What is NPS? Net Promoter Score formula, examples, and limits
Net Promoter Score is a common way to track customer loyalty because it comes down to one question and one number. That simplicity is also its weak spot: the score can tell you that customer sentiment moved, but not why.
What is Net Promoter Score (NPS)?
Net Promoter Score (NPS) measures how likely customers are to recommend a company, product, or service. It is based on one question, which Bain & Company, where NPS was developed, words as: “How likely are you to recommend us to a friend or colleague?”
Customers answer on a 0 to 10 scale and fall into three groups:
| Group | Score | What it usually signals |
|---|---|---|
| Promoters | 9 or 10 | Enthusiastic customers who are likely to recommend you |
| Passives | 7 or 8 | Satisfied customers who are not especially loyal |
| Detractors | 0 to 6 | Unhappy customers who may leave or warn others off |
Passives count toward the total number of responses, but they don't add to or subtract from the score.
How to calculate NPS
NPS = % promoters − % detractors
Say 120 customers respond to your survey:
| Group | Responses | Share of responses |
|---|---|---|
| Promoters (9 to 10) | 54 | 45% |
| Passives (7 to 8) | 30 | 25% |
| Detractors (0 to 6) | 36 | 30% |
| Total | 120 | 100% |
NPS = 45 − 30 = 15
That's an NPS of 15, not 15%. Scores range from −100, where every respondent is a detractor, to 100, where every respondent is a promoter.
With a small number of responses, report the counts and the full spread of ratings next to the score. A few answers can move a small sample a long way.
What is a good NPS score?
Any score above 0 means you have more promoters than detractors. Past that, what counts as good depends on your industry, your customers, and how you run the survey. A score from a survey sent right after a support chat isn't comparable to one from a quarterly relationship survey.
The most useful comparison is usually your own trend, measured the same way each time. Bain reports that companies with sustained, profitable growth have Net Promoter Scores about twice as high as the average company.
Relationship NPS vs. transactional NPS
Both use the same question, but they measure different things:
| Relationship NPS | Transactional NPS | |
|---|---|---|
| What it asks about | The customer's overall relationship with the company or product | One specific interaction, such as a purchase, delivery, or support contact |
| When it's sent | On a regular schedule, such as quarterly | Right after the interaction |
| Use it to | Track loyalty over time and compare customer segments | Find which journeys, channels, or touchpoints help or hurt |
Keep the two in separate trends. A score taken right after a support contact and a quarterly relationship score come from different customers at different moments.
How to run an NPS survey
- Keep the question and scale the same. Changing the wording or the 0 to 10 scale breaks the trend.
- Ask why. Follow the score with one open-ended question about the reason for the rating. The answers are often more useful than the number.
- Decide who gets surveyed, and when. Pick the customers, the trigger or schedule, and the channel, then keep them consistent.
- Limit how often customers are asked. Frequent contacts shouldn't get several surveys while quieter customers get none.
- Track response rate. The score reflects the people who answered, not everyone you surveyed.
- Keep the raw ratings. The 0 to 10 spread shows things the three groups hide.
- Don't ask for a 9 or 10. Coaching customers changes the number without improving the experience.
If any of these change, note it on the trend so a change in method isn't mistaken for a change in how customers feel.
How to read a change in NPS
Start with what's underneath the number. The same score can come from very different groups of customers: a split between many promoters and many detractors can produce the same NPS as a crowd of passives.
When the score moves:
- Check that the question, the customers surveyed, the channel, and the timing stayed the same.
- Look at the number of responses and the spread of ratings, not just the score.
- Break the change down by customer segment, product, journey, and channel.
- Read the reasons customers gave and group them into consistent themes.
- Compare with what customers did, such as renewals, repeat purchases, and repeat contacts.
The score points you to where to look. The reasons and the behavior data tell you what to fix.
NPS vs. CSAT
NPS and CSAT are often confused. NPS asks how likely customers are to recommend you, on a 0 to 10 scale, and tracks the overall relationship. CSAT (customer satisfaction score) asks how satisfied a customer was with a specific interaction, usually on a 1 to 5 scale. Many teams use both: NPS for the relationship, and CSAT or customer effort score (CES) for individual interactions.
Limitations of NPS
It throws away detail
Grouping 11 possible answers into three categories makes the score easy to share, but a 0 and a 6 both count as detractors, and a 7 and an 8 both count as passives. Keep the raw ratings for analysis.
It doesn't show cause
If NPS rises after a service change, the change may not be the reason. Product releases, pricing, customer mix, and survey timing can all move the score too.
It only reflects the people who answer
Who you survey, how, and when all affect the result, and customers who don't respond may feel differently from those who do.
It isn't a full picture of loyalty
Researchers have questioned how much NPS alone can predict loyalty or growth. A 2023 study in The TQM Journal found that identical scores can come from very different mixes of customers, which can make the index misleading on its own.
How Sierra customers use NPS
When an AI agent handles customer conversations, NPS is one way to check that customers are happy with the experience, not just that the conversation ended. After SoFi launched its Sierra AI agent, NPS for chats the agent handled on its own rose 33 points. The agent contains 61% of conversations and handles more than 50,000 a week.
A relationship NPS can move for reasons outside customer service, such as pricing or product changes. To see what's driving a score, Sierra customers look at it alongside conversation-level measures like resolution rate, repeat contacts, CSAT, and why customers asked for a person. Ghostwriter, Sierra's agent for building agents, can report how these metrics are trending and suggest changes to the agent to improve them.
See how Sierra customers use AI agents to get better results.
