Section 1
Find out which stories are already circulating
Before you define anything, collect what exists. Ask people who joined in the last six months what they were told about how things work here, phrased as a question about advice rather than culture. What did someone tell you to watch out for. What did they say never to do. Who did they say to ask when something is stuck. You will get a short list, and some of it will contradict the values you have published. That contradiction is the actual state of the culture and it is the only useful starting point. The stories that circulate are almost always about consequences: what happened to the person who tested a rule. That is why culture cannot be defined by assertion. An assertion has no consequence attached, and people are asking about consequences. Founders carry more weight here than they expect, for the reasons set out in [Why Storytelling Is a Critical Leadership Skill for Founders](/blog/why-storytelling-is-a-critical-leadership-skill-for-founders).
Section 2
New canon is made by decisions, not by writing
You add to the set of circulating stories in one way: by making a costly decision visible. Refuse revenue that would have compromised a commitment and say publicly that you refused it and what it cost. Back someone who made a defensible call that turned out badly, and say why the call was defensible. A decision that costs nothing produces no story, because there is nothing to tell. This is the reason values workshops feel hollow to the people in them. Everyone in the room knows the list will be tested at some point by a real trade-off, and nobody in the room believes the list determines the outcome. It is cheaper to skip the workshop and spend the attention on making one hard decision legible.
Section 3
Auditing the gap
The table below places each stated value beside the story people actually retell about it and the most recent decision that either confirmed or contradicted it. Values with no story attached are aspirations. Label them as aspirations and the whole document gains credibility.
Section 4
Making the good stories durable
Stories decay as the people who witnessed them leave. Preserve the ones that carry a decision rule, in writing, with the specifics intact: what the situation was, what the options were, what was chosen, what it cost, what would be decided differently now. Sanitised versions do not survive retelling because they are not interesting enough to retell. Put them where they are encountered in the course of work rather than in an induction folder: attached to the process they explain, in the runbook, in the review template. And prune. A story that celebrates a heroic weekend rescue teaches people that heroics are rewarded, which produces more emergencies. Nonprofit communications teams face a sharper version of the same editorial problem, since their material involves real people with real exposure, and [How Nonprofits Use Storytelling for Fundraising](/blog/how-nonprofits-use-storytelling-for-fundraising) is worth reading on that discipline.
Section 5
Three ways this goes wrong
Manufacturing a culture story is the first and it never survives contact. Somebody was there, and their version circulates alongside yours with the correction attached. The second is confusing culture with atmosphere. Social events and shared jokes are pleasant and tell nobody how a disagreement between two teams gets resolved, which is the thing people actually need to know. The third is letting the founder be the protagonist of every story, which teaches the organisation that judgement lives in one place and should be routed there. If you want to know which stories genuinely circulate rather than which ones you hope do, the text your company already produces in tickets, reviews and survey comments is the honest record, and [Natural Language Processing Use Cases for Startups](/blog/natural-language-processing-use-cases-for-startups) covers reading it at scale.