For decades, many investors believed it took one sort of leader to start a company and another to scale it. Especially in Silicon Valley, where I work, the stereotypical founding CEO would be a young engineer who'd start the company and reach product-market fit, and the second CEO would be a gray-haired executive with a business or sales background. Then, around 10 years ago, the common wisdom began to change. Many investors now favor keeping founding CEOs for as long as possible.

I've seen this change up close. I started my first company, Udemy, as a product-focused founder. I got to the hypergrowth phase and then handed the reins to a more experienced executive, my COO. Today, I'm running my second company, Carbon Health, with a new set of assumptions. I'm the founding CEO, and I will remain CEO for as long as possible.

I think this new attitude will be a good thing for me, for Carbon Health, and for other founders and their companies. But there's still truth in the idea that it takes one kind of CEO to start a company and another kind to scale it.

Leading a team of fewer than 200 is different from leading a team of 2,000-plus. During the early stages of a company, everybody's either building or selling. You, as a founder, need to contribute to at least one of those things (and ideally to both). But when your organization grows, keeping everybody aligned becomes your main job--and that requires a new set of people skills.

Tell Good Stories

One way to look at a startup is as a group of people uniting around a story. The Carbon Health story is that the smartest doctors, designers, engineers, and operators together can make the health care experience 10 times better for average Americans--without increasing the cost. Udemy's story was that the internet would one day become the primary learning destination, and that learning a new skill is the most effective way to improve one's life.

Stories matter--for your company's values, strategy, operating principles, and brand. But to tell the right story and tell it well, you need to find simple keywords or tag lines that people can remember. You need to frame the story for your team, and you can't outsource that. It has to come directly from you, their leader.

Early on, I wrote out four company values that would define the culture I wanted to create at Carbon Health. I still repeat them constantly. One of those values, the three-word sentence "Assume karma exists," has done wonders in aligning a large organization under the idea that doing the right thing will yield disproportionately good results, even if it's not immediately clear how.

When the pandemic hit, for example, we provided free virtual Covid consultations, set up testing trailers in underserved communities, and open-sourced our clinical data sets for researchers, all at an operational loss. Those efforts won us superfans. Hiring got easier. Governments linked to our resources, which boosted our search-engine rankings. Google offered us exclusive advertising terms. We ended up with massive brand recognition and commercial value that we never would have gotten otherwise.

Ideas and stories are great, but you also need lived experiences to help people internalize the ideas. Iif i don't admit mistakes, why would anyone else? And if nobody admitted mistakes, how would anybody learn from them?

More broadly, framing your story is about helping everyone see the big picture and understand how they fit into it. We don't do "conversion optimization" at Carbon Health; we reduce friction for people who are trying to get health care. Our legal team doesn't talk about "compliance" as something we need to do to avoid being sued; we talk about it as a set of practices to protect patient privacy.

It's your job as the CEO to make sure everyone knows that their job matters. By framing your story in this way, you not only lead people to make the right decisions, but also inspire them to believe in their work.

Establish Rituals

We start every all-hands meeting the same way. We read some patient stories, meet the people who have recently joined us, announce promotions, and celebrate the things we do that show our authentic commitment to our mission. This is a ritual. During each all-hands, we also have something we call the CEO corner, where I try to display the kind of candor I expect from the rest of the organization by discussing my own mistakes.

Back in 2016, for example, I was very vocal about feeling that we should design patients' charts around their health issues rather than around their visits. It just made more sense, because the information would be structured in a more useful way. Over the years, though, we learned that doing it my way hurt doctors' productivity, and we ended up having to rebuild the system with a more standard workflow. It was an expensive project and offered an important lesson--for me and for the team--in how assumptions can be proved wrong.

The key idea behind such rituals is reinforcement. Ideas and stories are great, but you also need lived experiences to allow people to internalize the ideas. If I don't admit mistakes, why would anyone else? And if nobody admitted mistakes, how would anybody learn from them?

Rituals aren't something you want to clone from another company. When we hit an inflection point in Carbon Health's growth and realized it was critical that we more formally establish our culture, we gathered the people and executive teams to brainstorm our cultural tenets and share those with regional leaders, asking them to vote on the principles that resonated most. The result is an immersive, two-day onboarding program called Ascent that uses interactive lessons and hands-on activities to take new hires through Carbon Health's values and culture and how we do things differently. The program reinforces rituals and is itself an important new ritual.

Make Sacrifices

Your values don't mean anything if you follow them only when it's convenient. One of our values is to hire people who genuinely care. Often, we see people with super-impressive backgrounds who lack that sense of mission. It's extremely tempting to make exceptions, but it's vital that we don't. We sacrifice that convenience and allure to uphold our values.

Similarly, it would be convenient if we followed the paths of many large organizations and created two classes of employees: corporate employees and everyone else. Amazon's corporate employees and warehouse employees don't have the same benefits. Retail companies don't treat their retail employees the same way they treat their corporate employees. For many companies, service employees and laborers are simply easier to replace than the execs and engineers at headquarters.

At Carbon Health, we decided to treat everyone alike. Yes, people get paid differently, ­depending on their jobs and experience, but we try to keep everything else as closely aligned as possible. This often means our Silicon Valley employees don't get the same lavish perks that other companies in the region provide--and we're very transparent about that. The payoff is that we get people who truly want to be here.

Every time you can take a hit for your principles, consider it an opportunity. People will remember when you sacrifice short-term convenience to defend your values. Is there a user interface problem that happens to be decreasing churn? Fix it anyway, and then share your reasoning with the team. They might even share the story--and they'll certainly respect the leadership.