Something strange happened at work a couple of months ago: We parted ways with an employee. That doesn't happen too often at 37signals, the Chicago software company I co-founded. In 11 years, we've lost just five people—and one of them came back seven years later.

But the really strange thing was the reason this employee and I decided it was time to part ways. The issue was ambition—not a lack of it, but more of it than we could use.

She had been in our customer service department for about three years and had always done quality work. Her smarts, initiative, and productivity had never been in question. Now, she wanted that performance to be rewarded with managerial responsibilities and a new job title. It wasn't about the money.

You may be wondering how this possibly could be a problem. After all, what business owner doesn't crave an office staffed with motivated people looking to take on more work? In fact, companies often go out of their way to create positions, responsibilities, and job titles to hang on to their most ambitious employees.

At 37signals, however, we have a different position on ambition. We're not big fans of what I consider "vertical" ambition—that is, the usual career-path trajectory, in which a newbie moves up the ladder from associate to manager to vice president over a number of years of service. On the other hand, we revere "horizontal" ambition—in which employees who love what they do are encouraged to dig deeper, expand their knowledge, and become better at it. We always try to hire people who yearn to be master craftspeople, that is, designers who want to be great designers, not managers of designers; developers who want to master the art of programming, not management.

Instead of rewarding high performers with managerial responsibilities—which often drives people further away from the job they are actually good at—we reward with responsibilities closer to the work. We also throw in above-market salaries and generous benefits, including four-day workweeks in the summer, as much time off as they would like (within reason, of course), and plenty of freedom to make their own decisions about the projects they're working on.

This has served us well over the years. But recently, as we've brought on more people, that model has been showing signs of strain. We're now at 26 people. And, as many entrepreneurs have learned, once your business reaches a certain size, matters you didn't have to consider before become difficult to ignore. In our case, HR terms like departments, managers, and titles have begun to pop up more often.

Besides being small, 37signals has always been a flat organization. In fact, flatness is one of our core values. We have eight programmers, but we don't have a chief technical officer. We have five designers, but no creative director. We have five people on our customer support team, and no customer support manager. And because we don't have a marketing department, we don't have a chief marketing officer.

Even as we've grown, we've remained a lean organization. We do not have room for people who don't do the actual work. Nearly everyone at 37signals touches our products at one point or another. From writers writing and updating support documentation to designers designing the user interfaces to programmers writing the code to our operations people, who make sure the servers keep on serving, we don't have delegators who get paid to tell other people what to do.

We've experimented with promoting a few people to manager-level roles. In some cases, this has worked out; in others, it hasn't. But one thing we've found is that groups that manage themselves are often better off than groups that are managed by a single person. So when groups do require structure, we get them to manage themselves.

For example, back when we had three people working in customer service, we hired someone to manage the team. His responsibility was to review everyone's tickets, keep an eye on tone, make sure our customers were getting prompt and proper responses, interact with our developers to let them know what our customers were requesting, and measure our overall support performance. He might jump in and handle a few tickets here and there, but his main responsibility was to step back and improve the department.

It didn't work out. This isn't a knock on that particular manager (he was a great guy who knew how to run a department, and we helped him find another job). It wouldn't have mattered who was in the role; the role itself was unnecessary. But because we expected the department to grow, we thought it would be a good idea to get some structure in place. After all, everyone knows that as you add people, you add structure.

What we learned is that adding a dedicated manager and creating a hierarchy is not the only way to create structure. Instead, we decided to let the team be entirely self-managed. There's still a team leader, but that role rotates among the team every week. Each week, a new leader sketches out the agenda, writes up the notes about problems and performance, and steps up to handle any troubled customer interactions.

One of the things I like about the arrangement is that it frees us from the often toxic labor-versus-management dynamic, in which neither party truly understands what it's like to be on the other side. This is where you'll find a lot of conflict in companies. But because we rotate management duties weekly, everyone is more empathetic toward one another. When you'll be management soon, you respect management more. Same with labor. It reminds me of one of my favorite quotes, by the philosopher John Rawls: "The fairest rules are those to which everyone would agree if they did not know how much power they would have." Our support has gotten better, and our customers are happier. We've measured the difference, and we know it works.

Observing the performance of the customer service team reminds me that promoting horizontally instead of vertically can benefit everyone involved. Moving ambitious people upward tends to lock down other capable people on the team. If there are three or four people capable of managing (in a traditional sense) and we promote just one of them, the teammates might feel stuck. We would rather everyone work together in an environment in which everyone has a chance to move proudly and thoroughly sideways.

In the end, no one here is happy when an employee leaves. But as a business owner, I have to think about the long term. Moving someone up to a managerial position just because he or she outgrew his or her current position isn't reason enough. Adding managers to the mix sends a strong cultural statement about flatness giving way to hierarchy. We're definitely not ready for that now. I hope we never will be.

As it happens, this story has a happy ending. Our ex-employee didn't find another job—she launched her own business. She runs the show, and she's having a blast. We've even provided advice and helped her promote her new business. It definitely was the right move for everyone.

Jason Fried is co-founder of 37signals, a Chicago-based software firm, and co-author of the book Rework.