eNPS: How to Calculate It Correctly, What to Show the CEO, and Why 40 Isn't 'Excellent'

The company's eNPS is 22. Sales has an eNPS of 47, engineering has an eNPS of −8. On the overall chart it all looks like "fine, slightly above average." In reality the company lives in two different worlds, in one of which half the team is already updating their resume. You cannot average it.

eNPS is the most popular and most frequently confused metric in HR. It has a reputation as "one question — the whole truth about the company." That's convenient for slides, but dangerous for decisions. In one cycle it can throw a team into euphoria, in the next into panic, and both times it won't be because of real change but because of the sample and the wording.

Let's break it down without fluff: what eNPS actually measures, three typical traps, why industry benchmarks are deceptive, the one chart to show the CEO, and the cases where the metric is better thrown out entirely.

The Formula — and What It Actually Measures

eNPS — Employee Net Promoter Score — is an adaptation of the classic NPS to the employer–employee relationship. One question:

"How likely are you to recommend our company as a place to work to your friends and acquaintances? On a scale from 0 to 10."

Respondents are split into three groups by their answer:

Group	Score	eNPS formula
Promoters	9–10
Passives	7–8	eNPS = % promoters − % detractors
Detractors	0–6	(range: −100 to +100)

The key thing often missed: eNPS measures only the intention to recommend, not engagement, not satisfaction, and not engagement in the broad sense. It's about loyalty as a mild behavioral intention, nothing more. When HR writes "engagement = 47" on a slide and it's actually eNPS, that's a rename, not new information.

And there are three traps almost every company falls into when they start calculating.

Trap 1. A small sample. Fewer than 30 respondents in a cohort is statistical garbage. You have a department of 12, six scored it a 9, six scored it a 6 → eNPS = 0. The next week one of the sixes bumps to a 9 → eNPS = +16.7. That's not growth, it's noise.

Trap 2. A one-off spike from an event. The survey ran a week after a salary increase → eNPS rose 15 points. The survey ran a week after two people were let go → it dropped 12. To interpret it, you need a baseline — at least 4 quarters.

Trap 3. Averaging across segments. The most common. The overall figure "company eNPS = 22" hides a breakdown of "sales +47, engineering −8," and decisions made on the overall number hit the wrong part of the team. Without segmentation, eNPS can't be interpreted.

Why Industry Benchmarks Are Deceptive

The most common question after the first measurement: "Is that a lot or a little? What eNPS counts as good?"

And the typical HR-consultant answer: "In tech the average is +30, you're at +22 — you need to grow." That's a bad answer.

Multiple methodologies. NICE Satmetrix, Bain, numerous internal methods. The scales can be 0–10, 1–10, 1–5 — and each gives different numbers.

Cultural differences in responses. Respondents in some countries, on average, score 1–2 points lower than American respondents on the same scale for the same actual satisfaction — a stable finding in cross-cultural research. So an eNPS of +10 in one market can correspond to an eNPS of +30 in the US for the same real employee experience.

Anonymity vs identification. Companies that run anonymous surveys (with k≥5 anonymity assurances) get lower but more honest numbers. Companies where employees feel their answer can be traced get inflated ones.

Selection bias. Who responded? The 30% most loyal, or 90% of the whole company? A real eNPS from 30% is the eNPS of the loyal, not of the company.

The honest comparison is your eNPS against your own previous quarter, with a stable methodology. Your own trend → you can trust it. "The industry benchmark" → better to ignore, or at least not put it forward as the deciding argument.

And separately: "eNPS = 40 is excellent" is a marketing cliché, not data. At one company with +40 the climate will be ideal; at another, a toxic culture held together by one star salary everyone is afraid to lose. Same number, different reality.

What to Show the CEO — One Chart Instead of Twenty

If you have a quarterly review today and you're preparing an eNPS dashboard — don't bring the CEO twenty charts. This is the most common HR mistake in working with stakeholders.

The CEO scans the dashboard in 10 seconds and is left with one impression: "all fine" or "all bad." Twenty charts don't help them make a decision, they dissolve attention.

Instead — one composite chart with three layers.

Layer 1 — the trend line (company eNPS by quarter). The CEO sees whether the number is moving up or down. One line, four points — easy to read.

Layer 2 — the range across departments (vertical min–max bars). This shows how widely the segments diverge. If the bar is short, the company is homogeneous. If it's long, you have two different worlds, and the average is deceptive.

Layer 3 — the response rate (under each quarter). This is the data-quality indicator. If the response rate is 30%, eNPS doesn't represent the company at all. The CEO sees "RR 73%" and understands the data is reliable.

There should be no industry benchmarks on this chart. They distract and don't help decisions. You're talking about your company over time.

And one more important gesture: before you bring the chart to the CEO, talk it through with yourself — "what will we do differently depending on which number comes out?" If the answer is "nothing, we'll just look," you don't need to show that chart.

When eNPS Is Useless

In several situations eNPS carries no information:

• A company under 50 people. The sample is too small for significance. Use a 1:1 with each person.
• Right after mass layoffs, M&A, or a major restructuring. The event's noise drowns out the signal; wait 2–3 quarters before trusting the numbers.
• Strong seasonality. In industries with heavy seasonal hiring, eNPS will swing with the season, not the culture.
• An obviously non-anonymous survey. If people fear their answers can be traced, the numbers are inflated by 10–20 points and don't reflect reality.

In these cases it's better to use behavioral metrics:

• Quit rate / voluntary turnover per quarter — far more honest than the intention to "recommend."
• Internal mobility rate — shows whether employees see a future here.
• A weekly pulse question: "How was your week?" on a 1–5 scale. This gives a more frequent, more honest signal than a quarterly eNPS.
• Portal signals: peer-recognition frequency, canceled 1:1s, activity in discussions. They lag changes by less than surveys do.

eNPS is a tool. Useful in some conditions, useless in others. Don't hesitate to tell the CEO "this quarter eNPS carries no information, let's look at retention."

The "eNPS That Doesn't Lie" Checklist

Before making a decision on the number, run through 7 points:

1. Sample ≥ 30 in the cohort. Less is noise, not a signal.
2. Anonymity guaranteed and visible. If people don't believe it, the numbers are inflated.
3. The same question wording quarter to quarter. Rephrasing changes answers.
4. Response rate ≥ 60%. Less, and you have selection bias.
5. Segmentation by department and tenure. Without it, the average is deceptive.
6. Comparison against your own past quarter, not the industry. The trend is your own story; the points don't exist in someone else's methodology.
7. One open "why?" question per measurement. A number without a reason is under-informative.

If all seven points are covered, eNPS works. If even one is missing, it either lies or shows nothing.

The Bottom Line

eNPS is the most popular HR metric, and used correctly it works. But "used correctly" means: one segmented number over time, without industry benchmarks, with a clear sample and a meaningful consequence for a decision. Without that, eNPS turns into a pretty number for a report that leads no one anywhere.

And one more thing — you cannot call eNPS "engagement," however much you'd like to. They're different metrics, and mixing them misleads both you and the CEO.

Next, HR-tech predictions for 2026. Where AI is headed, where deepfake interviews fit in, and what it all means for anyone building a corporate portal.

If you want the eNPS calculation template in Google Sheets with an automatic segment breakdown, it's linked in the card. No signup, copy it and adapt.