Skip to content

eNPS and employee satisfaction: what the results tell you and what remains to be found out

How eNPS is calculated, how much it varies by chance alone, why identical scores hide different situations and which questions complement the result.

By the Pulsika team Internal communication Published 10 minutes read

A single eNPS score, with no value, and three questions that explain it: in what context it was given, how many people responded and what the answers say.

eNPS (Employee Net Promoter Score) shows how willing employees are to recommend the company as a place to work: the percentage of those who give 9–10 minus the percentage of those who give 0–6, on a scale from 0 to 10. The result is a single figure, between −100 and +100, easy to follow over time. It does not tell you, however, why people answered as they did or what needs to change.

This article shows the full calculation on an example, how much the score can vary by chance alone (with the formula for the margin of error), why identical scores hide different situations and which questions complement eNPS.

Quick answer. eNPS is calculated from the question “How likely are you to recommend the company as a place to work?”, on a 0–10 scale: the percentage of promoters (9–10) minus the percentage of detractors (0–6). The score ranges from −100 to +100. Interpret it against your own previous results, with the number of responses alongside, and complement it with an open question about the reason.

What eNPS is and where it comes from

eNPS is the employee adaptation of the Net Promoter Score (NPS), a customer loyalty indicator. Frederick F. Reichheld presented NPS in the article “The One Number You Need to Grow”, published in the Harvard Business Review in December 2003 (HBR). Bain & Company, the consultancy where Reichheld developed the system, describes the employee version on the site dedicated to it.

  • Promoters are respondents who give a 9 or a 10.
  • Passives are those who give a 7 or an 8. They count in the total number of responses, but not in the difference.
  • Detractors are those who give between 0 and 6.
  • The score is the percentage of promoters minus the percentage of detractors; it is written as a number, not a percentage: +8, not 8%.

Bain’s page on eNPS words the main question as follows: “How likely would you recommend this company as a place to work?” and states that employee scores can be substantially lower than customer scores (Bain & Company, Employee NPS, checked on 18 October 2026).

How eNPS is calculated: a complete example

Hypothetical example. A fictitious company with 100 employees sends out the eNPS question. 80 people respond: 28 give 9 or 10, 30 give 7 or 8, and 22 give between 0 and 6.

  1. The percentage of promoters: 28 / 80 = 0.35, that is 35%.
  2. The percentage of detractors: 22 / 80 = 0.275, that is 27.5%.
  3. eNPS: 35 − 27.5 = +7.5. Rounded, +8.
  4. The response rate: 80 / 100 = 80%. It is always reported next to the score.

The denominator is the number of people who answered the question, not the number of employees. Someone who did not respond is neither a promoter nor a detractor; that is why an eNPS without the response rate can be misleading.

Identical scores, different situations

The difference between two percentages loses information. Three fictitious teams of 100 respondents each have the same eNPS, 0, but completely different problems.

Team (fictitious data)PromotersPassivesDetractorseNPSWhat is actually happening
A40%20%40%0A divided team: some very satisfied, some very dissatisfied. Look for what separates them (shift, manager, length of service).
B10%80%10%0Moderate satisfaction, with no enthusiasm and no open conflict.
C30%40%30%0An in-between situation; without the distribution you cannot tell it apart from A or B.

So report all three percentages, not just the difference. A chart with three bars says more than a single number.

How much the score varies by chance alone

An eNPS calculated on a small number of responses moves a lot from one month to the next without anything having changed. It is worth knowing by how much before you celebrate or panic.

The effect of a single person

With 80 responses, one detractor who becomes a promoter moves the score by 2 / 80 = 2.5 points. In a team with 20 respondents, the same person moves it by 2 / 20 = 10 points, and a detractor who becomes a passive, by 5 points.

The margin of error, calculated

Each response can be scored +1 (promoter), 0 (passive) or −1 (detractor); eNPS is the mean of these values, multiplied by 100. The variance of one response is p + d − (p − d)², where p and d are the proportions of promoters and detractors. The standard error is the square root of this variance divided by n.

