Your culture isn't your values. It's your operations.

Infinia ยท

Here's the uncomfortable part, before anything else:

If your company's culture is drifting, the founder is usually the last person who should be fixing it.

Not because founders don't care. Because the culture is the founder.

It's their habits, their hours, their blind spots and their way of making decisions, copied by everyone who joined early. Asking the founder to fix the culture is asking them to see themselves clearly while running the company. Almost nobody can.

Culture is what you reward, not what you write

Every company at $1M to $10M has a values page. Most of them say some version of "ownership, transparency, customer obsession."

None of that is culture. Culture is the answers to questions like these:

Those are operating decisions. Hiring, firing, compensation, decision rights, meeting cadence, how incidents are handled.

Culture is just operations, repeated until people stop noticing.

Which is why it's an operator's job.

The best-known culture document was co-written by an operator

The most famous culture document in tech is Netflix's "Freedom & Responsibility" deck. It said things most companies wouldn't say out loud.

"We're a team, not a family."

"Adequate performance gets a generous severance package."

It wasn't a values poster. It was a set of operating rules: how Netflix pays, how it decides, how it handles average performance. And it was written by Reed Hastings with Patty McCord, Netflix's chief talent officer, an operator whose job was turning those ideas into how the company actually ran.

That's the model. The founder sets the direction. An operator turns it into rules people live by.

Culture decisions have a price. Somebody has to run them.

Two well-known examples show what it looks like when culture gets treated as an operating decision, not a mood.

Coinbase, 2020. Brian Armstrong published a post declaring Coinbase a "mission focused company" that would stay out of broader social and political causes at work. Then he offered severance to anyone who didn't want to work that way. Around 60 employees, roughly 5% of the company, took it.

Basecamp, 2021. The founders announced they would no longer allow societal and political discussions on the company's internal tools, along with other policy changes. They offered buyouts. About a third of the company left.

You can think either decision was right or wrong. People still argue about both. But notice what they have in common: a culture decision turned immediately into an operations problem. Severance terms, who covers whose work, hiring to replace a third of a team, keeping customers served while it happens.

Most $3M companies make culture decisions by accident and never run the operations behind them. That's how culture drifts.

What happens when nobody owns it

The opposite example is just as well known.

At Uber in 2017, a former engineer, Susan Fowler, published an account of how harassment complaints were handled internally. An outside investigation followed, with a long list of recommendations about how the company managed people, reviews and complaints.

Within months, the CEO had resigned.

Uber was a huge company. But the failure pattern shows up at every size:

the culture was whatever the most powerful people tolerated, because no one owned the systems that should have caught it.

HR reported into the same power structure it was supposed to check.

A $5M company won't have Uber's headlines. It will have the same pattern on a smaller scale: the high performer everyone's afraid to manage, the founder's friend nobody can give feedback to, the person who quietly does the work of three and burns out.

The AI-native version: culture by exhaustion

AI-native companies are writing a new chapter of this right now.

Some of the best-funded AI startups are open about extreme hours. One CEO wrote that his team routinely works "through the weekend," and that "many of us literally live where we work."

Job posts ask candidates to "thrive under high urgency, with AGI timelines approaching."

At the other end, Henry Shi, who tracks lean AI companies, describes that grind as "exhausting and unnecessary." Plenty of founders running $5M companies with ten people would agree.

Here's our take: neither one is a culture until someone operationalizes it. "We work hard" without rules about on-call, time off, decision rights and who covers whom isn't a high-performance culture. It's a burnout schedule with a mission statement.

"We're lean" without clear ownership is how a ten-person team ends up with everyone saying "I'm dying, we need one more person."

Why a fractional COO

The EOS world has a simple way to describe this: every company needs a visionary and an integrator.

The visionary sets direction. The integrator turns it into how the company runs. Most founders are visionaries. Most $1M to $10M companies have no integrator.

That's the COO's job, and it's why a fractional COO is often the right fit for culture work:

  1. They're outside the founder's shadow. They can see the habits everyone copied without being one of the people who copied them.
  2. They own the systems, not the slogans. Hiring process, performance reviews, decision rights, meeting cadence, how conflicts escalate.
  3. They can say what nobody inside will. "Your best engineer is the reason two people quit." "Your leadership meeting is a status update with no decisions."
  4. It's a project with an end. Culture work at this stage is mostly design and installation: set the systems up, train the team to run them, then step back. That's a few days a month for a few months, not a full-time hire.

Here's the part founders won't like: a fractional COO doing culture work will, at some point, tell you that you're the problem. If you hire one who never does, you hired the wrong one.

Where the CFO and CTO come in

Culture lives in money and in code as much as in meetings.

The fractional CFO makes culture show up in pay. If your values say "ownership" but your compensation rewards tenure, people will believe the pay. The CFO builds comp bands, equity refreshes and bonus structures that reward the behavior you actually want, and models what it costs before you promise it.

The fractional CTO makes culture show up in engineering. Engineering culture is code review norms, on-call rotation, how deploys happen, and what happens after something breaks. Blameless incident reviews versus finding someone to blame shapes how honest your engineers are more than any values page.

The COO owns the whole system. The CFO and CTO make sure the money and the code reinforce it, instead of quietly contradicting it.

When you don't need this

If you're in one of those groups, we'll tell you so.

Three questions to answer this week

  1. Who was the last person promoted, and could everyone on the team explain why?
  2. When two leaders disagree, who decides, and is that written down anywhere?
  3. What behavior do you tolerate from a high performer that you'd fire anyone else for?

If the honest answers are "not really," "me, eventually," and "more than I'd like," your culture isn't a values problem. It's an operations problem, and nobody owns it yet.

If your culture is drifting and you're the one who's been trying to fix it, we'll introduce you to one fractional COO who has done this before. And, if the money or the engineering side needs it too, the CFO or CTO to support them. Free: infinia.fun/for-businesses/intake

Tell me where I'm wrong. Founders who built great cultures themselves, I expect to hear from you.

Sources: Netflix, "Freedom & Responsibility" culture deck (2009) by Reed Hastings and Patty McCord; Brian Armstrong, "Coinbase is a mission focused company" (September 2020) and follow-up reporting on roughly 60 departures; reporting on Basecamp's April 2021 policy changes and buyouts; Susan Fowler, "Reflecting on one very, very strange year at Uber" (February 2017) and the Holder report; Gergely Orosz, The Pragmatic Engineer, "New trend: extreme hours at AI startups" (quoting Cognition's CEO and a Lovable job post); Henry Shi, Lean AI playbook; the visionary/integrator model from Gino Wickman's EOS and the book "Rocket Fuel."