eNPS: how to calculate your employee Net Promoter Score and what to do with the result

What eNPS is and how to calculate it

eNPS (employee Net Promoter Score) is an employee loyalty index. It shows how willing people are to recommend the company as a place to work. The metric grew out of customer NPS and works the same way: one question, a 0 to 10 scale, three groups of respondents.

• Promoters — gave a 9 or 10. Ready to invite friends to join and defend the company in conversation.
• Passives — 7 or 8. Generally satisfied, but will leave if offered clearly better conditions.
• Detractors — 0 to 6. Unhappy, and most likely talking about it — to colleagues, friends, in employer reviews.

eNPS = % promoters − % detractors

Passives aren't part of the formula, but they count towards the total number of respondents. That's why percentages are calculated from everyone who answered, not from total headcount.

A worked example

Say a call centre in Astana has 52 operators and 40 of them responded. Of those, 18 gave 9–10, 14 gave 7–8 and 8 people gave 0 to 6.

eNPS = 45 − 20 = +25. It's written as a number without a percent sign: "eNPS +25".

Always state the response rate next to it: 40 out of 52, or 77%. If only a third of staff responded, the result describes the people who chose to answer, not the whole workforce.

How to word the question and what else to ask

Word the main question the same way every time, or you won't be able to compare results:

Don't switch to a 1–5 scale or add labels to every number: the formula is built for 11 points. Don't put the question at the end of a long survey — by item forty, people are answering on autopilot.

Open questions

Right after the score, you need at least one open question. That's where comments like "the evening shift has been short-staffed for two weeks" or "my manager only ever talks about numbers" show up. That's the material you make decisions with.

If you want more structure, add 3–5 closed questions on specific topics: workload, manager, pay and recognition, development, working conditions. Then the comments and scores will show which topic is dragging the index down. Keep the whole employee engagement survey to 3–5 minutes.

How often to run the survey

For most companies, eNPS once a quarter works well. That's enough to see the effect of changes without the survey turning into a routine.

• Once a year — too rare. Between surveys, managers change, key people leave, and you won't know what affected the number.
• Once a quarter — eNPS plus 1–3 open questions. A solid baseline.
• Once a month — a short pulse survey of 3–5 questions, where eNPS can alternate with questions on a specific topic. Works if the company is growing fast or going through big changes: an office move, a new schedule, a new manager.

There's a simple rule: don't launch the next survey until you've shown people what you did after the previous one. Otherwise each new score will be lower than the last — and the survey itself will be the reason.

New hires are worth surveying separately: they have a different starting point, and their answers are best viewed separately, broken down by tenure. If your company has high turnover, compare department eNPS with where people leave most often — read more about calculating it in our article on employee turnover.

Anonymity: how to keep it in a small team

If an employee suspects their answer can be traced back to them, they'll give an 8 and write nothing. The index will look decent and be useless.

In a company of 300 people, anonymity is easier to guarantee. The difficulties begin when a department has 4 people and the manager knows how each of them writes.

If the whole company is 15 people, you don't need a breakdown by department. Calculate overall eNPS and work with the open answers. For a team like that, an honest conversation at an all-hands meeting after the survey often gives more than the index itself.

How to read the result

eNPS ranges from −100 (all detractors) to +100 (all promoters). A positive score means there are more promoters than detractors. Beyond that, it gets nuanced.

Don't look for a "normal market eNPS". The figures floating around the internet were collected in different countries, industries and in different ways. A warehouse with night shifts and an IT team of ten don't need to have the same score. Compare the company with itself.

What matters more than the absolute number

• The trend. +10 after +30 is a reason to dig in. +10 after −15 is noticeable growth, even if the number looks modest.
• The breakdown by department. An overall +25 can be made up of +50 in sales and +5 in the warehouse. The company as a whole is doing well, but the warehouse isn't.
• The structure of responses. An eNPS of 0 can come from 10% promoters and 10% detractors, or from 40% and 40%. In the second case the company is split, and that's a completely different situation.
• Response rate. If 50% of a department responded instead of the usual 90%, that's a signal in itself: people either don't believe in the anonymity or don't believe anything will change.

The threshold for a meaningful change is higher for small groups. In a department of 10, one employee moving from passive to detractor shifts the index by 10 points. Don't launch a post-mortem over a swing that comes down to one or two people.

What to do after the survey

A survey with no follow-through is worse than no survey at all: it teaches people that nobody cares what they think. Here's what to do after you measure.

1. Sort the comments by theme. List 5–7 recurring themes and count how often each one comes up. Separately flag themes that concern specific departments.
2. Choose 2–3 priorities. Don't try to fix everything. It's better to take one quick win (for example, water and air conditioning in the warehouse) and one systemic issue (for example, regular feedback from managers).
3. Assign an owner and a deadline. Every action needs a named owner and a date. "HR will work on engagement" is not a plan.
4. Discuss with department heads. Show them their results and ask them to come back with two actions of their own. If the problem is the manager, that's a one-on-one conversation between them and the HR director, not a topic for a general meeting.
5. Give the team feedback. Within two to three weeks of the survey, tell people how many responded, what the result was, what you heard and what you'll do. Be honest about what can't be changed and explain why.
6. Show the changes before the next survey. Before the new survey, briefly remind people what was done after the last one.

If the comments show the problem lies with specific managers, it's worth adding a more precise tool to eNPS — for example, 360° feedback, where the manager is rated anonymously by their direct reports and peers.

Mistakes that make eNPS worthless

• Calculating the average score instead of eNPS. An average score of 7.5 and eNPS are different metrics. The average smooths things out; eNPS shows the imbalance between the satisfied and the dissatisfied.
• Changing the wording of the question. "Would you recommend the company" and "Are you satisfied with your job" get different answers. After a wording change, you have to start the trend from scratch.
• Linking eNPS to a manager's bonus. Then the manager will start "preparing" the team for the survey, and the index will stop measuring anything.
• Hunting for the authors of negative comments. One such case is enough to put an end to honest answers — in that department and the ones next to it.
• Surveying in the middle of a crunch or right after bad news. Run surveys at the same point in each quarter so the comparison is fair.
• Publishing department results as a ranking. A "best and worst department" table on the big screen turns the survey into a competition and the worst department into the guilty party.
• Going quiet after the survey. The most common and most costly mistake: the response rate drops next time, and there are fewer comments.

Where to start tomorrow

The first survey is a baseline, not a verdict on HR's work. Its value becomes clear with the second and third surveys, once there's a trend and you can see which actions moved it.