The Culture Audit

An afternoon, done alone, using material the organization already has. No survey, no external facilitator, and nothing announced.

The point is not to produce a report. It is to establish what the operative culture actually is, which most leaders have never determined and most believe they already know.

Question 1 — Where is turnover concentrated?

Pull voluntary departures for two years and sort them by the supervisor the person reported to, not by department.

Then add a column nobody normally adds: how good each of those people was. If the strongest are the ones leaving, the culture is selecting against capability regardless of what any survey reports.

What it tells you: where the culture is actually administered badly, and whether the organization is retaining the people it needs. Aggregate figures are specifically designed to obscure both.

Question 2 — What distinguished the last five people promoted?

For each, write one sentence on what actually decided it. Not the stated rationale. The operative reason, honestly.

Then read the five sentences together as a stranger would, and ask what somebody would conclude about the path here.

What it tells you: the organization’s real statement about what it values, which everybody below leadership has already read and acted upon.

Question 3 — What is the worst unaddressed instance of each stated standard?

Take the three or four standards that are supposed to be firm. For each, name the worst violation in the last year that produced no consequence.

If you cannot recall one, ask a supervisor. They will name it immediately, which is itself a finding.

What it tells you: the actual location of each standard. Everything above that instance is aspiration.

Question 4 — How long between the problem and leadership hearing about it?

Take four past incidents of any size. For each, establish when somebody in the organization first knew and when it reached the person who could act.

The interval is retrievable from emails, logs, and asking two people. Expect it to be longer than your impression.

What it tells you: the price the organization attaches to delivering bad news, which is the single most consequential cultural variable and the one most directly set by the leader’s own conduct.

Question 5 — How many exceptions were granted?

Against each firm standard, count the exceptions in the last twelve months. Include the ones granted for entirely good reasons, since those are most of them.

Then note who received them. An uneven distribution toward people who are articulate or senior is a finding in its own right.

What it tells you: whether each standard is a boundary or an opening position, and whether outcomes here are determined by who asks.

Reading the result

Do not aggregate the five into a judgment about whether the culture is good. That question has no useful answer.

Look instead for the two or three specific findings that were not what you expected. Those are what the exercise was for, and a leader who finds nothing surprising has almost certainly measured their own assumptions.

Then choose one. Not a programme, and not five simultaneous corrections. The single finding that would most change how the organization behaves, addressed through consequences rather than communication, and sustained past the point where you have stopped finding it interesting.

Cadence and handling

Once a year is right. More often and the same findings recur before anything has had time to move.

Keep it private. The instrument works because nobody is performing for it, and an audit that becomes a reported exercise will produce presentable answers within one cycle.

Where a finding concerns a specific individual’s conduct, that is a separate conversation held privately and properly, not an item in an audit. Merging the two destroys both.