Skip to main content

Net Promoter System: from analysis to action in a multinational

Net Promoter System: from analysis to action in a multinational


All organised businesses interview their customers and all claim to be Customer Centric, but the real challenge is how to pass from data analysis to action capable of improving the Customer Experience (CX) of not one, but thousands of customers. This case study describes how a multinational operating in the service sector – specifically the maintenance of installations – managed to transform 140,000 interviews with its B2B customers into three concrete actions, implemented globally.


The client is a leading multinational in the design, production, installation and maintenance of electromechanical installations used in the construction sector. The largest revenues and the greatest number of customers are associated with its installation maintenance and repair activities. These customers – over 400,000 around the world, are typically building, facility or property managers, airports, train stations, hospitals and hotels. Each year, the multinational interviews about 30% of its customer base in the local language e.g. English, French, Spanish, German, Portuguese, Chinese etc., using the Net Promoter System (NPS) methodology and asking essentially just two questions:

  1. How much would you recommend [company name], from 0 to 10, to a friend or colleague?
  2. What is the main reason for giving this score?

The problem we had to tackle was how to manage 140,000 answers to an open-ended question or, in other words, how to find a common denominator within an enormous mass of unstructured data? Historically, the multinational had always used manual classification to analyse this customer feedback, but the process was very slow, subjective, costly and ineffective in identifying key drivers with a positive impact on Customer Experience at a global level. The problem we had to tackle is typical of Big Data Analysis i.e. how to manage 140,000 answers to an open-ended question and find a common denominator within an enormous mass of unstructured data? Historically, the multinational had always used manual classification for this feedback analysis, but the process was very slow, subjective, costly and ineffective in identifying key drivers with a positive impact on Customer Experience at a global level. The multinational wanted:

  • To be able to identify and improve the key drivers of customer satisfaction in a systematic manner
  • To implement an analytical methodology consistently in all countries and branches around the world
  • To define quickly an effective action plan that could maintain its competitive advantage, reduce the number of critics and, therefore, lower the churn rate (customer loss rate)

We immediately identified manual data analysis as the principal weakness in the entire process and introduced the use of powerful AI software, whose algorithm – based on the enormous progress made in the field of NLP (Natural Language Processing) – has the great advantage of consistency. This eliminated at a stroke all the inconsistencies in the analyses carried out in each country, enabling our team to focus on the definition of action plans for the resolution of problems and avoid having to justify, in never-ending discussions, the methodology underlying the interpretations that caused it to focus on certain issues rather than others. The introduction of AI to the text analysis process had another tremendous advantage: the creation of powerful visual reports that are easy to read and available at many levels, from the macro-zones to the smallest branch. This granularity and the consistency of the analytical model adopted were, without doubt, the trump cards of the new methodology. Consequently, management was able to “buy in” and, critically, transition smoothly from presentation of the results obtained from measuring the Customer Experience, to implementation and monitoring of the actions identified in order to ensure their effectiveness.


By contrast with the expectations of management, the principal problem – common to all business units – found via analysis of the Big Data was not the speedy or effective resolution of the faults reported by customers, but rather “COMMUNICATIONS”. The identification of communications as a principal element on which to focus was an important step, but in order to define an action plan it was necessary to drill down further to the operational details. Fortunately, customers know everything about supplier behaviour and their voices are clear and unequivocal: we only have to listen with the tools to understand. The generic “communications” problem was therefore broken down into three areas:

  1. Communications with the customer when a fault is reported
  2. Communications with the sales person responsible for the customer immediately after the satisfaction interview, in order to improve the effectiveness of the subsequent closed-loop meeting
  3. Communications with the technician, who often interacts directly with the customer without being informed about all the non-technical aspects influencing the company/customer relationship


The process of resolving faults in the installation maintenance sector has not changed since the early 1900s:

  1. The customer reports the fault by telephone, using the toll-free number
  2. The call centre operator takes note and promises to send out a technician
  3. The technician, after a variable amount of time, range from 30 minutes to several days, visits the installation, identifies the fault and, after a variable amount of time – ranging from immediately to days – repairs it.
  4. The technician notifies the company that the fault has been resolved and, based on the contractual conditions, the company invoices (or not) the work and any spare parts used

