Workplace culture survey: questions and interpretation

A satisfaction survey asks whether people like their job. An engagement survey asks whether people are willing to put in effort. A culture survey asks something different: whether the stated values of the organization match how people actually behave day to day, and whether people feel safe enough to say so.

That makes culture surveys more sensitive than the other two, in both senses of the word — the questions probe more personal territory, like psychological safety, fairness, and whether people feel they belong, and the answers are more sensitive to how safe respondents feel giving an honest answer in the first place.

• “The way decisions actually get made here matches what leadership says we value.” (agree/disagree)
• “I've seen someone held accountable for behavior that went against our stated values, even when that person was a strong performer.”
• “New hires are told the truth about what it's really like to work here during onboarding.”

• “I can disagree with my manager without fear of negative consequences.”
• “I feel like I belong on my team.”
• “Mistakes here are treated as something to learn from, not just something to blame someone for.”
• “I can bring up a problem with how our team works without it being held against me.”

• “It's clear who is responsible for making a given decision.”
• “People here are held to the same standard regardless of tenure or relationships.”
• “When something goes wrong, this organization looks at what happened, not just who did it.”

Frame culture questions around specific, observable behavior instead of abstract values language. “Do we live our values” is hard for anyone to answer meaningfully; “I've seen someone held accountable for behavior that went against our stated values” gives people something concrete to evaluate.

Pair every values-alignment question with room for an example in an optional open-text follow-up. The specific stories in the open text are usually more useful for identifying what to fix than the aggregate score is.

Culture questions ask people to describe things leadership might not want to hear, which means the honesty of the answers depends heavily on whether respondents trust the confidentiality of the process. On a team small enough that a comment could be traced to one person, expect answers to skew safer than the real sentiment — that's a limitation of the method, not evidence that everything is fine.

Report culture survey results at a level where no individual can be identified by elimination. Team-level results typically need a minimum group size — many organizations use five or more respondents — before they're broken out separately; smaller groups should be rolled into a department or region view instead.

A single organization-wide culture score is the least useful number a culture survey produces. The more useful comparisons are: this team against the org average, tenure under one year against tenure over three years, and this quarter against the same questions last cycle. Large gaps between subgroups are usually more actionable than the overall average.

Be cautious about comparing culture scores across teams with very different jobs — a warehouse team and a sales team can have equally healthy cultures and still answer psychological-safety questions differently because of the nature of the work, not the quality of the culture.

Start with the individual questions that show the widest gap between groups, not the organization-wide average. A values-alignment question that scores well below the org average for people under one year of tenure points at something specific — onboarding, a particular manager, or a promotion decision people watched happen — in a way the overall number never will.

Pull a small group of respondents into a listening session built around that specific question, not the survey as a whole, and re-ask the identical question next cycle before calling the problem fixed. One cycle of improvement can be noise; the same question moving the same direction across two or three cycles is a signal worth acting on.