What is eNPS? How to calculate and use the Employee Net Promoter Score

The eNPS question

eNPS grew out of the Net Promoter Score, which Fred Reichheld of Bain & Company introduced in the Harvard Business Review in 2003. Where NPS asks customers whether they would recommend you, eNPS asks your own people, about you as an employer. Bain puts the question like this:

How likely would you recommend this company as a place to work?

Employees pick a number from 0 (not at all likely) to 10 (extremely likely), and that number puts them in one of three groups:

• Promoters (9 and 10) are your advocates: they recommend you without being asked.
• Passives (7 and 8) are content enough to stay, but not enthusiastic enough to recommend you.
• Detractors (0 to 6) would sooner talk a friend out of joining you.

Keep the wording identical from one round to the next. "Recommend as an employer" and "recommend as a place to work" sound interchangeable, but even a small change muddies the comparison with last time.

How to calculate eNPS

There are two steps. First work out what share of all responses came from promoters and what share from detractors. Then subtract the second from the first.

A worked example

An organisation with 80 employees runs an eNPS survey and 60 people respond:

The eNPS is 35 − 25 = +10. Three things to note:

1. Passives count towards the total. You divide by all 60 responses, not by the 36 promoters and detractors.
2. The result is a number, not a percentage. Write +10, not 10%.
3. An average tells a different story. A mean score of, say, 7.4 looks perfectly respectable and hides the fact that a quarter of your people would advise against joining you.

Rather not do the sums yourself? Our eNPS calculator does it for you, even from a pasted column of individual scores, and gives you the margin of error too.

What is a good eNPS?

It's the question everyone asks, and the honest answer is that your own previous result matters more than any national average. Plenty of benchmarks are in circulation, but they come from different sectors, countries, question wordings and points in time, and it's rarely clear how they were put together. So we don't quote any here.

A few things are certain, though:

• Above zero means more promoters than detractors. It's the only threshold that follows from the calculation itself.
• eNPS often runs lower than customer NPS. Bain & Company notes that employee scores can be substantially lower than customer scores. Employees see you from the inside and hold you to a higher standard. Don't compare the two.
• Your own trend is the best yardstick. Same question, same time of year, same group: then a shift means something.

Margin of error: how much does a change mean?

An eNPS from a sample is an estimate. In the example above, 60 of 80 employees responded. Work out the margin of error (at 95% confidence, corrected for having heard from most of the workforce) and the true eNPS lies somewhere between roughly 0 and +20. Had the same company employed 600 people, with the same 60 responses, the margin would have been almost twice as wide.

That has a practical consequence: a team that moves from +10 to +6 probably hasn't got any worse. Small shifts in small groups are mostly noise. The eNPS calculator works out the margin when you enter the number of employees you invited.

Common eNPS pitfalls

Turning the score into a target. Tie a manager's bonus to a higher eNPS and you'll soon find employees being asked to "just give it a 9". The score then measures something other than what you wanted to know. Use eNPS to learn, not to judge.

Reporting on groups that are too small. Show the eNPS of a team of three and everyone in it knows who the detractor was. Only report on groups large enough to keep individuals out of view. Our guide to anonymous employee surveys and the GDPR explains how.

Measuring without acting. People notice when nothing happens after a survey. Next time fewer respond, and those who do are less candid. After every round, say what you heard and what you'll do about it, even if that's "nothing yet, but here's what we're looking into".

Changing the question or the scale. A 1 to 10 scale instead of 0 to 10, or a different wording, makes your new score incomparable with the old one. Timing matters too: a survey straight after a restructure measures something different from one in a quiet month.

Asking only the eNPS question. One question is quick to answer, and that's its strength. But a number without an explanation leaves you guessing at the cause.

After the score: ask why

eNPS tells you where you stand, not what to change. A +10 might mean people love their colleagues and can't keep up with the workload. A −5 might be largely about one decision that went down badly last month. You can't tell from the number.

That's why the eNPS question is nearly always followed by an open one. These work well:

Alongside the eNPS question, Bain often suggests a second statement, my job inspires me, which it finds a better predictor of the energy and enthusiasm people bring to their work. It's a useful addition, but it doesn't tell you why either.

The trouble with open questions in a survey is that the answers tend to be short: "workload", "fine", "pay". A good interviewer would follow up: workload because of what, since when, in which kind of work? That's exactly the difference between a form and a conversation, which our guide to conversational surveys explores in more depth.

How often should you measure eNPS?

According to Bain, companies most commonly ask for eNPS feedback monthly, which gives a steady stream of input rather than one snapshot a year. For many organisations, quarterly is a more manageable rhythm: often enough to see a trend, not so often that people tire of it.

Whatever you choose, three rules of thumb:

• Don't measure more often than you can respond. Every round deserves feedback. If you can't give it monthly, measure quarterly.
• Pick a fixed moment. Compare the first quarter with the first quarter, not a summer survey with the December rush.
• Combine it with a broader survey. eNPS is a thermometer. Once a year, a fuller employee survey shows what lies underneath.