The entire process OMITS communications with the customer who reported the fault, who remains in the dark about the work performed – whether easy or difficult – and who may well receive an invoice for work about which nothing is known. The new analytical methodology highlighted this evident communications gap and the above process was modified accordingly, introducing the concept of ETA (Estimated Time of Arrival) that the technician is required to send the customer, via a text message generated automatically by the system, on accepting notification of the work to be performed. Specifically, we added a mandatory ETA field to the application already developed by the multinational, requiring the technician to input the estimated time of arrival at the installation. Not content with this, we also also added another two text messages containing essential information for the customer. In the first case, the technician now confirms arrival on site with a click in the internal application, which generates another automatic text message for the customer; in the second, another text message is sent when the technician departs, stating either “problem resolved at time xx.yy” or “problem not resolved due to xxx, next steps TBA”.


In accordance with the internal procedure that implemented the NET PROMOTER SYSTEM, the sales person responsible for the customer received a somewhat cryptic, SAP-generated message following the interview, notifying the score given by the customer together with some comments typed by the operator who conducted the interview. This message contained a link to the management software (SAP) used by the multinational, where – in principle – it was possible to read the full interview, as typed in by the operator. This SAP link only worked if the sales person was in front of the computer and already logged into SAP with username and password. As sales persons live on the road, dedicating their time to visiting customers, it took them several days to read the interview and contact the customer again (CLOSED LOOP), thus extending the time taken to discuss the results of the interview with the customer by more than one week. Management recognised the importance of prompt, good quality responses by the company to the feedback given by customers during their interviews. This specific problem was addressed by two main actions. NPS interviews are now recorded and a notification e-mail is sent to the mobile phones of the sales persons concerned. This message specifies the score, details of the customer and the installations with maintenance contracts, as well as their value, and allows the recorded interview to be played back in real time. This solution is much appreciated by customers and the entire sales team, providing an additional tool to demonstrate proactive professionalism by contacting the customer promptly and in a fully informed manner, having already listened to the recorded interview. This attention repays the time dedicated by the customer to respond to the satisfaction survey.


Whenever technicians visit an installation for any reason – maintenance, repairs or action following a reported fault – internal procedures REQUIRE them to contact the local customer both on arrival and on departure. This is known internally as the Check-in, Check-out process. At that time, after having responded to a satisfaction survey, customers frequently asked visiting technicians what the company was doing about the matters discussed during the interview. Of course, they had no idea. This was unsatisfactory for the customers concerned, who often thought that the technicians were there to resolve the problems explained during the telephone interview. To solve this problem, we made the entire text of each interview – score, questions and the answers given by customers – available on the mobile phones of the technicians concerned. We then asked the technicians to checks if the customer had been interviewed recently by accessing, as usual, the information available about the installation and reading the text of the interview. Here too, customer appreciation was unanimous and, no less important, so was that of the technicians, who demonstrated how much they care for their customers and who were happy to be able to show their professionalism and interest, thanks to the new information that was readily available.


The definition and implementation of these three actions made it possible to improve the global Net Promoter Score by 8 percentage points, to 64% for Promoters, while reducing the score for detractors to less than 10%. These results were obtained over a three-year period, following implementation in greatly different markets on all continents. The establishment of such an effective and long-lasting application was only possible due to the clarity of the guidelines and immediate understanding of its importance by all stakeholders: customers, sales persons and technicians. Altogether, a somewhat rare outcome for a Customer Experience project in a multinational enterprise.

CX in a multinational

Sagres used artificial intelligence to analyze the verbatims of 140,000 interviews/year and define 3 main actions to improve the Customer Experience. The global NPS improved beyond expectations, from 46% to 55%.


your challenges

Siamo pronti ad

Raccontateci le
vostre sfide

Complete the form
Receive our business presentation immediately.
An advisor will contact you soon.


Logo Sagres spa

© Sagres S.p.A. All rights reserved.

P. iva: 07676810968