What is a good NPS score?

Short answer

Above 0 is a reasonable starting point, since it means promoters outnumber detractors. Scores above +50 are commonly read as strong within the NPS score range of -100 to +100. But there is no single good NPS score across industries. Judge your result against your own baseline over time, not an outside number.

Above 0 is a reasonable starting point, since it means promoters outnumber detractors. Scores above +50 are commonly read as strong within the NPS score range of -100 to +100. But there is no single good NPS score across industries. Judge your result against your own baseline over time, not an outside number.

ยท 6 min read

How is NPS calculated and what do the score bands mean?

NPS is calculated by subtracting the percentage of detractors from the percentage of promoters, based on responses to a single 0-to-10 question: โ€œHow likely is it that you would recommend this company to a friend or colleague?โ€ Fred Reichheld introduced this question in Harvard Business Review in December 2003. Respondents scoring 9 or 10 are promoters, 7 or 8 are passives, and 0 through 6 are detractors. See the full net promoter score definition for the underlying mechanics.

Here is how to calculate NPS with real numbers. A team collects 1,000 responses: 450 score 9 or 10 (promoters), 350 score 7 or 8 (passives), and 200 score 0 through 6 (detractors). Promoters make up 45%, detractors 20%. Subtract 20 from 45 and the result is a Net Promoter Score of +25. Passives sit in the denominator but add nothing to the score, which is why two companies with different passive rates can land on the same NPS despite different underlying sentiment.

The bands below follow a widely repeated convention, usually attributed to Bain & Company: above 0 is good, above 20 is favorable, above 50 is excellent, and above 80 is world-class. Other versions of the same convention use different cutoffs, such as 30 for good, 50 for great and 70 for world-class, and the attribution rarely comes with a dated original source. Treat the bands as conventions in wide circulation, not statistical thresholds from a specific study.

Score band

What it usually signals

What to check next

Below 0

Detractors outnumber promoters

Review detractor comments for recurring friction points

0 to +20

A favorable tilt under the common convention

Segment by journey stage to find where passives form

+20 to +50

Solid performance by common industry convention

Compare against your own prior-period baseline

+50 to +80

Considered excellent under the same convention

Confirm the result holds across channels and survey timing

Above +80

Labeled world-class; some versions set that label at 70

Verify sample size before treating the score as stable

Why do NPS benchmarks differ so much between industries?

NPS benchmarks by industry differ because the question measures a mix of sector expectations, survey mechanics, and timing, not one consistent sentiment. Forresterโ€™s Global Net Promoter Score Rankings, 2025, surveyed more than 275,000 customers across 478 brands, 13 industries, and 13 countries, and found that globally 23% of brands declined, 5% improved, and 72% were statistically unchanged. Forrester cautions against comparing an internal NPS to its national rankings, since methodology and sampling differ enough to make the comparison misleading.

Four mechanics break comparability between any two NPS figures. Industry and category expectations set a different baseline tone before a survey goes out. Survey channel and mode matter, since phone-administered surveys tend to pull more extreme, positive answers than web-form surveys for the same population. Timing relative to the interaction shifts the score depending on whether a customer is reacting to one transaction or summarizing a longer relationship. Question wording and scale localization change how respondents read the same numeric scale across languages and regions.

This is also why transactional NPS and relationship NPS cannot be averaged together. Transactional NPS is collected right after a specific interaction, such as a resolved support ticket. Relationship NPS runs on a recurring schedule, commonly quarterly, and reflects the overall brand relationship. The same company can run both and land on two different numbers for the same customer base, because each measures a different scope.

Keiningham et al., writing in the Journal of Marketing in 2007, replicated Reichheldโ€™s original analysis using longitudinal data from 21 firms and more than 15,500 interviews, and did not find NPS superior to other loyalty metrics for predicting growth.

What is a good NPS score for your own team?

A good NPS score is judged against your own trend line, not an industry number, which requires a disciplined process rather than a single reading.

  1. Fix the survey design and timing, then leave it alone. Changing the wording, channel, or trigger resets your baseline and invalidates future comparisons.

  2. Set a baseline from at least two consecutive measurement periods before drawing conclusions. One score is a snapshot, not a baseline.

