Westrum Organizational Culture: Pathological, Bureaucratic, Generative

Oleh Victor Hoang · Diperbarui 11 Oktober 2026

Westrum organizational culture is a typology by sociologist Ron Westrum that sorts organizations into three types by how information flows: pathological (power-oriented), bureaucratic (rule-oriented), and generative (performance-oriented). The type shows up in how people treat bad news, share responsibility, handle failure, and react to new ideas.

Most culture models ask what an organization believes. Westrum asks something more practical: when someone knows something the organization needs to hear, does it travel? That one question predicts a surprising amount. It tells you whether a small problem gets flagged on Monday or discovered by a customer on Friday, and whether a failed launch becomes a lesson or a witch hunt.

Where the Westrum Model Came From

Westrum published the typology in 2004 in the paper "A typology of organisational cultures" (Quality & Safety in Health Care, DOI 10.1136/qshc.2003.009522). He wrote it for safety-critical settings such as healthcare. His argument: the way information flows both shapes and reflects other parts of a culture, so it can help predict how an organization will respond when problems emerge. He also noted that information culture is associated with error reporting and with performance, including safety, while cautioning that the link needed further study.

The model moved into business when the DORA research program (DevOps Research and Assessment, now part of Google Cloud) adopted it to measure software teams. DORA's page on generative organizational culture describes good information as information that answers the receiver's questions, arrives on time, and can be used effectively. That definition is what makes the model portable. A sales team, a finance team, and a hospital ward all depend on information reaching the right person in time to use it.

The Three Culture Types

Pathological culture is power-oriented. Information is a resource to hoard or a threat to suppress. People protect their turf, and the safest move is to say nothing.

Bureaucratic culture is rule-oriented. Information moves through channels, and the channels decide what gets heard. It isn't hostile, but it's slow and narrow. People do their own job well and stop at the edge of it.

Generative culture is performance-oriented. The mission sets the priorities, so information flows to wherever it helps most. People expect to be told the truth, and they expect to tell it.

These aren't personality types for companies. A single organization can contain all three: a generative engineering team inside a bureaucratic finance function inside a company with a pathological executive layer. That's why you measure at team level, not just company level.

The Six Behaviours That Separate Them

DORA's page lays out six behaviours that differ across the three types. This table follows that structure.

Read the table as six small tests. You can run each one on your own team this week.

Cooperation

Look at who actually works together. In a pathological culture, teams guard their budgets and headcount. In a generative one, a delivery process includes every function that touches it, from the start.

Messengers

Think about the last person who brought you bad news. What happened to them? "Punished" doesn't have to mean fired. It can be an eye roll, a missing invitation, or a slow loss of influence. "Trained" means the organization teaches people how to raise problems, and makes doing so safe at any time.

Responsibilities

Bureaucracies draw tight boxes around each role, and anything outside the box belongs to nobody. Generative teams treat quality, reliability, and risk as shared concerns, so a gap between two roles gets filled instead of admired.

Bridging

Bridging is about the links between groups. Silos are tolerated in rule-bound cultures and actively policed in pathological ones. Generative cultures build bridges on purpose: shared channels, early involvement of downstream teams, and informal relationships.

Failure

This is the sharpest single test. After something breaks, does the first question concern who did it, or what allowed it? The first produces scapegoating or a formal blame process. The second produces inquiry. For more on this contrast, see blame culture vs learning culture.

Novelty

New ideas are crushed, routed into a problem, or implemented. In practice, ask how many suggestions from frontline staff changed something in the last quarter.

What the Research Actually Shows

DORA's research links generative culture to performance, with a few caveats worth knowing.

The 2018 DORA report page states that generative culture is essential, and describes it as a culture of cooperation, trust, voice, and a focus on learning that leads to better software delivery and operational performance and better organizational outcomes. The 2022 report page lands on a similar finding: high-trust, low-blame cultures lead to better organizational performance.

DORA's capability page adds that high-trust, generative culture predicts software delivery and organizational performance. It cites a multi-year Google study and the 2019 State of DevOps Report, which tied psychological safety to delivery performance, organizational performance, and productivity.

