How to measure employee engagement: eNPS and beyond

Employee engagement is something everyone calls important and few companies measure regularly. The reason given by those who do not is usually the same: "we run surveys but nothing comes of them". The problem is rarely the survey itself; it is the loop around the survey.

Engagement is not satisfaction

Satisfaction is an employee being content with their job. Engagement is an employee voluntarily putting extra effort into their work and their company. The two do not always travel together: a comfortable but indifferent employee may be satisfied without being engaged.

This distinction changes the questions too. "Are you satisfied with your pay?" measures satisfaction; "Would you recommend working here to a friend?" gets closer to engagement.

How eNPS is calculated

eNPS (Employee Net Promoter Score) rests on a single question: "On a scale of 0–10, how likely are you to recommend this company as a place to work to a friend?"

Answers fall into three groups:

• 9–10: promoters
• 7–8: passives
• 0–6: detractors

The score is the percentage of promoters minus the percentage of detractors, a number between –100 and +100. Passives do not enter the subtraction but do count in the total.

Example: of 100 responses, 40 are 9–10, 35 are 7–8 and 25 are 0–6, so eNPS = 40 – 25 = +15.

The strength and the limit of eNPS

Its strength is simplicity: one question, quick to answer, high participation, and a number that is easy to compare over time.

Its limit comes from the same place. A single number does not say why. That is why eNPS should always be asked together with an open question: "What is the main reason for your score?" The real information is in those answers.

The second limit is benchmarking. Comparing with industry averages is tempting but unreliable because of methodological differences. The meaningful comparison is your own history.

Annual survey or pulse?

The classic annual engagement survey is comprehensive but slow: a problem measured in March has an action plan by September.

Pulse surveys shorten that delay: two or three questions every two weeks or every month. They show the trend early and keep the window for intervention open.

The model that works best combines the two: one comprehensive survey a year, short pulse measurements in between. We covered the practicalities in employee surveys and pulse measurement.

Anonymity has to be real

The single most important factor in participation is whether employees genuinely believe in anonymity. If they do not, they either skip the survey or give the safe answer; both make the data useless.

Providing anonymity technically is not enough — it has to be explained in a way people understand. Breakdown reporting in small teams is a specific risk: showing the result of a four-person team separately effectively removes anonymity. A minimum-response rule (groups under five are not reported separately, for example) closes that gap.

A survey whose results are not shared is the last survey

The most critical step in engagement measurement is sharing the results. For most employees a survey is a test of hope: will what I said change anything?

If results are not shared, the answer is "no" and participation in the next survey drops. Bad results have to be shared too; hiding a bad result signals that something already known is being concealed.

The share-out has three parts: what came out, which three topics we are focusing on, and when we will report back.

Action happens at team level

Company-wide action plans are usually ineffective because problems differ by team. In the same company one team may be struggling with workload and another with uncertainty.

The most productive model is therefore for every manager to see their own team's results and agree two or three concrete actions with the team. The central plan sets the frame; the team produces the change.

Participation rate is the first indicator

Before looking at survey results, look at the participation rate. Below 40%, results do not represent the whole workforce and usually reflect the most engaged or the most frustrated group.

What raises participation is predictable: a short survey, one that can be completed during working hours, mobile access, credible anonymity, and the results of the previous survey having been shared.

The last is the strongest. An opening message that recalls last year's results and the actions taken raises participation noticeably.

Read by segment

The company-wide eNPS is a single number and usually carries no actionable information. The real information is in the breakdowns: department, location, tenure, age group, working pattern.

Two breakdowns are especially instructive. Tenure shows at which stage engagement drops; in most companies the six-month to two-year window is the weakest point. Working pattern reveals whether there is a gap between remote and office staff.

Do not turn the number into an obsession

Making eNPS a corporate target corrupts the measurement. Managers expected to raise the score start managing the survey instead of solving the problem: influencing who fills it in, choosing the timing, steering expectations.

The healthier approach is to use the score as a thermometer. Make the thermometer the target and you will break the thermometer rather than lower the fever.

Combine with leaver data

The most powerful use of engagement data is comparing it with actual attrition. Do teams with low eNPS really have higher turnover?

If the two line up, the measurement is working and gains predictive power. If they do not, either the survey is not asking the right questions or people are leaving for other reasons. There is something to learn in both cases.