Organizational culture: what it is, how to observe it and turn diagnosis into action

1. What organizational culture means

Organizational culture can be understood as assumptions, values, norms and meanings that guide how situations are interpreted and how people act. Part of culture is visible—rituals, stories, language, symbols and processes—and part is discovered by observing which behaviours are rewarded, which problems can be named and how decisions are made when two values compete.

A value statement therefore does not demonstrate culture on its own. If a company declares collaboration but its targets reward individual results at any cost, practice communicates a different rule. Assessing culture means examining consistency between what is said, what is decided and what people learn will work.

• What is stated: purpose, values, policy and messages.
• What is practised: decisions, routines, incentives and coordination.
• What is tolerated: repeated behaviour that contradicts formal rules.
• What is learned: responses people consider safe or effective.

Culture does not live only in Human Resources: it is reproduced through operating decisions, leadership, systems and incentives.

2. Culture and organizational climate: related, not identical

Organizational climate focuses on shared perceptions of experienced policy, practice and behaviour: clarity, support, fairness, participation, respect, workload, safety or collaboration, among other dimensions. It is closer to asking, “What is it like to work here now?” Culture seeks deeper meanings and assumptions: “What is considered important, and how did we learn to operate this way?”

The distinction prevents excessive claims. A climate survey may show that one department perceives little autonomy, but it does not establish the cause by itself. Listening, work-design analysis and approval criteria may reveal whether the pattern relates to actual risk, micromanagement, ambiguous roles or missing resources. Available systematic review evidence also shows wide variation across instruments and dimensions; there is no universal scale that works equally well for every organization and decision.

• Culture: learned meanings, norms, assumptions and consistency.
• Climate: shared perception of current practices and conditions.
• Satisfaction: an individual evaluation of work experience.
• Wellbeing: a broader construct that includes work, health and life outside work.

3. Which dimensions to assess

Dimensions should follow the objective and context rather than the number of questions a platform allows. NIOSH organizes its worker-wellbeing questionnaire around work evaluation and experience, workplace policy and culture, physical environment, health and circumstances outside work. Its Quality of Worklife Questionnaire includes safety climate, psychosocial safety climate, respect, trust and management relationship alongside workload, autonomy, resources, roles and teamwork.

The ILO lists poor communication, unclear objectives, limited participation, insufficient support, authoritarian supervision, role ambiguity and job insecurity among psychosocial hazards. COPSOQ III provides an international framework covering demands, work organization, relationships, leadership, social capital and outcomes. These frameworks can guide dimension selection; they do not authorize copying items without reviewing licensing, language adaptation, population and measurement properties.

• Direction: purpose, priorities and consistency between messages and decisions.
• Work: workload, autonomy, role clarity, tools and execution capacity.
• Relationships: respect, trust, support, inclusion and conflict handling.
• Leadership: expectations, participation, fairness, recognition and response to concerns.
• Coordination: information, dependencies, decisions and cross-team collaboration.
• Safety and wellbeing: physical and psychosocial conditions the organization can influence.

4. Employee climate survey questions: examples and limits

A useful question expresses one idea, uses language the population understands and specifies a period when needed. Avoid double-barrelled statements such as “My manager communicates and recognizes people well,” because someone may experience strong communication and weak recognition. Also avoid wording that suggests the desired answer. A small comprehension test before launch can reveal ambiguity without collecting sensitive answers.

The examples below help organize dimensions but do not constitute a validated scale by themselves. Responses may use a consistent frequency or agreement scale, with “not applicable” where appropriate. Once a validated instrument is modified, its original measurement properties should not be assumed to remain intact automatically.

• Clarity: “I know which outcomes are expected from my role.”
• Resources: “I have the information and tools needed to do my work.”
• Voice: “I can raise a work problem without fear of retaliation.”
• Coordination: “The departments my work depends on make clear agreements.”
• Leadership: “My manager explains the criteria behind decisions that affect the team.”
• Respect: “My team corrects behaviour without humiliating people.”
• Open question: “Which concrete change would make your work easier?”

Asking creates an expectation of listening. If the organization cannot respond, reduce scope and explain what can realistically change.

5. How to gather evidence without breaking trust

Anonymous and confidential do not mean the same thing. An anonymous survey does not collect an identity that can link an answer to a person. A confidential process may hold identifiable information under restricted access and use. The organization should explain which model applies, which fields are collected, who can access them, how long data will be retained and under which rules results will be shown.

Cuts by department, tenure or role can identify someone even after names are removed. Reporting therefore needs minimum group sizes, segment combination and special review of open comments. Interviews and group sessions need similar rules for purpose, consent, notes and attribution. Trust comes from coherent technical and operating limits, not a vague promise of anonymity.

Survey results should be compared with relevant sources such as interviews, groups, process observation, policy and aggregated turnover, absence, approval time, incident, complaint or rework data. Those indicators neither replace employee voice nor prove causation; they help explanations to be formed and checked.

• Communicate purpose, use and limits before collecting answers.
• Collect only the information required for the defined decision.
• Avoid filters that allow identities to be inferred in small groups.
• Keep illustrative comments separate from representative conclusions.
• Document who accesses raw data and who receives grouped results.

6. From results to an action plan

Analysis begins by reviewing participation, missing data, distributions and relevant differences—not by finding the lowest score and blaming a department. An average can hide opposing experiences, and a small difference may have little practical meaning. Separate facts—what the population answered—from interpretations—what may explain the pattern—and decisions—what will be done. Sample and instrument limitations should remain visible.

Feedback should occur even when findings are uncomfortable. State what was heard, what cannot be concluded, which themes will be prioritized and which are out of scope for now. Then select a small number of actions with an owner, date, resources and review signal. Some responses will involve process or leadership changes; others may require training, formal investigation, role redesign or specialist care.

Later measurement is worthwhile only when it reviews an intervention or hypothesis. Repeating the same survey without action wears down participation. A broad baseline can be followed by short pulses on specific priorities and operational evidence, preserving enough stability for comparison without pretending to have precision the data cannot support.

• Priority: a relevant pattern connected with a feasible decision.
• Action: a concrete change to policy, practice, process or capability.
• Owner: a person with authority and resources to implement it.
• Signal: evidence for reviewing implementation and effect.
• Communication: an honest update on progress, limits and next steps.

Diagnosis does not change culture. Change comes from what the organization decides, executes, corrects and sustains after seeing the results.