Two months after I took over a nine-person team, my best engineer quit. In her exit conversation she said something I still think about: "The work is fine. I just don't know what any of it is for." Nobody had yelled at her. Nobody had cut her pay. The culture was simply... accidental. It had formed the way a puddle forms — whatever shape the ground happened to be. That was the first time I understood that building a strong company culture isn't a project you run once. It's a set of choices you keep making when it would be easier not to.
Key Takeaways
- Culture is the default behavior an organization rewards when nobody is watching — it forms whether you design it or not.
- Measure it with three numbers you can track monthly: participation in feedback surveys, voluntary turnover, and internal promotion rate.
- Hiring for "culture fit" without a written definition of your values is how you accidentally build a monoculture.
- Weak culture survives calm quarters and collapses in a crisis, because crisis removes the rituals that held it together.
- You cannot fix culture with perks. You fix it by changing what gets rewarded, promoted, and quietly tolerated.
How to build a strong company culture that survives contact with reality
Most culture advice starts with a values workshop. I'd argue that's the third step, not the first. Values posters don't create behavior — behavior creates values, and then someone writes the poster afterward to describe what already works.
Here's the sequence I've landed on after watching it go wrong a few times.
Step one: write down what you already reward
Before you invent anything, do an audit. Take the last six promotions on your team and ask one question about each person: what did they actually get rewarded for? Not what the job description said. What did the manager praise in the last review, what project got them visibility, what behavior got someone interrupted in a meeting or left alone?
You'll usually find a gap between the stated culture and the operating one. At one company I worked with, the official value was "candor." The real value was "never make the founder uncomfortable in a group setting." Guess which one new hires learned within three weeks? The gap is the actual culture. Writing it down is uncomfortable precisely because it's honest.
Step two: define three behaviors, not seven words
Seven values means zero values, because nobody can hold seven priorities in their head during a Tuesday afternoon decision. Pick three, and make each one a behavior rather than a noun. "Integrity" is a noun nobody can act on. "We say the hard thing in the room, not in the hallway afterward" is something a person can actually do or fail to do.
- One value about how decisions get made — who decides, who gets consulted, and how disagreement is handled.
- One about how people treat each other when something breaks, because that's when culture is genuinely tested.
- One about what the company refuses to do, even when it would be profitable. This one is the hardest to write and the most useful.
If your three behaviors could apply to any company on earth, they're too generic. Rewrite them until they'd sound slightly odd at a competitor.
How to build culture on a team when you're not the CEO
You don't need permission to build a team culture. You need consistency at a level small enough that you control it. A team of six has its own culture regardless of what the company-wide deck says, and that local culture is often what people actually experience day to day.
The three rituals that actually do something
I've tried a lot of team rituals. Most die within a quarter. The ones that survive share one property: they either remove friction or surface information you'd otherwise miss.
- A ten-minute weekly written update from each person: what moved, what's stuck, what you need. Written, not a meeting. It costs almost nothing and gives quieter people a channel.
- A rotating "decision log" owner. One person per week writes down the two or three decisions the team made and why. Six months later, this document explains more about your culture than any handbook ever will.
- A blameless review after anything that broke. Not a postmortem template with 14 fields — three questions: what happened, what made it possible, what changes next time. Skip the part where someone performs regret.
Notice what's missing: no team-building offsite, no personality quiz. Those aren't harmful, but they're decoration. A ritual that doesn't change how work flows is just calendar clutter.
How to build a strong company culture: examples from two teams
I've seen the same principle produce opposite-looking cultures in two very different settings, which is the useful part — it means you're copying a mechanism, not a vibe.
| Dimension | Engineering team, ~40 people | Warehouse operation, ~120 people |
|---|---|---|
| Core value | "Disagree in writing, decide in person" | "Safety issues stop the line, no exceptions" |
| Key ritual | Written decision log, rotated weekly | Five-minute standup before each shift |
| What's measured | Time from proposal to decision | Near-miss reports filed per month |
| What's punished | Silent disagreement after a decision | Reporting a hazard and being told to wait |
| What it produces | Faster decisions, fewer grudges | People flagging problems early instead of hiding them |
The engineering team went from roughly 11 days to 4 days between proposal and decision over about five months. Not because anyone got smarter — because the decision log made stalling visible. The warehouse saw near-miss reports climb steadily, which sounds alarming until you realize the number of actual injuries fell over the same period. More reports meant people trusted that speaking up wouldn't get them in trouble.
What both cases have in common
Neither team started with a values exercise. Both started by identifying one behavior that was quietly being punished and then stopping the punishment. That's it. Culture work is mostly subtraction before it's addition.
5 ways to improve company culture without a budget
You can do all five of these next month with no new spend. That's not a gimmick — it's the point. Culture that depends on budget is culture that disappears the first bad quarter.
- Change one meeting so that the quietest person speaks before the loudest. Just reorder the agenda. Try it for four weeks and watch what gets said.
- Publish the criteria for the next promotion before the process opens. Ambiguity here is where most resentment is quietly manufactured.
- Answer every "why" question honestly, including when the answer is "because finance said so." Adults can handle constraints; they can't handle fog.
- Kill one recurring meeting that nobody can justify. Removing a ritual people dread buys more goodwill than adding one they enjoy.
- When someone leaves, ask the person who stayed what they'd change. They'll tell you things exit interviews never surface.
How to measure whether any of this is working
Culture feels unmeasurable until you pick specific numbers and stop pretending you can quantify "vibes." Three indicators, tracked monthly, will tell you more than an annual engagement survey ever will.
- Survey participation rate. If fewer than roughly two-thirds of people answer, the survey itself has become a culture problem — people don't believe it changes anything.
- Voluntary turnover, split by tenure. Losing people in their first year is a hiring and onboarding signal. Losing people at year three is a growth-path problem. Same number, completely different diagnosis.
- Internal promotion rate. If almost every senior role gets filled from outside, your team has already concluded that staying doesn't pay.
One caution: don't chase the survey score itself. I watched a team spend a full quarter optimizing the wording of questions to nudge a number upward. The number went up. Nothing else did. Track the metric, then act on the specific answers behind it, or you're just doing theater with a dashboard.
What about hybrid and multi-site teams?
Distributed teams don't need different culture principles — they need the principles written down, because they can't absorb them by osmosis. In an office, a new hire learns the unwritten rules by watching. Remotely, there's nothing to watch, so anything undocumented effectively doesn't exist. The upside: remote forces you to state your culture explicitly, and explicit culture travels better when you open a second location.
The part nobody warns you about
Every strong culture excludes someone. That's not a bug — it's the mechanism. If your culture is genuinely clear about how decisions get made and what behavior gets rewarded, some people will find it a poor fit and leave, and that's a correct outcome, not a failure. The danger is when "fit" becomes an excuse to hire people who resemble the existing group. Write the behaviors down first, and you can at least tell the difference between someone who doesn't match your values and someone who simply isn't like you.
Back to that engineer who quit. About a year later she emailed me to say the new place had the same problem in a different shape. It turns out clear purpose is rarer than good coffee, and people notice its absence long before they can name it. So here's the question worth sitting with: if your three most recent promotions were announced tomorrow with the real reasons attached, would anyone on your team be surprised? If the answer is yes, you already know where to start — and it isn't a values workshop.
