Net Promoter Score (NPS) is a customer loyalty metric that measures, with a single question, how likely customers are to recommend a brand to people close to them. It was developed by Fred Reichheld and became widespread through an article published in Harvard Business Review in 2003 (see Sources). Today it is used as one of the key experience metrics in many industries, from retail to banking and from automotive to healthcare.
The NPS Question
In NPS measurement, the customer is asked a single question:
How likely are you to recommend us to a friend or colleague?
The answer is given on a scale from 0 to 10, where 0 means not at all likely and 10 means extremely likely. Labeling both ends of the scale on screen helps the customer interpret the scale correctly.
It is recommended to add an open-ended question after the score question: "What is the main reason for your score?" The number shows what changed, and the explanation shows why.
Customer Groups
Responses are divided into three groups:
| Score | Group | Meaning |
|---|---|---|
| 9-10 | Promoter | A loyal customer willing to recommend the brand |
| 7-8 | Passive | A satisfied customer with weak loyalty who may switch to a competitor |
| 0-6 | Detractor | A customer who was not satisfied with the experience and is likely to speak negatively about it |
Calculation
NPS is found by subtracting the percentage of detractors from the percentage of promoters:
NPS = Share of promoters (%) - Share of detractors (%)
The result is expressed as a number between -100 and +100, without a percent sign.
Example
Suppose 200 customers responded in a given period:
- 110 customers gave 9 or 10: share of promoters 55%
- 50 customers gave 7 or 8: share of passives 25%
- 40 customers gave a score between 0 and 6: share of detractors 20%
NPS = 55 - 20 = 35
Passive customers do not appear directly in the formula, but because they are part of the total number of responses, they affect the shares.
Interpreting the Results
NPS varies considerably by industry and country; therefore the most meaningful comparison is between your own past periods, branches and regions. NPS calculated from a small number of responses fluctuates widely: when making decisions at unit level, accumulating at least a few dozen responses for each unit is more reliable than interpreting the ups and downs of individual periods.
Because the same NPS value can come from different distributions, look not only at the result but also at the share of detractors; an increase in this share is a more important warning than a small change in the average. Finally, keep the moment of measurement consistent: relationship NPS (which measures the overall relationship at regular intervals) and transactional NPS (which measures after a purchase, delivery or service) answer different questions and should not be combined in the same chart.
What Should You Do With NPS?
NPS is an outcome metric; on its own it does not tell you what to fix. Its value emerges through the next steps:
- Follow up with detractors. Contacting customers who give a score of 0-6 (a low score) soon afterwards is an opportunity both to resolve the issue and to win the customer back. This approach is called closed-loop feedback.
- Classify the reasons. Sort open-ended responses and root cause questions into topics. The most frequently recurring reasons determine your improvement priorities.
- Track by unit. A breakdown by branch, region or channel shows whether the problem is general or specific to a particular location.
- Measure the change. Track the effect of the improvement in the NPS and share of detractors of subsequent periods.
When Is NPS Not Enough?
NPS measures overall loyalty; it is not designed to measure satisfaction at a specific touchpoint or how easy a transaction is. In these cases, CSAT or CES gives more accurate results. You can find the differences between the three metrics in the NPS, CSAT and CES comparison guide.