bottom-Up X Top-Down

Conducting NPS satisfaction surveys using the bottom-up technique means that you will give all customers the possibility to conduct the survey.
Every new transaction is triggered a new search is triggered. Thus it is possible to identify the employee and the store involved and to which groupings this specific store belongs.
Like this, both regional, how much districts and sales channels are evaluated and have their data consolidated. In this way it is possible to understand the company's NPS at all levels described above.

Já no Top-Down, often used in benchmark studies, the company determines the desired sample error in the survey. This data is provided taking into account the brand or flag you want to know about the NPS. We can say that the big difference between the bottom-up and the top-down is that the first presents a photo of the process and the second brings a film of the entire trajectory.

In this case, the bottom-up methodology is the most suitable for companies that wish to monitor their NPS daily. It will serve as an indicator to be pursued and improved by converting detrimental customers into promoters..

Carrying out Top-down research as an internal methodology and not as a benchmark is not to worry about treating detrimental clients.

Sample size and sample error

As mentioned above, sample size is required for Top-Down surveys, However, many companies due to lack of budget and mainly because they are in a period of internal awareness of their management, end up using sample size definitions to determine a search volume goal, considering a certain sample error.

It is also common for companies that carry out the bottom-up survey to monitor what is the sampling error achieved in the evaluated period..
Many companies set a target in terms of volume of evaluations per operation or even response rate.

Topdown Bottomup

Tip SoluCX:

The most appropriate when considering sampling error would be to consider a target, due to this indicator considering the transaction volume of each unit / operation.

To set the sample size or find out which sample error of the evaluated period, we recommend reading our specific article on the topic at the link above.

Calculate sample error and sample size yourself:

We prepare calculators for you to experiment at will and calculate the sampling error it's the sample size ideal for your research goals.


Based on GDPR (General Data Protection Regulation) from Europe, a LGPD (General Data Protection Law) is the Brazilian version of the law created to protect customer data. This law requires, among other details, that companies:

• ask your customers for authorization to send satisfaction surveys;
• provide an option to delete the registration if the customer wishes;
• option to edit personal registration data;
• elimination of personal data whenever the relationship ends;
• issuing reports that prove data exclusion.

Read here about General Data Protection Law.

NPS and CRM integration

When the company chooses to close the cycle (NPS 2.0), that is, not only classifying customers into Promoters, Neutrals or Detractors, but also promoting actions that seek to convert Detractor customers into Promoters, some good practices are necessary.
The first one, in order to make the process productive, would be the integration of the tool chosen for collecting NPS data with CRM (Customer Relationship Management). This integration takes place through the Rest API.

Habit Power and NPS

Satisfaction survey is not a procedure loved by consumers, for this reason, ensuring great participation is always challenging. That is why, it's important to understand the power of habit to increase response rates for surveys sent.

look here an article that we talked about the power of habit.

Benchmarking: What is a good NPS?

So that this question makes sense, we need to be more specific. For example:
1. What is a good NPS for the telephony sector?
2. What is a good NPS for the retail sector?
3. What is a good NPS for the auto industry?

There are those who want to simplify NPS levels and unify it for all segments, however this does not make sense when we analyze the results in a segmented way.
Below is an example of how NPS has different results in each business segment.

NPS Horizontal

NPS in Brazilian retail

In expanded Brazilian retail, SoluCX has more than 30 big customers. Through all the acquired know-how it is possible to affirm that the NPS in these brands is among 50 and 90.

Step by step for implementation

Top management awareness:
Without leadership commitment to making NPS a priority, the effort spent on implementation will not be effective. One of the relevant conditions is related to data collection.
Initially, a top-dow collection is proposed, in order to present the power of the tool. With this application it is possible for the leadership to understand and feel really comfortable for the implementation.

Customer registration:
For many companies the initial challenge is to have their customer data systematically organized. When that is the challenge, unfortunately, companies do not use good practices for acquisition.
Usually what happens are internal campaigns for “to motivate” employees to hit goals, rewarding new registrations.
However,, it is noticed that many brands establish an incorrect registration volume as a goal, collaborating for irregular registrations in the system.
Some examples are:
• E-mails: naotem@naotem.com.br, joao@qualquerdominio.com.br, 123@ domain.com
• Phones: (11)12345679, (21)987654321, (11)111111111

The power of habit must be considered at that time. The question that needs an answer is: What the customer gets by registering?

Bulk or real-time data loading by integration:

It is important to work with a platform that allows detailed uploads of transactions and customer profiles. This way you can avoid asking the customer everything that is already registered. Research that asks for data that has already been reported is very unpleasant.

Consolidation of data:

Obtaining data is a time-consuming and time-consuming process. However,, if carried out in a systemic way through platforms such as SoluCX, this acquisition happens to be in real time.

Detractor customers:

“The best advertising is done by satisfied customers”. The phrase is by Philip Kotler, considered the father of marketing. So if selling is scalable, treating detrimental customers is also! Being patient and analyzing complaints are important starting points for improving relationships with detrimental customers..

As explained not NPS 2.0, detrimental customers are identified, these should be directed to treatment via SAC. Anyone who thinks it is impossible to retain a dissatisfied customer is wrong. Showing attention to flaws that have been noticed can be a point of engagement and customer loyalty.

Application of necessary improvements:

It is now perceived in the market that many companies measure, but few take effective action to resolve complaints.
An important point to initiate corrective actions in this process would be the application of the PDCA cycle:
P – Plan (to plan)
D – Do (Perform)
C – Check (Check)
A – Act (Agir)

This is the secret of NPS deployment. These techniques found in the industry should also be adopted to improve the process:

– Pareto Diagram
– Diagrama de Ishikawa
– 5W / 1H or 5W / 2H
– 5S
– Lean
– Six Sigma
– PDCA cycle

Communication to customers (real reward):

Unfortunately, consumers have not yet incorporated a culture of voluntarily joining the interaction with satisfaction surveys..
Within today, even with the awareness of the consumer's power over a brand, the return is far from ideal. That is why, when a customer comes to interact with a brand, answering the survey and exposing your pain, without any stimulus for that (in consumer earnings), the least that can be offered to him is his pain healed as a prize.

Answering a satisfaction questionnaire and not being answered will only serve to accentuate the customer's level of dissatisfaction. In cases where the problem has already been solved, but the client continues with a detrimental stance, what he expects as a prize is to be heard and that the brand generates improvements through its dissatisfactions. The best thing to do is to notify customers who have given you feedback on improvements made to the process.

Team awareness:

With conscientious leadership, it will be possible to disseminate the importance of the indicator throughout the organization.
For this reason, some companies carry out some promotion actions that are:
• associate NPS with PLR (Profit and Results Sharing Program) of employees;
• set goals at different levels of the organization (matrix, districts, regional, units and collaborators);
• distribute results in real time to all levels that have goals.

About a SoluCX

A SoluCX is a startup born in São José dos Campos (SP) that offers solutions for managing the experience of
client (CX).
With its innovative methodology, companies of all sizes have access to fundamental information to understand customer behavior and its relationship with the brand. With this it is possible to outline strategies to improve financial results from the loyalty and improvement of services and processes. So it is possible to create a closer relationship with the communities where they are inserted.

SoluCX logo

Improve the experience and enhance your results

Join our list and receive exclusive content

4 Responses

  1. to get better, only one app was missing. This makes the query more dynamic., I can even access my company's valuation from home

Leave a Reply

Your email address will not be published. Required fields are marked *