The Great Flattening: What It Means for Founder-Led Companies

Amazon and a long list of other large companies have spent the last two years removing layers of management and saying it will make them faster. If you run a company of 30 to 120 people, you've probably read those headlines and wondered whether you have too many managers, or whether you should hold off on hiring any. When a layer goes away, the decisions it used to make usually move up to someone more senior. Whether the company gets faster depends on who ends up with those decisions, and most companies don't work that out before they cut.

What is the Great Flattening?

The Great Flattening is the label the business press has put on large employers cutting middle managers and whole layers of management. The stated reason is usually bureaucracy, and lately AI.

Live Data Technologies found that middle managers made up almost a third of U.S. layoffs, up from about 20% in 2018 (Bloomberg, March 2024). In September 2024 Amazon told employees in a memo that it would raise its ratio of individual contributors to managers by at least 15% by the end of the first quarter of 2025 (Fortune, November 2024). Korn Ferry's 2025 workforce survey found 44% of U.S. employees saying their company had cut manager levels.

Each layer adds a handoff, which slows things down, and the original intent gets garbled along the way. McKinsey surveyed about 1,200 managers and executives on decision making. At organizations with one to three reporting layers, 70% of respondents said their company makes high-quality decisions and 61% said it makes them quickly. At organizations with seven or more layers, those numbers were 45% and 38% (McKinsey, 2019).

Where do the decisions go when a layer comes out?

Usually they go up. Julie Wulf at Harvard Business School studied more than 300 large U.S. companies over about 15 years. Her study looked at flattening near the top of the company, and her conclusion was that it "looks more like centralization." CEOs got more involved with the people below them, and decision rights moved toward the top instead of spreading out. In her words, "flattening can lead to exactly the opposite effects from what it promises to do" (HBS Working Paper 12-087, 2012). Her data covers big companies only, though I've seen the same thing happen at companies of around 50 people.

Say two teams both need the same engineer next week. The person who used to sort that out is gone, so the conflict sits until someone with the authority makes the call. If that person is the founder, it can sit for days while both teams wait.

In McKinsey's survey, respondents whose companies made decisions at the right level, often by delegating them, were 6.8 times more likely to be in the top group on decision quality and speed (McKinsey, 2019).

What happens to the managers who are left?

They're managing bigger teams, and Gallup's numbers show it's wearing on them. The average manager had 12.1 direct reports in 2025, up from 10.9 the year before and nearly 50% higher than when Gallup started measuring in 2013 (Gallup, January 2026). The median is still five or six, so most of that increase comes from a smaller group of managers with very large teams, and 13% now have 25 or more.

The number I'd watch most closely is how much of their own work managers still carry. Almost all of them have some, and the median manager spends about 40% of the week on it. Above that line, Gallup sees engagement drop, and it drops faster as the team gets bigger. In a smaller company that's the normal setup, since your best engineer becomes a manager and keeps shipping code. A lot of the leaders I coach are in that spot, recently promoted and still carrying a good piece of their old job.

Manager engagement fell to 22% in 2025, down five points in one year, according to Gallup's State of the Global Workplace report from April 2026. Individual contributors held roughly steady, so Gallup puts most of the drop in global engagement, now 20%, on managers. Korn Ferry found that 40% of U.S. employees feel a lack of direction at work (April 2025). Korn Ferry didn't ask why, but my guess is it's the same thing Gallup is seeing in managers.

Team size matters less than the manager

Gallup found that manager talent matters more than span of control. Highly talented managers stayed engaged even with 25 or more reports and a heavy individual workload. Feedback made a big difference too. Employees who strongly agreed they got meaningful feedback in the past week were about 70% engaged regardless of team size, compared with about 25% for everyone else (Gallup, January 2026).

What the manager layer was doing

A lot of what middle managers do isn't written down anywhere. When a company removes the role without deciding who picks up each piece, those pieces drift up to the founder or get dropped. This table is the checklist I'd want a founder to go through before cutting a layer, or before deciding not to build one.

Founders skip this a lot, because the payroll savings show up right away and it usually takes a few months before anyone connects the slowdown to the cut.

What this means for a founder-led company

You may already be flat by default

Many of the founders I work with have 40 or 80 people and a structure where every decision still routes through them. It just happened as they grew. Now the founder is making calls other people could make, and the managers they do have are too busy to manage. Copying Amazon won't help much here. A company like this is already about as flat as Amazon is trying to get, and it has the problems Wulf and Gallup describe. In companies under 100 people, I think the savings from not having a manager are usually smaller than they look, because the founder is the one picking up that work.

At Fitbit, a lot of operations still ran on hallway conversations. The teams I worked with in Korea and Germany, and the contract manufacturers across Asia, couldn't walk down the hall and ask, so we had to write down who decided what. A founder-led company runs into the same thing once it has people in more than one place, or more people than the founder sees in a week.

What made it work at Blue Apron

At Blue Apron in spring 2022, four senior executives had a standing two-hour meeting every week. We retired it by building a RACI so ownership was written down, dashboards so people could see the numbers without asking, some automations, and roadmaps everyone could see. Once those were in place, most of what the meeting handled could be settled by whoever owned it.

A smaller company can use the same order. Settle who owns which decisions and get a regular operating cadence going before you change the org chart. If you change the org chart first, expect a lot of those decisions to land back on your desk.

The people below have to be willing to make the call

Moving decisions down has a catch, and it comes up a lot in my coaching work. People bring habits with them from past bosses. I've coached people who spent years being penalized for questioning a leader's decision, so they learned to keep their heads down and execute. Then they joined a company with confident leaders who actually wanted to be challenged, and they still held back out of habit.