Three limits matter when you quote this. First, these are software-delivery findings. Applying them to a sales floor or a hospital is a reasonable extension of the model, but it isn't what the surveys measured. Second, the pages above report the direction of the finding, not a single headline percentage, so treat any "generative culture makes you X% better" claim with suspicion unless it names its source. Third, DORA's newest edition has shifted focus. The 2025 report centers on AI, with the finding that AI acts as an amplifier of an organization's existing strengths and weaknesses. If that's right, a culture that suppresses bad news will amplify that too. For the wider evidence base, read the link between culture and performance.

Key Facts

• Ron Westrum published the typology in 2004 in "A typology of organisational cultures," in Quality & Safety in Health Care. Source: PubMed Central
• DORA defines the three types as pathological (power-oriented), bureaucratic (rule-oriented), and generative (performance-oriented). Source: DORA
• DORA measures the culture with six survey statements on a 1 to 7 agreement scale, averaged into one score. Source: DORA
• DORA's 2018 report page states that generative culture is essential for better software delivery and operational performance. Source: DORA 2018
• DORA's 2022 report page reports that high-trust, low-blame cultures lead to better organizational performance. Source: DORA 2022

Westrum vs Other Culture Frameworks

Westrum isn't a rival to the better-known culture models. It answers a narrower question.

The practical difference: Competing Values says no type is best. Westrum disagrees on its own turf. A pathological culture isn't a style choice. It's a risk.

How to Measure and Shift Your Culture, Step by Step

Step 1: Pick the unit

Choose a team, not the whole company. Eight to twenty people is a workable size. Whole-company averages hide the variation that matters.

Step 2: Run the six-item survey

Send the six statements (see the template below) as an anonymous survey, unlabeled and untitled, in the order shown. DORA recommends only minor wording changes so the instrument keeps its statistical properties. Keep it anonymous, or scores will drift upward for exactly the reason the culture is worth measuring.

Step 3: Score it

Average the six answers per person, then average across the team. Also look at the lowest-scoring item, because the average can hide a single broken behavior. For broader measurement approaches, see how to measure company culture.

Step 4: Cross-check with behavior

Surveys report perception. Compare them with records: how many incidents were reported by staff rather than customers, how long bad news took to reach a decision-maker, how many ideas from the team shipped. Disagreement between the two is itself information.

Step 5: Pick one behavior to shift

Don't try to change six things. Choose the weakest one and work on it for a quarter. If failure is the weak spot, start with blameless reviews. If messengers are the weak spot, start by visibly thanking the next person who brings you bad news.

Step 6: Change what leaders do first

Culture moves when leaders' reactions change. If the first reaction to bad news is a question about blame, no policy will fix it. For why people hold back in the first place, see why teams stay silent in meetings.

Step 7: Re-run in a quarter

Same unit, same six statements, same timing. Look for movement in the item you targeted, not just the average.

Template: Westrum Survey and Scoring Sheet

DORA publishes these six statements, scored from 1 (Strongly Disagree) through 4 (Neither Agree nor Disagree) to 7 (Strongly Agree). They're taken from DORA's generative culture page. Copy the table into a spreadsheet and give each respondent a row.

Scoring note: average the six items for the team's single Westrum score. A higher average indicates a more generative culture. The sheet is a diagnostic, not a verdict, so pair it with the behavioral checks in Step 4.

Examples by Business Function

These are illustrative scenarios, not case studies.

Do's and Don'ts

Do:

• Measure at team level, and report the item scores, not just the average.
• Reward the person who surfaces a problem, publicly and quickly.
• Run reviews after failures that ask what allowed it to happen.
• Build links between teams before a crisis forces them.

Don't:

• Don't use the survey as a ranking of managers. Used that way, it teaches people to answer defensively.
• Don't assume a bureaucratic culture is the enemy. Rules have value; the issue is a narrow flow of information.
• Don't announce "we're a generative culture." The behaviors define it, not the label.
• Don't stop at the average. A high mean can hide one broken item.

Where to Go Next

Westrum gives you the diagnosis. The next question is why information gets blocked in the first place. Start with psychological safety at work for the individual side of speaking up, and blame culture vs learning culture for what to do after something goes wrong.

Westrum's point is simple and uncomfortable: a culture is only as good as the worst news it can hear without flinching. Measure how your team handles that, and the rest of the culture conversation gets a lot more concrete.