  3. Require a minimum NPS sample size before calling a change real. The exact figure depends on the confidence level and margin of error you choose, but a small sample makes normal noise look like a meaningful swing.

  4. Split the score by segment, journey stage, and channel. A flat headline number often hides one segment sliding while others hold steady.

  5. Judge direction and consistency over the absolute number. Three consecutive periods moving the same way is stronger evidence than one large jump.

  6. When the score moves, read detractor comments and passive comments, group them into causes, and check the conversations behind them in tickets and calls.

After that, fix one cause, re-measure the affected segment, and confirm the movement held in the next period.

Teams with no existing NPS or CSAT process do not need to start with a survey. Start by reading the support conversations already logged, group the recurring complaints into causes, and add a survey only where those conversations leave a gap a transcript cannot fill.

How does OnClarity help with this?

OnClarity (www.onclarity.com) Voice of Customer reads 100% of conversations across calls, chats, tickets, reviews, surveys and social, and ties NPS and CSAT movements to topics and root causes. That is the step between seeing a score drop and knowing why: a team can trace a dip straight to the detractor comments driving it. Research Copilot answers questions in plain language over the full dataset, and alerts flag a topic when it starts trending, before it shows up as a quarter-over-quarter swing.

OnClarity analyses survey results; it does not send or build surveys. It is SOC 2 Type II certified, GDPR and HIPAA Ready, and Saudi PDPL aligned.

Frequently asked questions

Is a negative NPS always bad?

Not necessarily. A negative NPS means detractors outnumber promoters, but context decides whether that is a problem. A company in a low-expectation category, or one surveying right after a complaint-driven interaction, may sit below zero while still retaining customers and growing revenue. What matters more is whether that score is improving against its own baseline or sliding further over consecutive periods.

How many responses do you need before an NPS change is meaningful?

There is no single fixed answer, since the right NPS sample size depends on the confidence level and margin of error a team chooses to apply. A small sample can swing 20 to 30 points between periods from statistical noise alone. Set a minimum response count in advance for each segment before deciding a movement reflects actual experience rather than sampling variation.

How often should a team measure NPS?

Frequency depends on the type of NPS a team runs. Transactional NPS fires after each defined interaction, such as a resolved support ticket, and accumulates continuously. Relationship NPS runs on a fixed recurring cadence, commonly quarterly, to track the overall brand relationship. Running both gives a fuller picture, but neither should be averaged together, since they measure different scopes of the customer relationship.

Key terms

CSAT

CSAT, or customer satisfaction score, is a post-contact rating averaged across responses and usually collected with a single question immediately after an interaction closes. Because only a small fraction of customers respond, CSAT reflects the experience of people motivated to answer rather than the whole contact population.

Detractor

A detractor is a Net Promoter Score respondent scoring between zero and six, treated as an active churn and word-of-mouth risk. Detractor comments are usually the most useful part of an NPS programme, because they name specific failures that aggregate scores cannot surface on their own.

Net promoter score

Net Promoter Score is a likelihood-to-recommend score used as a loyalty benchmark, calculated by subtracting the percentage of detractors from the percentage of promoters. It produces a single comparable number, but the free-text comment attached to each score carries most of the diagnostic value.

Transactional NPS

Transactional NPS is a Net Promoter Score survey triggered immediately after a specific interaction, such as a resolved ticket or a completed call, rather than sent on a fixed calendar cycle. Because it fires per interaction, sample sizes build quickly. Teams get enough volume to compare scores across channels, queues, or agents. This contrasts with relationship NPS, which asks about the overall brand relationship on a periodic basis. Sent too often, it risks over-surveying customers and depressing response rates. Example: asking for a score right after a ticket closes, then comparing chat resolutions against phone resolutions. Done well, it flags a declining queue within days rather than waiting for the next quarterly relationship wave.

Support teams are no longer a cost centre.

See how AI Support Agent, Agent Assist, and VoC Analyst work on your own ticket volume.

Support teams are no longer a cost centre.

See how AI Support Agent, Agent Assist, and VoC Analyst work on your own ticket volume.

Support teams are no longer a cost centre.

See how AI Support Agent, Agent Assist, and VoC Analyst work on your own ticket volume.