A founder can write a clear decision-rights list, but people who learned somewhere else that pushing back gets punished will keep sending decisions up until they've seen, more than once, that it's safe to make the call themselves. That takes a while, and some people need coaching to get there.

When cutting a layer is the right call

If a manager role mostly passes information up and down and doesn't own any decisions, it slows things down and doesn't add much, so removing it is reasonable. The same goes for a layer that exists because someone needed a promotion and there was nowhere else to put them. In those cases, removing the layer probably will speed things up. To check, write down what that role decides on its own. If you can't get past three or four things, the role probably isn't pulling its weight.

Before you hire your first managers, or remove them

If you're deciding whether to add a management layer at all, do these first, whatever you end up deciding about structure:

• Write down the 15 or 20 decisions that come to you most often, and decide which ones someone else should own.
• Set a weekly rhythm that doesn't depend on you being in the room, with a short review of a handful of numbers.
• Give the managers you have time to manage. If they're spending well over 40% of their week on their own deliverables, their teams will feel it.
• Get your managers some coaching. Gallup found manager talent matters more than team size, and most people promoted into management at a growing company got the job because they were good at something else.

Where fractional leadership fits

A fractional COO or Chief of Staff is one way to cover the gap while this gets built. The role sits between the founder and the team for a set number of days a week, takes over routine decisions, and puts the operating rhythm and decision rights in place so they keep working after the engagement ends. It's a reasonable option when a full-time COO feels a year or two early. If your structure is already in good shape and you just need someone to run it, a full-time COO is the better hire. There's more on that choice in What Is a Fractional COO or Chief of Staff, and Why Do You Need One?

What about AI?

AI is the newest reason companies give for flattening. Gartner predicted in 2024 that one in five organizations would use AI to flatten their structure by the end of this year.

AI does make people more productive. I've seen it firsthand. At one professional-services firm I worked with, AI-integrated tools cut manual work by about 40%. Over the same twelve months, the firm's revenue run rate grew about four times.

Whether AI is why so many managers are being cut is another question, and I'm skeptical. A lot of these cuts look like short-term cost decisions to me, and AI makes a convenient explanation. If AI really does save your people that much time, cutting headcount is only one way to use it. You could get products to market faster, or offer more products with the same team. Those options take longer to show up in the numbers, which is probably why they get passed over. I'd rather see a company put that time into the roadmap it never had the people for.

Some of a manager's work can be handled by better tools: status updates, pulling reports together, scheduling, first drafts. The parts of the job that stay with people are the judgment calls in the table above, like settling a conflict between two teams or noticing that a good employee is struggling. Use AI for the reporting work, but someone still has to make those calls. At a small company without managers, that's usually the founder.

How this looks in the Research Triangle

Hiring in the Triangle has been uneven this year. Raleigh-Cary job postings on Indeed were at 109.5 in early October, only modestly above February 2020 (100), and Durham-Chapel Hill's were below their pre-pandemic level at 94.7, according to the Triangle Hiring Brief. That makes it tempting to leave manager roles open. If you already know you need one, I wouldn't wait for the market to pick back up. The latest Triangle job market numbers are updated with each issue of the Brief.

Frequently asked questions

What is the Great Flattening?

The Great Flattening is the trend of large companies cutting middle-management layers to reduce cost and speed up decisions, increasingly with AI cited as the reason. Middle managers made up almost a third of U.S. layoffs in a 2024 Live Data Technologies analysis, up from about 20% in 2018, and 44% of U.S. employees said in 2025 that their company had cut manager levels.

Does removing middle managers make a company faster?

It depends on whether someone clearly takes over the decisions those managers were making. Research by Julie Wulf at Harvard Business School found that flattening often moved decisions to the top. McKinsey found that companies making decisions at the right level, often by delegating, were 6.8 times more likely to be strong on decision quality and speed.

How many direct reports should a manager have?

There's no universal number. Gallup's 2026 research found the average is now about 12, with a median of five or six, and that manager talent matters more than team size. A manager with a big team can do well with time to manage and regular feedback conversations with their people.

Should a small company cut its managers?

Most companies under 100 people don't have extra layers to cut, and many are already flat with the founder as the bottleneck. Cut a manager role only if it doesn't own any real decisions and mainly relays information. Otherwise, sort out decision rights and the operating rhythm first.

What happens to decisions when you remove a management layer?

They move to whoever has the authority to make them, which is usually someone more senior. If that person is a busy founder or executive, decisions wait and work slows down. Write down who owns which decisions before you change the structure.

Why is manager engagement falling?

Gallup found manager engagement dropped to 22% in 2025, down from 27% the year before. Managers are leading larger teams while still doing a lot of their own individual work, which leaves less time for their people. Gallup named falling manager engagement as the main driver of the global engagement decline.

Is AI the real reason companies are cutting middle managers?

AI is often the stated reason, but many of these cuts look like short-term cost decisions. AI does make people more productive, and that gain can also be used to ship faster or offer more products with the same team. Cutting headcount is only one way to use it.

Can AI replace middle managers?

AI can take over parts of the job, like status reporting, scheduling and pulling information together. The judgment calls and the coaching still need a person. The best use is taking reporting work off managers so they have more time with their teams.

What is the alternative to hiring a full-time COO when the founder is the bottleneck?

A fractional COO or Chief of Staff works with the company a set number of days a week to take on routine decisions and build the decision rights and operating rhythm the company is missing. It fits when a full-time hire feels early. If the company already runs well and just needs someone to keep it running, hire full-time.