The most expensive thing a growing company can get wrong is what it rewards, and the tell is not turnover. It is who applies.
A growing software company competes for senior people against businesses with more money and more certainty. Its one durable advantage is that the work is more interesting and the people are more trusted.
Without a clear purpose, it pays for the gap in hiring cost, in turnover, and in a slow slide in who it can attract.
The people you want are curious, comfortable with ambiguity and unafraid to think past their remit. Hiring alone does not produce them. They are what you reward.
The threat to new ideas is not a shortage of them. It is a culture that disempowers the people who have them.
Whatever you think of Netflix's culture, its premise is worth taking: people walk in the door with power, and the company's job is to remind them of it, then make room for them to use it.
So the question is not what your values are. Every company has a values page. The question is what your company rewards. Look at who was promoted last year, and why. Look at what gets applause at the all-hands. Look at what happened to the last person who ran a good experiment that failed.
Those answers are your values. If they disagree with the ones on the wall, the ones on the wall are decoration.
Alignment is what makes autonomy possible. Without it, autonomy is people moving in different directions at speed.
There is a distinction I keep coming back to. It is cruder than reality, and useful.
In a WE culture, anyone can own an initiative. A champion gets support from leadership, gathers the right people, runs the experiment and shares what was learned. The work moves in iterations: the scientific method and design thinking at their best.
A ME culture talks a lot about ownership. Answers arrive before problems are understood. Structure matters more than process. Underneath, reliably, is mistrust. People worry about roles, boundaries and stepping on toes. Will the process be fair? What happens if I am not included?
A ME culture is usually what is left when a company has used structure as its main tool for too long.
Your role is the hat you wear on a team.
A responsibility to the role, an accountability to the company. They are not the same, and the difference is where most friction lives.
The clearest case in software is product, marketing, sales and customer success. Product-led growth needs all four working as one system.
What happens instead, almost every planning cycle, is that each group tries to own its piece of a company metric, and to own it, optimizes for its own function. Everyone behaves rationally. The system gets worse.
Watch for it in your OKRs. When key results are activities instead of outcomes, or a KPI has quietly become the objective, you may be building a ME culture without meaning to.
How leaders frame things decides how much of that autonomy a team will take. Kahneman and Tversky showed that a loss hurts more than an equal gain pleases.
In a company the cost is direct. Under a message framed as loss, or with no information at all, people become risk-averse and look for certainty. They stop proposing things.
Software companies are more exposed to this than most, because the work is solving customer pain, and pain, problems and gaps are loss frames for what could as easily be called opportunities, hypotheses and experiments.
Which reading your team hears is mostly up to you.
Some tension is worth keeping. A CEO accountable for growth beside a CFO accountable for risk. Product, optimizing for learning and adoption, beside design, optimizing for less customer pain and higher quality.
Those frictions surface options neither side would find alone. The job is not to resolve them. It is to keep them civil and in the open.
What I would do on Monday is small. State the purpose of every meeting before it starts: sharing, solving or generating ideas. Most bad meetings are two of those fighting. Work in initiatives of six weeks or less, and run more experiments than feels comfortable.
Kill the three-letter acronyms: shared language is cheaper than shared process and does more work. Reward the team outcome over the individual save. Titles and pay only reach so far.
Structure keeps presenting itself as the answer because it is the part you can see. Most of the time the constraint sits upstream, in what you reward and how you frame it.
First published on Medium in January 2021, as part of a longer piece. Edited in 2026; the argument is unchanged.
Revisited, October 2026
I wrote that your role is the hat you wear on a team before AI started taking the hats. In 2026 I went back to the idea: which hats it takes, and the two someone still has to own. Hire the translator, not the title.