Posted on by & filed under leadership.

Recently I have been spending a lot of time thinking about non-hierarchical teams. I love the idea of empowering employees. I hate bureaucracy and inefficiency, so the concept of removing the middleman and letting people get things done sounds pretty darn awesome to me.

But even though I love the utopian idea of “no managers”, there are some key functions that, at the end of the day, someone has to do. And it is those responsibilities (and their effects) that lead me to pursue excellence in technical leadership.

This post is an exploration of sorts, into the concept of management and how it works today. It details my current thoughts as to why large organizations need managers, or at least why they need a way to serve many of the same functions (if disguised in other titles or descriptions).


When does having “no managers” makes sense?

There are definitely times and places where “no managers” makes sense.

For those of you in disbelief, let me explain why.

As an organization grows, each person tends to know the most about their domain (everyone is a localized expert in a way), and they have a solid grasp on the state of their systems, the demands and requests placed on them, and the holes or missing pieces.

If you assume employees want their company to be successful (otherwise, you have some hiring mistakes to handle), then it follows that they would want to do the work that would lead to that successful outcome. In order to do this, though, the employee must know the goals of the organization and how the work they do fits into those goals. That is, they must know how to prioritize their work to achieve the highest level of success for the company.

However, in the above line of reasoning lie a whole lot of assumptions.

We assume that employees have all the information they need to prioritize. We assume that each person is motivated and capable of executing on those prioritize, and that their motivations and skills align with the company’s success. Each of these assumptions presents a challenge that has to be addressed directly in order to have a thriving “no manager” environment.

At scale, is it possible for everyone to prioritize without hierarchy?

In our startup we have no managers. Kate S decides what to do and when to do it. She keeps me in the loop, and I keep her in the loop with my work. Each week we share our plans and goals over email.

There are moments I have felt like a manager, though. This happens when I haven’t done a good job sharing goals, plans, or objectives for the company. In these moments, I find myself telling Kate something about a customer meeting or experiment I did that I forgot to share with her that would impact her plans or designs. When you are moving this fast, without a proper system for sharing these kinds of communications, occasionally things get lost.

Over the course of this year, we have worked out some implicit systems. We try to send the TLN on Saturday nights.  We aim for two blog posts per week, one for each of us. We send our status updates on Mondays. We have a shared calendar where we put planned launches, promotions, and our time off. We work together to regularly review and revisit our plans and priorities.

In big companies, on the other hand, understanding your work in the context of the whole company can be very challenging. When I was at Microsoft (admittedly this was over 10 years ago, so it may be different now) we had no idea how people used our work, or even how our giant organization contributed to the bottom line. In my experience at Amazon, there was more access and visibility, but even then it was hard to say if one team was overstaffed or understaffed when compared with another. Even as a manager, with visibility across organizations, it was challenging to weigh the contributions of each team to the company, so I can’t imagine doing it as an individual contributor.

This is why big companies create divisions. It is much easier to understand how your work fits into a vertical or area of the company. And thus the hierarchy begins.

In small companies, like startups, your goals are clear. You have limited runway, you have clear customer segments, and you have a plan to get you from where you are to success. The smaller the team is, the more obvious it is how each person contributes to the plan.

This makes it easier for each person to make decisions and act independently from one another, yet still understand how their work is moving the company forward. Moreover, the impact of each person is often felt quickly, so wrong moves or missteps can be easily rectified.

Therefore, one big requirement for eliminating managers in a larger org is to establish clear priorities, and to have the right systems to streamline communication in place so everyone can say on the same page.


Are people in the right roles?

In a non-hierarchical organization, people still have roles and responsibilities, though they may expand and contract at times. When teams are big, people become more specialized.

What happens if you happen to hire someone who isn’t adept at being self-directed? What about if you hire someone who is self-directed, but they are more motivated by solving interesting problems that advance their skills, than the non-glamorous project that the company needs finished? In these situations, you have a people problem.

In small teams, hiring mistakes tend to happen less frequently than in organizations that are comprised of hundreds of people, because in a small company every hire is subjected to careful review, usually by the founding team. At some point in the growth of an organization, though, it isn’t possible to review every single hire, and so you have to trust others to make those decisions. But even if you have great leaders, we all make mistakes now and again.

This means that in any big organization, there are people that aren’t great hires. This is the whole principle driving the concept of the vitality curve, which brought in the awful practice of stack ranking. At any given time in a large organization there are people at the bottom of the pile, and so as an organization you can improve the quality of your staff by topgrading.

Of course, I have worked in small teams where every person was an A player, where adopting practices like stack ranking or topgrading would have been downright silly. However, in large organizations, it is very difficult to maintain all A players over a long period of time. If you have B players, or even C or D players, can you still have no managers?

There reaches a point where you need someone with oversight to come up with a way of handling people that aren’t the right fit, or that were the right fit but aren’t anymore (whose goals and motivations changed, for example).

