Subcultures

Ask people in two departments of the same organization what it is like to work there and you will frequently get accounts that cannot both be true.

Both are accurate. They are describing different cultures that happen to share a payroll, and the difference between them is usually larger than the difference between either one and a comparable organization down the road.

Culture is local because consequences are local

If culture is the behaviour that gets rewarded, tolerated, and repeated, then the relevant question is who administers those consequences day to day.

For nearly everybody in an organization, that is their immediate supervisor. The chief executive sets the ceiling and the floor. The supervisor sets what actually happens on Tuesday, which is what people are responding to.

Which means an organization has roughly as many operative cultures as it has people with real authority over others, and they vary by however much those individuals vary.

This is why organization-wide culture work so often produces nothing. It addresses a level at which culture is not primarily determined.

Which variation is legitimate

Not all of it should be flattened, and a leader who attempts uniformity across genuinely different work will do damage.

Legitimate variation follows the work. A maintenance crew and a finance team should not have identical norms about interruption, about how disagreement is expressed, or about the pace at which decisions are made. Those differences reflect real differences in what the work requires, and imposing a single register on both makes one of them worse.

Illegitimate variation concerns the things the organization has said are not negotiable. Whether a safety standard is observed. Whether people are spoken to civilly. Whether bad news travels. If one unit holds these and another does not, that is not local character. It is a standard being enforced in one place and not another, and everybody in the organization can see it.

The practical rule: a small number of things must be identical everywhere, and everything else should be permitted to differ. Most organizations get this backwards, mandating uniformity on style and tolerating variance on standards.

The good unit inside a poor organization

A supervisor can build something considerably better than the organization around it, and this happens more often than the literature suggests.

Such units are recognisable: people transfer in and do not leave, the work is done properly, and the supervisor absorbs a great deal of friction from above so their people do not have to. It is real leadership and it is frequently invisible to anybody senior.

It is also fragile in a specific way. It depends entirely on one person, it is not encoded in anything, and it ends when they are promoted, retire, or leave. The successor inherits a team accustomed to a standard the organization does not require, and the reversion is fast and demoralising.

For senior leaders the implication is direct. Where a unit is visibly better, the question is what that supervisor does that the others do not, and whether any of it can be made structural rather than personal.

The poor unit inside a good organization

The inverse is more consequential and harder to see, because organizational indicators average it away.

One supervisor whose people are treated badly, in an otherwise decent organization, will produce turnover in a single unit while every aggregate measure looks fine. The organization’s overall survey result is satisfactory. The people in that unit are having an entirely different experience and have concluded, reasonably, that the organization tolerates it.

Which produces the most useful diagnostic in this entry: look at turnover, grievances, and transfer requests by supervisor rather than by department or organization. The distribution is almost never even, and the concentration is the finding.

Aggregate numbers are specifically designed to obscure this, and a leader who has never disaggregated them does not know what they have.

Working through the layer that matters

If culture is administered by supervisors, then culture work is supervisor work, and it consists of three things rather than a programme.

Be explicit about the small number of things that do not vary, so a supervisor knows exactly where their discretion ends. Hold them to those, including the well-liked ones and the ones whose numbers are good. And select for it, because whom you promote into supervision is the single largest determinant of what the organization’s culture will be in five years.

That last point is worth sitting with. An organization that promotes on technical output alone will fill its supervisory layer with people selected on a criterion unrelated to the thing that determines its culture, and will then run culture programmes to address the result.