Organizational Culture

Every company has rules. Working hours, how to request time off, spending limits — all written down somewhere. But very few of the decisions made on an ordinary day have an answer in that document. Whether to speak up when a manager's call looks wrong. Whether to leave a team's data open by default or hand it over only on request. Whether to cut quality or push the date when a deadline gets tight. No policy covers any of these. And yet inside one company the answers tend to land in the same place across dozens of people. Something wider than the rulebook is already running.

Before that something had a name, companies called it the atmosphere. The word carries a quiet assumption: that it can't be touched. It arrives like weather, and you're lucky or you put up with it. So when two companies in the same industry building similar products had attrition rates three times apart, there was no language for the gap. One side said they'd been lucky with hires. The other side said people these days don't stick around. Neither explanation suggested anything to do on Monday morning, which is the practical cost of calling something weather.

At Pagedeck the problem surfaced during an incident review. Payments had been down for 40 minutes, and while tracing the cause Theo mentioned that he'd noticed something odd before the deploy and hadn't said anything. Grace asked why. Theo said he wasn't certain, and slipping the date looked like the bigger problem. Nobody in that room had ever told Theo to keep quiet. No such rule existed anywhere. But Theo chose silence, and when the team looked into it afterward, several other engineers said they would have made the same call.

Organizational culture names this phenomenon so it can be worked on. The move that makes it useful is a reversal of where the explanation sits. Instead of treating people's personalities or their generation as the cause, it looks at what the company's past decisions have taught everyone. That is also why culture occupies one of the nine branches the people field uses to organize everything from hiring to departure. Hiring decides who comes in and compensation decides what they're paid; culture decides what they do when nobody is watching.

The terms under organizational culture fall into four branches: what the standard is (composition), what conditions let that standard hold (environment), how to tell where things stand (measurement), and what it looks like when things go wrong (risk). The first two define what you're trying to build. The last two tell you what you actually have.

Composition is the branch that writes the standard down, and it holds one term. Core values are the beliefs a company commits to in advance so that split decisions have a tiebreaker. One test separates values that get used from values that decorate a wall: does following the sentence cost you anything? "Put the customer first" gives up nothing, so it never changes a decision. "We don't hand over customer data, even when revenue depends on it" costs a signed contract. Only companies that have paid that price once reach for the sentence again. A useful set is also short. Eight values mean that any decision can be justified by at least one of them, which returns the company to having none.

Environment is the branch covering the conditions under which the standard can actually hold, and it has four terms. Each removes a different obstacle.

1. Psychological safety — the sense that admitting a mistake or disagreeing carries no penalty. It comes first because without it the other three are decorative. Open all the dashboards you like and nobody asks a question they're afraid will make them look uninformed. Grant all the autonomy you like and people still come for approval, because being wrong alone is worse than being slow.
2. Transparency — the practice of sharing information rather than holding it. It puts the raw material for judgment in everyone's hands. Someone who knows the quarter is at 60 percent of target, and why a feature was cut, cannot reach the same conclusion in a meeting as someone who doesn't. It works better when the exceptions are named too. Individual salaries and unannounced personnel moves usually generate more speculation open than closed.
3. Autonomy — how much of the how a person gets to decide. It isn't a license to pick your own work. The goal is set together and the method is handed over. Cutting churn by two points this month comes from above; whether that means fixing notifications or reworking onboarding belongs to the team.
4. Diversity, equity, and inclusion (DEI) — the effort to let people of different backgrounds work on fair terms. The three words do different jobs. Diversity asks how mixed the group is, equity asks whether opportunity and pay follow the same rule for everyone, inclusion asks whether the people who arrived have room to be heard. Handle the first two and skip the third and you get a company that hired well and then watched exactly those hires leave first.

Measurement is the branch that puts a number on the current state, and it has two terms with opposite shapes, which is why they're usually run together.