Someone needs to give feedback to people, to coach them to develop them, and to be their advocate and sounding board. There are some companies that try peer reviews to solve this problem, but most people are reticent when they know their praise or criticism will be used for other people’s raises and reviews.

Most of the time employees are not all bad, or all good. There are a lot of gray areas – and so it can be hard for peers to make the call if the person is a good fit for the organization. At the end of the day who can be responsible, and how will that scale with more people? Managers, perhaps?


What else do people need from managers?

So far we have established that people need clear priorities, and that someone needs to be responsible for forcing hard decisions and handling people that happen to be in the wrong roles. The third area to cover is rewards and recognition.

We know we want employees to be motivated by the company’s success. How do incentives work in this new paradigm? Do raises, promotions, and recognition need to align with company success? This seems like a good idea….in theory.

What happens to the employee that does their very best on a product that doesn’t resonate in the market as expected? Or what if one member of the team did a great job building the product, but the designs just didn’t hit the spot?

Should the great work go unrewarded because the business results aren’t worthy? Maybe, but people will leave an organization if they feel underappreciated, or if they don’t have the control to impact rewards and recognition for their work.

What happens if you are on a team of people, and engineering misses their deadline, causing marketing to change their optimum launch plans, resulting in lower sales? If you are a superstar it sucks to lose out on rewards or recognition for things you couldn’t possibly control. It is like everyone in a class at school getting the same grade. Sure, it encourages teamwork, but at some scale there is a point where your locus of control can’t actually impact the end result.

You want to treat everyone equally, but if you treat everyone the same no one will be happy. 

However, you can still be fair and give people what they deserve in an equitable way. And this is where managers can help. They have the visibility and insight into each person’s performance. and they can be responsible for recognizing a job well done, even if the company incentives don’t want to reward it. Great managers treat each person like the unique snowflake that they are. They can learn what matters to them and figure out how to deliver those things, because what matters to people isn’t always the money.

A manager has the ability to understand what each member of their team needs to stay motivated, and they have the power to deliver those rewards when they feel good work should be recognized (outside of bigger business goals). So, another point for managers?


Hierarchical or not?

There are definitely great things about empowering people and letting them do what they think is best, without the supervision or direction of managers. Smart people will be happier and it increases collaboration.

However, there are key elements that are central to these decentralized organizations. All people need clear priorities that fit into larger company goals, and they need clear channels to communicate changes. Employees need coaching, and ultimately there needs to be someone who is responsible for handling problem performers. And finally, we all want to be recognized for a job well done, so there needs to be someone whose role is to understand the best way to do that for us.

Do those roles need to be filled by managers? No. However, at scale it isn’t clear to me how to do it without them.

What do you think? I would love to hear what *you* think about getting rid of managers, if it’s possible, and how you’d do it at scale. Leave your thoughts in the comments!


Click here for Part 2 of this series on flat organizations and teams without managers, called “Succeeding on a team with no managers”.


Tags: better leader, change, leadership, leadership practice, management, team, thinking,

4 Responses to “Why I want to get rid of managers, but why I can’t [Part 1 on flat orgs]”

  1. Eric Kingery

    I would call it ‘all managers’, as opposed to ‘no managers’. If you drop the work that a good manager is supposed to do, the company will lose. If you can decentralize and distribute that work to every employee and not need a hierarchy, that’s great. That’s a lot of responsibility to take on. I do respect the attempt at Valve. I suspect there are many lazy implementations which lead to total disasters. Even given the amount of effort put into the concept at Valve, a recent employee said it “felt a lot like high school” there.

  2. katemats

    Thanks for the comment Eric. And I hear you. All managers does sound right. I am working on a new post now that talks about what skills someone might need to have to be successful in such an environment and it does sound a lot like “manager” skills :)

    • Luis Goncalves


      Today I published a blog about the same topic:

      :) I agree with a lot of things that you say but I believe a lot of them are not necessary true :)

      For example you do not need to reward everyone the same :) You can give that decision to the people… People can chose who will receive what :) There are companies using this, you delegate that to the teams. They know the best who did what :)

      You referred missed deadlines, I see what you say but If your company is agile enough to not have managers, sw development is the first one to be impacted and in this kind of context we always release something, maybe not with all features but we still release something :) we should not have missed deadlines.

      I know its an extremely complicated topic and I am trying to find more about it.

      In any case congratulations for your courage to talk about something so polemic :)


      • katemats

        Thanks for sharing your thoughts and your post Luis.
        I still don’t think it is scalable, and there more conversations I have the more I believe this to be true.

        All you do by making managers “leads” is just changing a job title. The job function is still there it is just named differently or the responsibilities are distributed differently – there is still people managing the organization, coaching poor performance (which does happen especially given time), paying attention to career advancement, and setting a vision.

        Agile works for tasks, but one of the big criticisms with agile is its failure for long term planning and long term strategy, so even in your model someone needs to own those things. And that means someone needs to step up and lead, or be a manager – depending on how you choose the names of your roles and responsibilities.