Back to Blog
LeadershipMarch 18, 20266 min read

I Was Wrong About Leadership Structure

I built leadership structures that mirrored the Army for years. It made sense. I served. The model worked: clear chain of command, redundancy at every level, defined roles with defined authority. If someone goes down, the next person steps up. Survivability. Aggressive execution. Structure that absorbs chaos.

I defended that model in every coaching engagement. Build your org chart like a military unit. Define the chain. Create redundancy. It works.

I was wrong.

Not about all of it. The discipline transfers. The accountability transfers. The bias toward action transfers. But the structure itself, the hierarchy, the layers, the information flowing up and decisions flowing down, that's optimized for a world where information is scarce and decision-making is centralized. We don't live in that world anymore.

What Jensen Huang showed me

I started studying Jensen Huang's leadership at Nvidia not because I was looking for a new model but because the results demanded explanation. Nvidia didn't become the most valuable company in the world by running a traditional hierarchy. Jensen has something like 60 direct reports. No one-on-ones. Information flows to everyone simultaneously. The organization is almost aggressively flat.

My first reaction was that this shouldn't work. My Army-trained brain said: that's a span of control problem. No one can effectively manage 60 people. The information bottleneck alone should kill execution speed.

But it doesn't. Because the flatness isn't the absence of structure. It's a different kind of structure. One where information moves horizontally instead of vertically. Where decisions happen at the point of contact instead of getting escalated up a chain. Where the leader's job isn't to make decisions but to make sure everyone has the context to make good ones.

That reframing broke something open for me.

The two things I changed

I used to operate under two principles I thought were permanent:

First: build hierarchical redundancy. If your ops manager quits tomorrow, someone beneath them should be ready to step in. Layers of backup. Military survivability applied to business.

The problem: layers create distance. Every layer between the person doing the work and the person setting direction is a translation layer. Information degrades. Context gets lost. By the time a frontline insight reaches the decision-maker, it's been filtered through three people's interpretations. The redundancy that's supposed to protect you is actually insulating you from reality.

The shift: build the flattest possible organization imaginable. Not flat as in no leadership, flat as in minimum viable distance between the work and the direction. Everyone has context. Everyone has ownership. The redundancy comes from shared understanding, not backup personnel.

Second: celebrate in public, correct in private. This is leadership orthodoxy. Praise people where everyone can see. Pull them aside for the hard conversations. Protect their dignity.

I believed this completely. And I was wrong about the second half.

Why I correct in public now

Not to shame anyone. Not to make examples. Because the alternative is worse: everyone else makes the same mistake because they never saw the correction.

When you correct in private, the lesson reaches one person. The other fifteen people on the team who are about to make the same error don't benefit. They'll each need their own private conversation. You're doing the same work sixteen times instead of once, and the team never develops a shared understanding of what good looks like because all the calibration happens behind closed doors.

When you correct in public, with the right intent, something different happens. The correction becomes a learning event for everyone present. Not "Jon screwed up and got called out." More like "here's a pattern we all need to see, and here's how we think about it differently going forward."

The key is intent. Public correction done to punish is toxic. Public correction done to teach is one of the most powerful compound learning mechanisms a team can have.

Think about it this way: every private correction is a withdrawal from the team's collective intelligence. You're investing in one person's growth while the rest of the team stays blind to the lesson. Every public correction, done right, is a deposit into the team's shared understanding. The learning compounds because everyone receives it simultaneously.

What this actually looks like

It's not a free-for-all. It's not brutality disguised as transparency. It's a culture where:

We don't pull punches about where we're weak. If the estimating department is consistently missing by 15%, we say that in front of everyone. Not to blame the estimators. Because the project managers need to know it, the field teams need to know it, and the estimators need to know that hiding it isn't an option. The weakness is shared. The fix should be too.

We explore together. When someone makes a mistake, the first question isn't "why did you do that?" It's "what did we all miss that let this happen?" The correction isn't aimed at a person. It's aimed at a gap in the system that the person's mistake revealed.

We celebrate and reward in public too. This part I always had right. But combining public celebration with public correction creates something neither does alone: a culture where growth is visible. People see both what excellence looks like and what the path from mistake to excellence looks like. That path is the most valuable thing a team can share.

The compound learning effect

Here's what I didn't understand when I was building Army-style hierarchies: the speed of organizational learning is limited by the speed of information flow. In a hierarchy, information flows through channels. In a flat organization with radical candor, information flows through the air.

When a correction happens publicly, every person present updates their mental model simultaneously. They don't wait for it to trickle down through a chain. They don't get a sanitized version from their manager. They get the raw lesson, in context, with the emotional weight of the real moment.

That's compound learning. Each public correction makes the next one less necessary because the team's shared calibration improves with every iteration. The organization gets smarter as a unit, not just as a collection of individuals who each had their own private coaching moment.

The military model I loved was built for survivability in chaos. Redundancy, chain of command, defined escalation paths. That's invaluable when people are shooting at you. But in a business context, where the goal is speed of learning and quality of decisions, the flat model with radical public accountability produces better outcomes. Not because the people are different. Because the information architecture is different.

Here's the question I'd ask any leader still running the model I used to run: what is your team learning right now that only one person knows? How many corrections happened this week behind closed doors that the rest of the team needed to hear? How much faster would your organization move if every lesson landed on everyone at once instead of trickling through a chain?

The discomfort of public growth is real. I'm not minimizing it. But the cost of private growth is invisible, which is why most leaders never question it. You can't see the lessons that didn't compound. You can't measure the mistakes that repeated because the correction only reached one person.

I spent years protecting my teams from that discomfort. It was the most expensive protection I ever provided.

JMJon Mayo

Jon Mayo

Executive coach, author, and creator of WayMaker.

Liked “I Was Wrong About Leadership Structure”?

Get notified when new Leadership articles are ready.

Subscribed to: