How to Calculate NPS — 5 Steps from Promoters and Detractors

The bottom line: NPS is calculated as "% of promoters − % of detractors." It is not an average. If you mistake this and report something like "an average of 8.2 out of 10," that is not NPS — it is just mean satisfaction.

NPS (Net Promoter Score) is a metric proposed by Reichheld (2003) that expresses customer loyalty in a single number. The calculation itself is simple, but people keep stumbling over how the question is worded, where the three-group boundaries fall, and how the percentages are taken. This article walks through the 5 steps to calculate NPS correctly, paired with the "this is where many go wrong" pitfall at each step. How to read the metric and benchmarks belong in Reading NPS and benchmarks; here we focus on "how to calculate it."

Step 1: Ask on a 0–10, 11-point scale

Calculating NPS starts with the right question. The format is fixed.

"How likely are you to recommend this product/service to a friend or colleague?" An 11-point scale from 0 (not at all likely) to 10 (extremely likely)

This 0-to-10, 11-point scale is the non-negotiable condition for NPS. Anything asked on a 5- or 7-point scale cannot be calculated as NPS.

This is where many go wrong: trying to derive NPS from a 5-point scale. NPS assumes a 0–10, 11-point scale, and the three-group boundaries are defined on top of it, so change the number of points and you get a different metric. For the difference from Likert scales, see the Likert scale design guide. One question is enough, but adding an open-ended "why" follow-up right after it surfaces hints for improvement.

Step 2: Sort responses into three groups — promoters, passives, detractors

Split the responses you collect into three groups by score. These boundaries are fixed.

• Promoters: 9–10. Enthusiastic fans — the people who stay and refer others
• Passives: 7–8. Satisfied but low-energy; easily lured away by competitors
• Detractors: 0–6. Dissatisfied customers, and a source of negative word of mouth

This is where many go wrong: mishandling the passives (7–8). A 7 and an 8 are not "promoters." If you fold them in because "they gave us an 8, so surely they're promoters," your NPS comes out higher than reality. It is also easy to miss that the entire 0–6 range is detractors (a 6 is not "pretty good" — it's a detractor). These boundaries were set by Reichheld based on the correlation with customer behavior, and the cardinal rule is to never move them on your own.

Step 3: Calculate each group's share (%)

Once you have the three groups, calculate each group's share of the total.

• % of promoters = number of promoters ÷ total respondents × 100
• % of detractors = number of detractors ÷ total respondents × 100

Example: out of 200 total respondents, 90 promoters, 70 passives, and 40 detractors gives

• Promoters = 90 ÷ 200 × 100 = 45%
• Detractors = 40 ÷ 200 × 100 = 20%

This is where many go wrong: getting the denominator wrong. The denominator for the share is always the total number of respondents (everyone, passives included). If you exclude passives and compute something like "promoters ÷ (promoters + detractors)," the figure swings wildly off. Passives never appear directly in the formula, but don't forget they are firmly included in the denominator.

Step 4: NPS = % promoters − % detractors

Now for the calculation. Just subtract the share of detractors from the share of promoters.

$$\text{NPS} = \text{\% of promoters} - \text{\% of detractors}$$

For the example above:

$$\text{NPS} = 45 - 20 = 25$$

NPS ranges from −100 to +100 (all promoters gives +100, all detractors gives −100). It carries no unit and is written as a whole number, as in "NPS is 25."

This is where many go wrong: adding a "%" and writing "NPS is 25%." NPS is a subtraction of percentages, but the result itself is treated as a score (points), not a percentage, by convention. Write it as "25 points" or simply "25."

Calculating in Excel

In practice, Excel or a spreadsheet can produce it in one pass. Assuming the responses sit in a single column (say column B):

• Number of promoters: =COUNTIFS(B:B,">=9")
• Number of detractors: =COUNTIF(B:B,"<=6")
• Total responses: =COUNT(B:B)
• NPS: =(promoters - detractors) / total responses * 100

Subtract detractors from promoters, divide by the total responses, and multiply by 100 — NPS pops out in a single step. The full workflow for importing a CSV and aggregating it is covered in the survey Excel aggregation guide.

Step 5: Read the score — but never judge it in isolation

How should you read the NPS you calculated? The practitioner's habit is to not get swept up by the raw number alone.

• NPS levels differ greatly by industry, country, and culture. "Is 25 good?" can't be answered in isolation
• Respondents in Japan tend to cluster toward the midpoint, so scores often come out lower than in the West
• What matters is the change over time and the relative comparison with peers, more than the absolute value

This is where many go wrong: judging "good or bad" from the absolute value of a single measurement. The value of NPS lies in tracking change through repeated measurement. There is also academic criticism of the calculation method itself (such as the way it discards passives), as in Dawes (2024). The correct way to read the score, the benchmarks, and the limitations are laid out in detail in Reading NPS and benchmarks, so once you can calculate it, be sure to read that too.

The editorial take — 3 things that really matter when calculating NPS

From the vantage point of continuously following industry cases and the voices of practitioners, here are 3 things that always matter when calculating NPS.

1. Never, ever confuse "average" with "NPS"

This is the most common accident. "An average of 8 out of 10, so NPS 80" is flat-out wrong. NPS is not an average; it is the share of promoters minus the share of detractors. If you report an "average" while calling it "NPS" in an executive meeting, every peer comparison and benchmark goes off the rails. This is the one point you can never miss.

2. Always report N (the number of responses) alongside it

Because NPS is a subtraction of percentages, a small N swings hard. Saying "NPS 30" at N=20 means a handful of responses can move it ±20 with ease. Always report the response count, as in "NPS 30 (N=20)," and treat scores with a small N as indicative only. For how to think about required sample sizes, see How to determine survey sample size.

3. Aggregate the open-ended "why," not just the score

The open-ended "why did you give that score" matters more for improvement than the NPS number itself. What are detractors unhappy about, and what do promoters value? Don't stop at producing a score — design the survey to collect the reasons as a set. For designing open-ended questions, see the open-ended question design guide.

Summary — the 5 steps to calculate NPS

1. Ask on a 0–10, 11-point scale — a 5- or 7-point scale isn't NPS
2. Sort into three groups — promoters 9–10 / passives 7–8 / detractors 0–6. Don't move the boundaries
3. Calculate the shares — the denominator is always the total respondents (passives included in the denominator)
4. % promoters − % detractors — the result is a score, not a percentage. −100 to +100
5. Read it over time and relatively — don't judge by the absolute value alone; confirm how to read it in the dedicated guide

Calculating NPS isn't hard once you've nailed "% promoters − % detractors." The stumbles are always the same three: confusing it with the average, the three-group boundaries, and getting the denominator wrong. Avoid those and anyone can calculate it correctly. Once you can, move on to reading the score and benchmarks.

---

If you want to build and aggregate an NPS survey, why not try Kicue, a free survey tool? Creating the 0–10, 11-point question, adding an open-ended "why" follow-up, and exporting a CSV with respondent IDs — you can start building and preparing to aggregate your NPS survey from a single account. (For computing % promoters − % detractors, the reliable approach is to run the exported CSV through Excel's COUNTIFS.)

References (2)

• Reichheld, F. F. (2003). The One Number You Need to Grow. Harvard Business Review, 81(12), 46-55.
• Dawes, J. G. (2024). The Net Promoter Score: What should managers know? International Journal of Market Research, 66(2-3), 182-198.