The pattern.
Burnout requires sustained effort under sustained load. That is worth reading twice, because it rules out the intuitive candidate. The employee who withdraws protects themselves. The employee who cares, absorbs, volunteers and stays late does not.
Commitment is not a defence against depletion — it is the mechanism that produces it. Engagement supplies the willingness to keep going; capacity is what is being spent. When the first is high and the second is low, the result is someone performing well right up to the moment they stop.
This is why the state is so often missed. It does not look like a problem. It looks like your most reliable people.
Why engagement scores hide it.
An engagement instrument measures whether people are motivated, connected and enabled. A committed-but-depleted employee answers those items honestly and positively: they do find the work meaningful, they do trust their manager, they are invested in the outcome.
Nothing in the instrument asks whether they can sustain it. So the survey returns a high score, the organisation concludes it is healthy, and the depletion continues unmeasured until it surfaces as a resignation.
Adding a few wellbeing questions to an engagement survey does not close the gap. Capacity items need behavioural phrasing and a bounded recent window — “over the past two weeks I have woken up feeling rested” rather than “my wellbeing is good” — because the second asks people to score themselves against a standard they invent, and correlates more with self-image than with any observable state.
What the data shows.
Two separate strain measures in our validation study separate sharply across the Thrive Index — and both are measures of the absence of strain rather than of burnout itself, which we do not diagnose.
6.8×
More people are not regularly overwhelmed — 15% in the bottom decile against 98% in the top.Overwhelm · bottom → top Thrive decile · n=10,000+
4.6×
More people are not running on empty — a separate energy measure, rising 17% to 79% across the same deciles.Energy · bottom → top decile
+36 / +50
Capacity moves 49.6 → 85.7 and engagement 48.0 → 97.5 across a 42-point Index spread — they correlate enough to combine, and diverge enough that combining alone hides the diagnosis.Index spread · decile to decile
What it looks like before the resignation.
Observable patterns, not a diagnosis. Each is individually unremarkable, which is why the combination matters more than any one of them.
Compensation
The same work taking longer, absorbed by working later rather than by asking for help. Output holds; the hours quietly extend.
No reset
Monday starts where Friday ended. A weekend stops changing anything — the clearest early signal available.
Narrowing
Discretionary contribution stops first. Still delivering the role, no longer volunteering beyond it.
Sudden exit
A resignation with no prior signal, from someone who never complained. The absence of complaint was the signal.
What to do about it.
The response that fails is motivational. Recognition, purpose messaging and engagement initiatives all ask for more of the resource that is already exhausted — and they are the natural response to a state that presents as high commitment.
01
Measure capacity separately, on the same instrument and in the same window as engagement. Two numbers measured six months apart cannot be read against each other.
02
Look at recovery first. It is the domain that declines earliest, and the one where a falling score is still reversible.
03
Reduce load or improve recovery conditions — out-of-hours norms, deadline design, predictability. Structural changes, not encouragement.
04
Check your highest-engagement teams specifically. On this pattern they are the ones to worry about, which inverts how most organisations read their own results.
The four states
Where this sits among the combinations of capacity and engagement: the Thrive Matrix.
What this does not claim.
These are associations from a single cross-sectional study, and they concern strain rather than clinical burnout. We do not claim that a capacity score diagnoses burnout, predicts an individual case, or substitutes for occupational health assessment.
The narrower claim is the useful one: strain concentrates where capacity is low, it is measurable at group level before people report being in trouble, and it is invisible to engagement measurement alone.
Method & sources.
Figures come from thrive at work’s own validation study (n=10,000+, 2025–26), fielded across the US and Europe. Comparisons are bottom decile versus top decile of the Thrive Index. The overwhelm and energy items measure the absence of strain, not clinical burnout.
These are associations measured within a single cross-sectional study, not causal claims. No score is displayed for any group of fewer than five people.
- Thrive Index validation study. thrive at work, n=10,000+, 2025–26, fielded across the US and Europe.
- Domain foundations. 140+ peer-reviewed papers underpinning the eight domains — see the methodology.