France has 8% engaged employees, against 12% in Europe and 20% worldwide (Gallup, 2026). Yet most startups run their employee experience on intuition, or on an annual survey nobody acts on. Eight metrics, one cadence, a one-page dashboard: here is how to measure what matters, without the bureaucracy.
Updated 17 September 2026 · 8 min read
Employee experience is the sum of everything a person lives through in your company, from the first contact with a job ad to the day they leave, and beyond. It covers hiring, onboarding, management, tools, pay, development, offboarding. It is not the foosball table or the afterworks. It is what someone feels on a Sunday evening when they ask themselves whether they want to go back.
You measure it for three concrete reasons:
Key point: measuring employee experience is not about launching a big annual survey. It is about tracking a few simple metrics, on a regular cadence, and above all acting on what they reveal.
These national averages are your benchmarks. A startup that measures already has an edge: most of its competitors in the talent market do not.
Each metric answers one precise question. If you do not know which decision a number will trigger, do not track it.
| Metric | Formula | The question it answers | Benchmark |
|---|---|---|---|
| eNPS | % promoters (9-10) − % detractors (0-6) on "would you recommend us as an employer?" | Are my people proud to work here? | > 0 acceptable, > 20 good, > 50 excellent (common scale) |
| 12-month retention | Employees still in place 12 months after hire ÷ hires in the cohort | Do I keep the people I hire? | About 1 permanent contract in 3 ends before one year in France, 1 in 5 for managers and professionals (DARES) |
| Early turnover | Departures before 12 months ÷ hires over the period | Does my hiring → onboarding chain hold? | Aim for < 10%. See reduce early turnover |
| Time-to-productivity | Days between arrival and reaching the 30-60-90 plan goals | Does my onboarding save time? | Set per job family; the trend matters more than the absolute value |
| Absenteeism | Days of absence (sickness, excluding leave) ÷ theoretical working days | Is a team cracking? | 4.3% in the French private sector in 2025 (Malakoff Humanis) |
| Participation rate | Respondents ÷ people surveyed, on each survey | Do my people still believe it is worth answering? | < 60% is a signal in itself |
| Internal mobility | Roles filled internally ÷ roles filled | Can people grow here without leaving? | Internal candidates convert 32 times better than inbound applicants (Gem, 2026) |
| Referral rate | Hires from an employee referral ÷ total hires | Do my teams recommend the company with their actions? | About 10% of hires in companies that run referrals (Hellowork, 2025) |
The eNPS (Employee Net Promoter Score) borrows the logic of customer NPS. One question, "on a scale of 0 to 10, how likely are you to recommend our company as an employer to a friend?", and one subtraction: the percentage of 9s and 10s (promoters) minus the percentage of 0 to 6 (detractors). 7s and 8s do not count. The score ranges from −100 to +100.
Example: 40 answers, 14 promoters (35%), 18 passives, 8 detractors (20%). eNPS = 35 − 20 = +15.
Three rules to make it useful:
Known limit: eNPS measures an intention, not a behaviour. Cross it with retention and referrals, which measure what people actually do.
The 60-question annual survey is the best way to do nothing with the results. Go for a short rhythm and few questions.
Day 30, 90, month 6
Three questions to each new hire: role clarity, quality of the welcome, intention to stay. The first eNPS is taken at day 90.
Every quarter
eNPS + 3 to 5 targeted questions (workload, management, recognition). Five minutes to fill in, results shared within 15 days.
Every month
Retention, early turnover, absenteeism, internal mobility, referrals. Pulled from payroll and the ATS, no survey needed.
At every exit
Six open questions, an exit eNPS, a written summary. The most honest data you will ever get.
For a startup of 30 to 150 people, one spreadsheet tab is enough. One row per month, one column per metric, a three-level colour code. What it should contain:
The fourth block is the only one that really counts. A metric that triggers no action is just one more number.
FAQ
eNPS ranges from −100 to +100. The commonly used scale treats a score above 0 as acceptable, above 20 as good and above 50 as excellent. The most useful reading, though, is the trend over time and the comparison between teams: a +10 that improves at every wave beats a +40 that erodes. Always add an open question to understand the score.
Eight are enough: eNPS, 12-month retention, early turnover, time-to-productivity, absenteeism, survey participation rate, internal mobility and referral rate. The first four are read by hiring cohort, the others company-wide and by team. Each should answer a precise question and trigger a decision.
An onboarding pulse at day 30, day 90 and six months for every new hire, a quarterly team pulse with eNPS and three to five targeted questions, a monthly review of HR metrics from payroll and the ATS (retention, absenteeism, mobility, referrals), and a systematic exit interview. Avoid the long annual survey: it comes too late and rarely gets acted on.
Ask: on a scale of 0 to 10, how likely are you to recommend our company as an employer to a friend? Count the percentage of 9 and 10 answers (promoters) and the percentage of 0 to 6 answers (detractors). The eNPS is the difference between the two. Answers of 7 and 8 are neutral. Example: 35% promoters and 20% detractors give an eNPS of +15.
According to Gallup (State of the Global Workplace 2026, 2025 data), 8% of French employees are engaged, 72% are not engaged and 20% are actively disengaged. The European average is 12% and the global average 20%. France ranks second to last among the 38 European countries studied.
Related guides
Sources
Contact
We set up your employee experience dashboard in four weeks: metrics, cadence, first pulse, feedback loop.
Book a call →