When culture stops explaining.
How organizations acquire the appearance of identity.

Can a company change its culture while sincerely believing it is preserving it?

Novo Nordisk may be running that experiment in public. Its CEO, Mike Doustdar, wrote on LinkedIn that he is "far too small to completely change a culture as strong as Novo Nordisk's." He has spent 34 years in the company, starting as an office clerk in Vienna. There is no reason to doubt he means it.

Around that sentence, the company has moved fast. Doustdar took over in August 2025. A month later the company announced it would cut roughly 9,000 of its 78,400 positions, and he named what the change was for: a more performance-based culture. When the board and the controlling foundation disagreed over how far to rebuild the board itself, the chairman and six other directors stepped down. A year in, Doustdar told Danish media the company needed to train more, work more, get up earlier and go to bed later. Two unions objected publicly. And throughout, the message has stayed the same: the essence of the culture will be preserved.

If culture is mostly what a company says about itself, there is no contradiction. If culture is also what gets rewarded, expected and repeatedly done, the question gets harder.

I know the question because I have watched a version of it happen.

I once worked inside an organization widely believed to have an unusually strong culture. It had a long history, its own vocabulary, a recognizable way of making sense of itself. Employees did not talk about the culture as a management programme. They talked about it as an inheritance.

Then, within roughly two years, much of the operating reality changed.

New senior leaders arrived with different assumptions. There was no declared break with the old organization. The language survived. The purpose and the values stayed in place. But the sequence of appointments changed. Different behaviours became promotable. Information travelled through different people and carried different consequences. Decisions that once required explanation required less of it.

Some employees who had known the old organization began saying its culture had disappeared.

Which creates a problem. If the culture was genuinely deep, how could it change so quickly?

The easy answer is that it was never as strong as people believed. Mostly branding, nostalgia, a flattering story.

But that answer fails elsewhere. Some organizations stay recognizable through decades of turnover and upheaval. A historical study of Carl Zeiss traced the company from 1846 to 1990 and found that the founders' principles became lasting features of the organization through identifiable processes: early consequential decisions, teaching, role modelling, written rules. The imprints outlived the men who made them. They survived two wars and the division of Germany.

So both things are real. A strong culture can transform in two years. Another can stay recognizable for a century. And we explain both with the same word. If an organization changes fast, new leaders replaced its culture. If it persists, its culture was deeply embedded.

A concept that explains every outcome may be telling us less than we think.

A pattern becomes a personality

Culture is not imaginary. People inside organizations really do learn which deadlines are negotiable, which risks are tolerated, who may challenge a decision, what happens when bad news moves upward. Those regularities shape behaviour.

The trouble starts with what we do next.

Watch a colleague over a long period. Every meeting begins prepared. Promises are kept. Bad news arrives early. Eventually you compress it into a sentence: she is reliable.

The compression is useful. Then the logic quietly reverses. She behaves reliably because she is a reliable person. The identity now explains the very actions it was inferred from.

Organizations get the same treatment. A company repeatedly takes risks and becomes entrepreneurial. A company repeatedly ignores warnings and acquires a poor safety culture. The descriptions may be accurate. Accuracy is not explanation.

The pattern has become the explanation of the pattern.

We usually imagine it the other way round: the organization has an identity, and the identity produces the behaviour. Often it starts at the other end. Particular ways of deciding, promoting, rewarding and interpreting get repeated. The repetition becomes recognizable. People, inside and outside, attribute an identity to it. That is the reversal.

The identity is still real once it exists. It attracts some people and repels others. It gives employees language to defend or resist a decision. It shapes what boards and regulators expect. What the organization produced starts helping to reproduce it.

But before a story can reinforce a pattern, something has to keep making the story plausible. At Zeiss, principles were translated into decisions, taught to newcomers, modelled by seniors, written into rules. Later generations inherited the machinery for enacting them again.

An organization does not preserve an identity the way a statue preserves a face. It re-enacts one, continuously, or stops.

Naming is not explaining

After the loss of the Space Shuttle Columbia in 2003, the investigation board concluded that NASA's organizational culture had contributed as much to the accident as the foam that struck the wing. Fair diagnosis. The disaster could not be reduced to a damaged part or one person's error.

But the investigators did not stop at the word. They reconstructed how repeated foam strikes had become normal, how requests for imagery lost force as they travelled, where the authority to escalate actually sat. Their recommendations addressed technical authority and organizational learning, not attitudes.

