Pathological, Bureaucratic, Generative: The Culture Underneath the Work
In brief
Where Westrum's three-cultures model came from, how it reached software, what the evidence honestly supports, and the permission it grants: to change your culture, accept it, or finally have it confirmed.
Cooperation that should have been easy stopped at a silo. A person carrying bad news took a risk in one place and earned thanks in another. Some rooms seemed to belong to nobody; in others, everyone stepped up and owned a piece of the work.
Teams that needed each other stayed behind walls. A failure triggered a hunt for someone to blame here and a serious attempt to learn there. Good ideas appeared, passed through a few meetings, and disappeared without anyone trying them. If you have spent enough time trying to change an organization, you have probably met some version of these contradictions.
I met all of them. I watched agile, lean, and DevOps practices land differently from one organization to the next, even when the people used the same language and followed similar methods. Some of it seemed dumb. Some of it felt like nonsense. For years I had no name for what I was looking at.
One model, three cultures
Ron Westrum spent his career studying safety and information in complex, high-risk systems. He defined culture as “the organisation's pattern of response to the problems and opportunities it encounters.” His typology describes three recurring patterns: pathological, bureaucratic, and generative. (Ron Westrum, “A typology of organisational cultures,” Quality and Safety in Health Care 13, Suppl. II, 2004.)
Leaders create much of the pattern through what occupies their attention. A preoccupation with personal power, personal needs, and glory produces a pathological culture. Where rules, positions, and departmental turf hold the leader's attention, the culture turns bureaucratic. When the mission itself is the preoccupation, the culture becomes generative. People learn what matters by watching what leaders notice and what happens afterward. In Westrum's words, “Rewards, punishments, and resources follow the leader's preferences.”
Information in Westrum's model has more than volume. Its relevance, timing, and fit for the recipient matter too. His generative organizations run on one question: “who needs the information now?”
Westrum put the effects into a table about how information moves through an organization. The columns show the three cultures. The rows show six places where the pattern becomes visible.
| Organizational behavior | Pathological | Bureaucratic | Generative |
|---|---|---|---|
| Cooperation | Low | Modest | High |
| Messengers | Shot | Neglected | Trained |
| Responsibilities | Shirked | Narrow | Risks are shared |
| Bridging | Discouraged | Tolerated | Encouraged |
| Failure | Leads to scapegoating | Leads to justice | Leads to inquiry |
| Novelty | Crushed | Leads to problems | Implemented |
A leader can read across a row to compare how one behavior changes with the culture, then read down a column to see the larger pattern. Six separate frustrations turn out to be one pattern wearing six faces, and the table gives people a shared vocabulary for conditions that official values statements blur together.
Organizations rarely fit one column perfectly. A company can respond generatively in one department and bureaucratically in another. The same group can also react differently when the stakes change. Westrum described types. A compulsory maturity ladder with one correct destination for every context goes beyond his evidence.
I understood the three types long before I understood all six rows. The columns alone gave me enough to start seeing recurring behavior as part of a larger system. The situations still made no sense as operating choices. I could at least comprehend what was behind them.
The return address
Westrum presented the first version of the typology at a World Bank conference in 1988. The paper came sixteen years later, in a patient-safety supplement of Quality and Safety in Health Care. He was studying how information travels when mistakes can injure or kill people, and his cases drew from medicine, aviation, nuclear power, and NASA. Management journals and software conferences came later.
The healthcare origin stood out to me. Software people usually meet this model in a DevOps book, where it reads like it was built for engineering organizations. It was built for places where a missed signal can cost a life.
Safety practitioners noticed before software did. Patrick Hudson's safety-culture ladder extended the three types for oil and aviation safety programs, keeping the pathological and generative ends. (Flight Safety Australia, “Safety in mind: Hudson's culture ladder”.)
The DevOps community kept importing from fields where failure has immediate human consequences. Christina Maslach, whose burnout research began with 1970s interviews of hospital and emergency-room staff, gave the 2018 DevOps Enterprise Summit keynote on job burnout. (IT Revolution, “Understanding Job Burnout”.) Westrum arrived through patient safety. Maslach arrived through human services.
