Employee wellbeing KPIs: the six numbers to take to the board
Six: adoption, active participation, average sleep, average stress and recovery, average activity, average BMI. All by group, never by person, and all read as a change against a baseline.
Why isn’t “how many people used the benefit” a KPI?
Because it measures consumption, not wellbeing. Knowing that 12% of employees opened the meditation app, or that 340 gym sessions were booked, tells you how much a service was used. It does not tell you whether anyone sleeps better, recovers more or moves more.
Usage rates have a structural flaw too: there is no baseline. The first month is always the best month, because novelty inflates it, so every comparison after that is a comparison with a peak. A wellbeing KPI has to answer a different question: compared with how the group was before it started, what changed?
Answering that takes objective, continuous data, aggregated by group. That is the definition of measurable employee wellbeing, and it is where the six numbers come from.
Adoption and active participation: how are they calculated?
These are the two programme KPIs. They come before the others, because without participation there are no averages.
Adoption = people who activated the programme / eligible people. The denominator matters: eligible means the people to whom the programme was actually offered, with a device available.
Active participation = people with data synced in the last 7 days / eligible people. This is the honest number. Adoption tells you who downloaded the app; participation tells you who uses it. The seven-day window is the one Vitality+ uses in the HR dashboard: short enough not to count people who have stopped, long enough not to punish a weekend without a wearable.
Two rules for reading them. First: active participation always falls after launch, and the number that matters is where it settles after 8 to 12 weeks. Second: the gap between adoption and participation is itself information. If many activate and few stay, the problem is the experience, not the communication.
The aggregated indicators: sleep, stress, recovery, activity, BMI
These are the four (or five) outcome KPIs. Each is a group average, calculated only when the group clears the aggregation threshold, and presented together with the change against baseline.
Average sleep: hours per night for the group. The share of the group above or below a threshold (7 hours, say) is useful too.
Average stress and recovery. Read them together: medium stress with high recovery describes a group that is coping; high stress with low recovery describes a group that is wearing down.
Average activity: activities, steps or active minutes per day.
Average BMI.
An outcome KPI only means something for a group. Vitality+ shows these averages for groups of 30 or more people, with at least 10 people with data on the day and a sufficient gender mix; below the threshold the value is hidden.
How do you read a trend by team and site?
The trend is the difference between the current average and the same group’s baseline, expressed in absolute terms (+18 minutes of sleep, say) and as a percentage.
Five habits avoid almost every mistake.
Use four-week rolling averages, not the day’s value.
Compare each group with its own baseline, never two groups with each other in absolute terms.
Check that group composition is stable. If half the team changed in three months, the trend compares different people.
Do not read small movements as results. ±2% on a group of 30 people is noise; +8% of sleep that holds for eight weeks is a signal.
One page for the board
The format that works fits on one page: six numbers, each with value, baseline, change and the threshold in a footnote.
| KPI | Current value | Baseline | Change |
|---|---|---|---|
| Adoption | 41% | 35% | +6 pp |
| Active participation (7 days) | 29% | 27% | +2 pp |
| Average sleep | 6 h 52 min | 6 h 34 min | +18 min |
| Average stress | 62/100 | 66/100 | −6% |
| Average recovery | 71/100 | 65/100 | +9% |
| Average activity | 7,900 steps/day | 7,100 steps/day | +11% |
The values in the table are an example of format, not real data. Below the table, three lines: what the data says, what it does not say, what you propose.
Related questions
What is a good adoption rate for an employee wellbeing programme? It depends on how it is offered and how many people already own a wearable, so treat generic benchmarks with suspicion. The useful comparison is between your own sites and over time: adoption that grows after launch, and active participation that settles instead of collapsing, tell you more than any industry average.
How often should wellbeing KPIs be updated? The dashboard updates daily, but the useful reading is weekly for HR (four-week rolling averages) and quarterly for the board.
Can you compare KPIs across two sites? Yes, on the trend, not on the absolute value. Two sites can start from different sleep or activity levels because of composition, age and type of work. What you compare is how far each has moved from its own baseline.