Culture named the pattern correctly. It did not explain how the pattern was made.

Four realities inside one word

When people say "the company culture", they are usually compressing four different things.

There is the formal organization: what is officially supported through policies, incentives, budgets, reporting lines and authority.

There is the expected organization: what people anticipate will happen when someone reports bad news or challenges a target.

There is the performed organization: what is repeatedly done, the routines, workarounds, silences and exceptions.

And there is the narrated organization: what the company says about itself, in values, stories and symbols.

These four influence one another. They do not automatically line up, and they do not move at the same speed.

A company can formally protect dissent while everyone expects dissent to end careers. It can narrate customer devotion while paying for behaviour that harms customers.

When the four reinforce one another, the company looks coherent and its identity looks self-evident. When they separate, people start saying the culture has become hypocritical, or weakened, or has vanished.

When the words outlive the system

Wells Fargo made the separation visible. For years the bank narrated a commitment to helping customers succeed financially. Meanwhile its compensation systems put enormous weight on cross-selling, and employees opened millions of accounts customers had never asked for. In 2018 the Federal Reserve capped the bank's growth until it fixed its controls.

Calling this a toxic sales culture is reasonable. Opening the term shows more. The narrated organization promised relationships. The formal organization paid for sales. Employees learned from the consequences around them what the system actually required, and the performed organization followed. The bank never had to abandon its values in public. It only had to stop maintaining the processes that made them credible.

The words stayed. The system that made them credible did not.

That difference in speed matters. A policy changes overnight. Expectations change after people watch whether the policy bites. Behaviour shifts unevenly. A public identity can survive for years.

It is why an owner and an employee can sincerely disagree about whether a company has changed. The owner, watching across generations, sees continuity in purpose and symbols. The employee, living through two years of new appointments and new consequences, is in a different company. They are not describing the same object. They are watching different realities move at different speeds.

Novo Nordisk contains both right now. From the level of ownership, the changes are a necessary correction in defence of the company's long-term purpose. From below, some employees describe something being dismantled. Both readings can be sincere.

Continuity is something an organization does

Think of a mature garden. It looks effortless. The work producing that order is invisible, and when the maintenance stops nothing collapses on the first day. The paths remain. Some flowers come back on their own. The visible order outlives the processes that sustained it. Only gradually do the edges soften.

We talk as though culture accumulates like sediment, harder to remove with every year. But age does not create resilience. A pattern survives only while the organization keeps recreating the conditions that make it probable. Continuity is renewed or abandoned one decision at a time.

That resolves the paradox. An organization stays recognizable for a century when the processes that reproduce its patterns survive changes in people. A deep culture changes fast when a small coalition takes control of several high-leverage processes at once: appointments, information, incentives, priorities, what counts as normal.

The moment that taught me most about culture was quiet.

A difficult decision was on the table. Someone invoked one of the values, the real ones, the ones everyone in the room had been hired under. Everyone recognized the words. Nobody disputed their importance.

And the outcome did not move.

That is the earliest sign, and it is never a new values statement.

The narrative is still running. The reproduction has stopped.

Artificial intelligence makes that moment easier to hide. The narrated organization is becoming nearly free to produce. Values, messages and corporate voice can continue fluently after the systems beneath them have changed. When words cost almost nothing, they prove almost nothing. Credibility has to be earned where it always was: in decisions, appointments and consequences.

The experiment is public

Which brings the question back to Denmark.

Doustdar may be entirely sincere, and on his own terms he may be right. No single person changes a culture built over a century. But that answers the wrong question.

The question his framing cannot ask is whether the company is still reproducing the culture he says he cannot change.

The four realities give us something to watch.

The narrated Novo says the essence will be preserved.

The formal Novo has already moved: a majority of the board replaced, incentives rebuilt around performance, one position in nine gone.

The expected Novo is starting to show movement: in union statements, and in what some employees now say publicly.

The performed Novo will only become legible over the next few years: which risks get taken, which behaviours advance careers, which warnings still get heard.

The test is simple and cold. Watch whether the decisions, appointments and consequences that once made the old identity credible keep occurring.

If they do, Doustdar will be vindicated.

If they stop, the language will still be there, fluent as ever, describing an identity the organization no longer reproduces.