A deteriorating climate with no identifiable cause: locate before acting
Everyone feels it is going badly, nobody can say where. Why the average hides the problem, and how to get from a general feeling to two or three precise zones.
7 min readUpdated August 21, 2026
"The climate is bad" is a shared and unusable statement: it says neither where, nor since when, nor where to start. The work consists in cutting it up until you have something you can decide on — and that breakdown is almost always possible.
Why the average hides the problem
A deteriorating climate rarely produces a general decline. It produces two or three pockets far below and a flat remainder, which shows up as a slightly lower overall score: the worst possible result, because it is both alarming and unactionable.
The same answers re-read team by team, role by role, site by site almost always show something else: the subject is located somewhere. That is why what each answer can be attached to matters more than how refined the questions are.
The corollary: an organization that cannot break itself down cannot analyse its climate. Teams, roles, sites, tenure — the analysis grain is decided before the questions go out, not after.
Separate what people live from what they observe
Two registers get mixed into the notion of climate. Experience: workload, recognition, meaning, whether it is safe to speak. Functioning: decisions that never come, priorities that keep changing, missing means.
They influence each other, but you do not act on both the same way. Experience degraded by broken functioning is fixed by fixing the functioning; treated as a morale problem, it produces a wellbeing plan that feeds the cynicism.
So ask both, and look at which move together. A team whose experience is low while functioning is judged fine poses a different question from one where both collapse.
What the annual survey cannot do
One measurement a year always arrives too late for the deterioration it records, and too early for the action you have just launched. It is a snapshot, not a steering instrument.
On a climate that is deteriorating, a short measurement narrowed to the zones concerned and repeated a few months apart teaches more than a sixty-question annual survey — provided you do not change the questions between waves. An unstable scale makes comparison impossible.
The other limit of the annual survey is social. If it is never followed by anything visible, the response rate drops with every wave and the answers flatten: the organization tells you what it is supposed to say.
From the zone to the mechanism
Once the zones are identified, they still have to be understood. Interviews here serve to establish a timeline: since when, and what changed at that moment — a scope, a departure, a project, a reorganization.
A plausible cause has to survive a simple question: why this team and not the one next door, which lives through the same thing? What distinguishes the two is usually the mechanism you are looking for.
Be wary of explanations that point at a person. They are sometimes right, often a simplification, and always a shortcut that spares you from looking at the actual working conditions: workload, means, decisions never made.
Report back without exposing anyone
A climate result is sensitive: it touches identifiable people as soon as the grain is small. Set a threshold — five answers, for instance — below which no team result is displayed, and hold it even when you are asked for an exception.
Report first to the teams concerned, with their managers, before any wide distribution. A manager who discovers their team's result in an executive meeting will not cooperate with the next measurement.
Finally, announce what will be done, by when, and what will not be addressed. On a deteriorating climate, silence is not neutral: it confirms the diagnosis.
Key points
- An overall climate score cannot be steered: the subject appears when the same answers are re-read by team, role and site.
- Separate experience from functioning — you cannot fix one with actions aimed at the other.
- A cause holds if it explains why this team, and not the one next door.
- A display threshold, the teams first, and a date announced before the first question.
These steps are tooled in Syntonie
Questionnaires, transcribed interviews, analysis by team and by role: the platform runs a firm's method on your organization, described once.