1. Engagement survey — a multi-item questionnaire about satisfaction and commitment. Because it splits the question into separate items, such as whether disagreeing with a manager feels safe or whether the next step in a career is visible, it shows where the weakness sits. The cost is length: past thirty items response rates fall, and running it quarterly wears people out.
2. Employee Net Promoter Score (eNPS) — a score built from one question, whether you'd recommend this company to someone you know. One item is cheap enough to ask often and good for watching a trend, but it never explains a drop. So teams use eNPS to catch when something turned and the survey to find what turned.

Risk is the branch describing what a damaged culture looks like in practice, and it has two terms. Both show up before anyone resigns.

1. Burnout — the state of running empty after sustained overload. Volume alone doesn't cause it. It arrives faster when the work has no visible end, or when nobody can say what the last three months were for. That's why it appears on teams that rarely work late. The usual signs come in order: the work still ships, then the enthusiasm goes, then the quality slips, and by the time output drops the person has been running empty for months.
2. Quiet quitting — staying, but doing the assigned share and nothing past it. Nothing looks wrong from outside, which is why it gets caught late. Attendance holds, deliverables land on time, and what disappears is opinions in meetings and any hand offered to someone else's problem. It is often less a decision than a conclusion someone reached months earlier and never said out loud.

The four branches form a chain. Core values decide what gets chosen, the four environment terms make that choice possible, the two measures report where things stand, and the two risks raise a flag while there's still time. Break one link and the rest stall. Without values the environment has nothing to serve. Without measurement a decline reaches the founder last. Without reading the risk signals, the first confirmation arrives as an attrition number, months after the point where anything could have been done about it.

"Candor is one of our values. Let's look at whether anyone got burned for being candid last quarter."

A proposal to check culture against real decisions rather than the stated version.

"The growth item dropped twelve points on this survey. Everything else held."

The problem sits in one area, not across the company, so start there.

"I'm worried 'culture fit' has turned into hiring people who are like us."

A flag that defending culture is sliding into sameness.

Pagedeck first treated culture as something to work on at 45 people. The trigger wasn't a decision to become a good place to work — it was a number. eNPS fell from 31 to 4 in one half, while only two people left, so nothing in the attrition data explained it. The first engagement survey came back with one item far below the rest: whether it's easy to speak up when you spot a problem.

Grace gathered the four managers and asked each to write down who had raised an objection over the past two quarters and what happened next. Three cases came back, and all three looked alike. The person who raised the issue was told it was a good catch, and the following week the work of chasing that issue landed on them. None of it was punishment. But after the third repetition, the engineering team had learned from experience that naming a problem adds work to your own plate.

What Mia proposed was not a rewrite of the culture statement. It was a change to how meetings ran: whoever finds a problem does not get assigned to fix it. The next survey brought that item back up. The sentences on the wall hadn't changed. One meeting rule had. That is the shape most culture work takes: not a new statement, but one repeated decision reversed.

The most common mistake is treating culture and perks as the same thing. Good snacks and a work-from-anywhere policy get a company called a good culture. There's a reason it reads that way — perks are visible, they fit in a job posting, and with budget they start next month, while culture is invisible and unpostable. But the two answer different questions. Perks answer what the company gives you. Culture answers what people do when the situation gets uncomfortable. Run them as one thing and you spend the budget without moving anything. A company that expanded the snack shelf and saw no change in attrition is usually a company where snacks were never why people left.

Treating culture as something that just happens, beyond anyone's reach, fails just as often. People gather, an atmosphere forms, and it isn't anyone's job to manage. Given that culture accumulates through experience rather than rules, the conclusion feels reasonable. But those experiences were produced by decisions the company made. Handing extra work to whoever raises an objection is a decision. Passing over someone who reported a mistake is a decision. Both can be made differently. Left alone, culture only hardens in the direction it's already drifting: the habits of whoever has been there longest become the standard, people who don't match those habits leave quietly, and the remaining habits grow stronger for it. Three years in, a culture nobody chose is running the company.