I have been at a few places now that seem to want to have a flat org when it's convenient and a hierarchy when it's convenient. I've found that hierarchical structures with established roles and responsibilities work best. Flat structures seem to allow no one to have accountability.<p>In your experience have you ever actually seen a "flat" org work? And if we are being honest is what you are describing a "flat" org, or just an org that claims to be flat but in reality there are clear power dynamics and decision makers with unofficial influence?
Yes. It requires a lot of trust. You cannot have any micromanagers. Typically small research companies do very well flat/shallow/leveled.<p>I really like galois and qumulo as examples of shallow companies.<p>Typically flat doesn't refer to no hierarchy, but instead no direct middle men. You may have three people who act as 'bosses' to different responsibilities you have at the company. They are actually the boss of that responsibility, not the boss of you. That means that individual is shared across a greater number of people as a resource relating to that responsibility.<p>Additionally flat doesn't prevent for conventional bosses, but also doesn't limit who can be that conventional boss. For this project Sam might be the boss, because of Sam's expertise. For the next project I might be the boss of Sam. Who is in charge is not defined by seniority or job title.<p>Another flattening idea is to distribute knowledge. Qumulo used to rotate which team you were were on every 4-7 months. I don't know if they still do this, I don't even know if they're still flat. But this forced you to understand a large amount of the codebase. You had to have a holistic understanding of their product. This made tenure highly correlated to product expertise.<p>For larger companies flat/shallow/leveled are harder to find, but grid companies do exist as alternatives to conventional hierarchies. JPL is A good example of a grid company. as a policy JPL allows you to assign any employee do your project as long as they have time and express interest. You often find this in parts of the national Labs as well.