How the model reached software
The Phoenix Project appeared in 2013. The first DevOps Enterprise Summit followed in San Francisco in 2014. That same year, the State of DevOps Report became the first report in the Puppet and DORA lineage to measure culture with Westrum's typology. The 2013 report contained no Westrum reference. The 2014 edition reproduced his table and cited the 2004 paper. (Puppet Labs, 2014 State of DevOps Report.)
By measuring culture with the typology and connecting it to how technology organizations perform, the report's authors carried Westrum from the safety literature into engineering leadership. DORA still uses a six-item version of the model.
Nicole Forsgren Velasquez, Gene Kim, Nigel Kersten, and Jez Humble wrote that report. Four years later, Forsgren, Humble, and Kim published Accelerate. Its third chapter returned to Westrum. In an IT Revolution adaptation, the authors wrote that their hypothesis about culture predicting software-delivery and organizational performance “proved to be true.” (IT Revolution, “Westrum's Organizational Model in Technology Organizations”.)
That lineage is how the model reached me: The Phoenix Project, my first DevOps Enterprise Summit, the free books that came with the conference, The DevOps Handbook among them, and Accelerate, which I read before it hit the mainstream. I still have a pre-release copy.
When I finally went back to Westrum's paper, the DevOps sources had already done the interpretive work, and the original read like a summary of things I knew. I skimmed it and moved on.
What the evidence supports
Westrum stated the limits himself. Case studies supporting the scheme were plentiful, while systematic tests were harder to find. Some studies found a relationship between culture and medical outcomes; others failed to find one. He called the correlation a hypothesis and treated climate as one statistical contributor to performance, with exceptions. Some effective organizations, he observed, fell outside the generative type.
He also wrote that the scheme captures only part of organizational culture. It stays silent on training, structure, and styles of problem solving. Information can help an organization achieve its goals regardless of what those goals are. A generative pattern carries no automatic claim that the mission itself is good, or that its costs fit every setting.
DORA's current culture page uses more categorical language. It says a high-trust, generative culture “predicts” software-delivery and organizational performance. No independent peer-reviewed test of DORA's six-item implementation has been published, as far as I can find. A statistical relationship remains the strongest claim those results can carry. (DORA, “Generative organizational culture”.)
I have used the model for years because it has earned its keep as a lens. A small table can explain only part of an organization and still be useful. Westrum left room for mixed cultures, local conditions, and exceptions. Leaders can test the pattern against what people actually experience.
Westrum himself put it in three words: “Culture is mutable.”
Years of looking through it
Four and a half years of consulting put me inside more than 20 client organizations across more than a dozen industries. I have seen cultures in transition and watched the surrounding culture shape a project's prospects. It also shapes how employees treat outside consultants and contractors, and how much candor survives across that boundary.
Outside consultants often feel the local rules before anyone explains them. The same practice can land well in one organization and stall in another for reasons that have nothing to do with the practice.
The typology became an influence map. It helped me judge where to push and where to hang back. Sometimes the available radius was an organization. Sometimes it was a department, a director's area, a set of teams, or one team. Some conditions made influence hard or impossible.
A senior vice president once answered a request I had made to readers about subjects worth exploring more deeply. He chose Westrum and generative culture. I think he was trying to build a more generative culture in a challenging environment.
DORA stays close to engineering. Team Topologies addresses organizational structure, and value-stream work traces flow. I see Westrum underneath all three: whatever a leader adopts from them lands inside an existing pattern of how people treat information, and that pattern decides what the adoption becomes. The typology deals with how people relate to one another and how work gets done, across organizations, domains, and contexts. AI changes some of the work; the questions about information, power, rules, and mission remain.
People still encounter harmful conditions and assume they are looking at isolated dysfunctions. The people inside still have to do the repair. Westrum can show why the same practices thrive in one place and stall in another, and help a leader judge the reach of their own influence. We need to keep finding ways to make work suck less.
Locate yourself
You now have a way to look at your organization and every other one you have worked in. The six rows make that view more specific. Each row asks where your experience sits across the three cultures, one behavior at a time. Each of the six rows gets a full essay of its own.
You may take the model as permission to change the culture, permission to accept where you are, or permission to have your experience confirmed. Confirmation can bring relief and frustration together. Sometimes the clearest response is simply, “I knew it.”