On the example above: p + d = 0.625; (p − d)² = 0.075² = 0.005625; variance = 0.619375. Divided by 80, it gives 0.00774; the square root is 0.088. The 95% margin is 1.96 × 0.088 = 0.172, that is ±17 points. The score of +7.5 is compatible with values between about −10 and +25.

Caution. The formula assumes a sample from a large population. When almost all employees have responded, the statistical uncertainty falls, but another question arises: who are the people who did not respond, and would they have answered the same way? Cross-check the response rate against units and shifts.

Comparing two quarters

Continuing the fictitious example: in the next quarter, 60 people respond: 21 promoters, 27 passives, 12 detractors. eNPS = 35% − 20% = +15, against +7.5. The standard error of the new score is 0.094 (p + d = 0.55; (p − d)² = 0.0225; 0.5275 / 60 = 0.00879). The difference has a standard error of √(0.00774 + 0.00879) = 0.129, so a margin of ±25 points. A rise of 7.5 points cannot be told apart from noise, and the response rate fell from 80% to 60%.

What eNPS measures and what it does not

InstrumentWhat it measuresHow many questionsStrengthLimitation
eNPSWillingness to recommend the company as a place to work1 rating + usually 1 open questionQuick, easy to repeat and to follow over timeDoes not give the cause; loses the distribution; sensitive to the number of responses
Satisfaction indexHow satisfied people are with aspects of their work (hours, manager, pay, conditions)10–25Shows where the problem isLonger; needs careful question design
Engagement surveyAttachment to the work and the organisation, discretionary effort10–40, often a standardised instrumentComparable with other companies when the instrument is standardisedCostly; copyright applies to some instruments
Stay interviewWhy a person stays and what would make them leave5–7, in conversationIndividual detail, concrete actionsNot anonymous; takes managers’ time

Satisfaction and engagement are different concepts: a person can be satisfied with their hours and pay without putting in extra effort. For the full satisfaction survey, see the guide to designing the survey.

The limits of eNPS

  • The validation comes from the customer field. A study published in the Journal of Marketing in 2007, on data from 21 firms and more than 15,500 interviews, did not replicate the “clear” superiority of NPS over other customer loyalty measures (Keiningham et al., 2007). The study concerns customers, not employees; do not extend it automatically, but do not assume that eNPS is better validated either.
  • Recommending is not the same as intending to stay. A person can recommend the company to a friend and still be looking for another job, for example because of pay or distance.
  • The score depends on anonymity. Bain notes that employee feedback must be kept confidential and that, unlike customer NPS, it is not followed by individual clarifying conversations. If people do not believe their answer is anonymous, the ratings may say more about their caution than about what they think.
  • There is no universal “good” threshold. Compare yourself with your own results and between comparable units, with the same question and the same period.

Which questions complement eNPS

eNPS tells you “how much”; the questions below tell you “why” and “where”. Choose 3–6 of them, not all.

What you find outQuestionScale
The reason for the rating“What is the main reason for the rating you gave?”Free answer
What would raise the rating“What would need to change for you to give a higher rating?”Free answer
The relationship with the manager“My direct manager tells me clearly what is expected of me.”1–5, from “not at all” to “completely”
Workload“Over the past month, my workload has been…”Too light / about right / too heavy
Intention to stay“Do you see yourself working here a year from now?”Yes / Don’t know / No
Recognition“In the past month, someone told me specifically what I did well.”Yes / No

For analysis, put the results next to operational data: staff turnover and absenteeism, for the same units and periods. A link between them is a lead to investigate, not proof of causation.

