Questioning teams and individuals

Two levels, two different questions

The collective answers questions about how things run: how decisions travel, where the handoffs jam, which rules are actually applied. These are shared facts, observable by several people, and a gap between teams means something.

The individual answers questions about experience: workload, meaning, recognition, intent to stay. These are personal states — averaging them loses what matters, and they are only worth anything when tied to a context: which team, which role, which site.

So the same subject gets asked twice. "Is the reorganization holding?" is, on the collective side, the clarity of scopes and the number of decisions escalating one level up. On the individual side, it is perceived workload and how legible one's own role is.

The questionnaire: cover, compare, locate

A questionnaire covers the whole scope and tells you where to look — not why. Its strength is comparison: between teams, between roles, between sites, and from one wave to the next.

Three conditions make it usable. The same questions, asked the same way, to everyone. A breakdown of the organization defined before it goes out: an answer with nothing to attach it to cannot be cross-read. A scale that stays stable over time, if you intend to measure again.

The main trap is the average. An overall 6.8 out of 10 cannot be acted on. The same answers re-read team by team usually reveal two or three pockets far below and a flat remainder: that is the subject, not the average.

The interview: understanding the mechanism

Interviews are how you understand why. They give you causal chains — "sign-off goes up to the director because the manager has no budget delegation" — that no questionnaire produces.

They do not need to be exhaustive. Fifteen or so well-spread interviews — managers and non-managers, mixed seniority, at least one secondary site — is enough to see the same mechanisms recur. When two consecutive interviews bring nothing new, you have enough.

They do require an explicit frame: what will be reported back, in what form, and what will never be attributed by name. Without that frame, you get the official version.

The order that works

In most situations: questionnaire first to locate, interviews afterwards on the areas the questionnaire pointed at. You walk into the interview with a hypothesis rather than a blank page, and you only spend human time where there is something to understand.

The reverse order is justified when you do not yet know what to ask: a few exploratory interviews then let you write a questionnaire that speaks the house language, rather than a generic one nobody recognises themselves in.

What to settle before the first question

The scope: who is concerned, and why them. Asking everyone so as not to offend anyone dilutes the signal and tires the organization.

Anonymity: anonymous does not mean unattached. You can not know who answered while still knowing which team the answer came from — that condition is what makes the analysis actionable. Set a threshold, five answers for instance, below which no team-level result is displayed.

The report back: announce the date before you ask the question. A consultation with no feedback costs more than no consultation at all — the next one will find nobody.

What comes next: what you will do with the results, and what you will not. An open-ended expectation turns into disappointment mechanically.