How to run an eNPS measurement

  1. Set the purpose and the decision. What will you do differently depending on the result? Without a possible decision, do not start measuring.
  2. Keep the wording fixed. The same question, the same 0–10 scale, with the ends labelled (“not at all likely”, “extremely likely”), at every measurement.
  3. Add the open question about the reason and, if needed, 2–4 questions from the table above.
  4. Guarantee anonymity and the minimum threshold. Do not report the results of a group with fewer than a minimum number of responses, set in advance (5 is a common convention).
  5. Choose the frequency. Bain notes that companies most often ask for eNPS feedback monthly. For small teams, where one person moves the score by 5–10 points, a quarterly measurement gives more stable figures.
  6. Report in full: the number of responses, the response rate, the three percentages, the score and the approximate margin.
  7. Communicate what you are doing. Within a month of closing at most, say what you have learned and 1–3 actions. Without this step, people have little reason to respond next time.

Common mistakes

  • The score reported without the number of responses. A +40 from 10 responses and a +40 from 400 do not mean the same thing. Fix: n and the response rate next to every score.
  • eNPS as a bonus target for managers. Managers end up asking their team for high ratings. Fix: use it as a signal, not as a paid target.
  • The average rating instead of eNPS. An average of 7.4 cannot be compared with an eNPS. Fix: choose one method and stick to it.
  • Comparing with the “industry average”. Figures published by vendors use different wording, populations and periods. Fix: compare yourself with yourself.
  • Results for teams of three people. Anonymity disappears. Fix: a minimum response threshold for any breakdown.
  • Reacting to every change. A 5-point fall with 40 responses is most likely noise. Fix: follow the trend across several measurements.

Frequently asked questions

What is a good eNPS?

There is no universal threshold that holds for every company. A positive score means more promoters than detractors, but the level depends on the industry, the wording and who responds. Bain notes that employee scores can be substantially lower than customer scores. Compare yourself with your own previous measurements and between comparable units, with the same question.

Why are passives left out of the calculation?

Passives count in the denominator, but not in the difference. The percentages of promoters and detractors are calculated from all respondents, passives included. That is why, if many passives become promoters, the score rises. Ignoring passives in reporting is a mistake: a team with 80% passives is in a different situation from one split between promoters and detractors.

How often should we measure eNPS?

The frequency depends on the size of the groups and on your capacity to act between measurements. Bain notes that companies most often ask for eNPS feedback monthly. For small teams, where a single response moves the score by several points, a quarterly measurement is more stable. Do not measure more often than you can communicate the results and the actions.

How many responses are needed?

There is no minimum that makes the score “correct”, only a margin that shrinks as the number of responses grows. With 80 responses, the margin can be about ±17 points; with 20, much larger. For anonymity, set a separate minimum reporting threshold per group, for example 5 responses, below which you do not show the result.

Is eNPS the same as employee satisfaction?

No. eNPS measures willingness to recommend the company as a place to work, through a single question. Satisfaction is measured across several aspects, such as working hours, the manager, pay and conditions. The two are related, but a person can be satisfied with some aspects and still not recommend the company, or the other way round.

Can we calculate eNPS on a scale from 1 to 5?

Not in the standard form. The split into promoters (9–10), passives (7–8) and detractors (0–6) assumes the 0 to 10 scale. On a 1–5 scale you can report the percentage of favourable and unfavourable answers, but the result cannot be compared with an eNPS. If your tool only has short scales, use a question with 11 options, from 0 to 10.

What do we do with a negative eNPS?

A negative eNPS shows that there are more detractors than promoters, but not the reason. First read the answers to the open question and the distribution by unit, with the anonymity threshold respected. Choose one to three causes that come up often, discuss them with the teams and communicate what you are changing. Measure again once the change has had time to show.

What to do next

If eNPS shows you that something is not working, the next step is to find out what exactly. Design a full survey with the help of the guide to the satisfaction survey, and for key people use the stay interview. To put the results next to your workforce data, read the article on calculating turnover.

Was this article useful?

See how Pulsika works for your team

Try every module free or book a demo with our team. No card needed.

or let us set up a trial company with your structure