In 2023, three McKinsey partners published a book called Power to the Middle, arguing that companies were about to make a serious mistake if they treated their middle managers as overhead. Bill Schaninger, Bryan Hancock, and Emily Field had the internal McKinsey data to back it up. Nobody who mattered seems to have read it, or if they did, they filed it under things to worry about later.
Later arrived fast. Bloomberg's analysis with Live Data Technologies found middle managers made up a third of all layoffs in 2023, the same year the book came out. Korn Ferry's most recent workforce survey has 41 percent of professionals saying their company cut manager-level roles in the past year. It isn't everywhere. Challenger, Gray & Christmas data through mid-2026 shows total corporate layoffs actually down 10 percent while tech layoffs are up 33, so this is a specific, concentrated bet, not a universal one. But where companies are placing it, they're placing it against the exact case Schaninger's team had just made.
The bet rests on a clean idea: middle managers mostly coordinate, AI can now do a lot of coordinating, so the layer is redundant. It's a tidy theory. It's also not what the data says middle managers were actually doing with their time.
McKinsey's own research on this, separate from the book, found managers spend less than a quarter of their time on strategy and less than a third on the part of the job that's actually about developing people. Nearly half their week goes to work that would belong to an individual contributor if anyone were honest about the org chart. Only one in five managers strongly agrees their company sets them up to succeed at the job. So the theory that got acted on wasn't really "AI replaces coordination." It was closer to "AI replaces coordination, and we'll also stop pretending the other half of this job existed," which is a different and much less flattering bet.
Gallup has been measuring what happens to the people left holding that job, and the trend line was already moving before this year's cuts. The average manager's span of control grew from 10.9 direct reports to 12.1 in a single year, part of a nearly 50 percent climb since Gallup started tracking it in 2013, based on a meta-analysis of over 92,000 teams. Cross a certain threshold of individual-contributor work on top of a growing team, and a manager's own engagement drops, and keeps dropping as the team gets bigger. Gallup's newest global workplace report has manager engagement falling from 30 percent in 2023 to 22 percent in 2025, the steepest single-year decline of any group they track, most of it landing in exactly the twelve months the flattening accelerated. Worldwide, low engagement is now estimated to cost the economy 10 trillion dollars a year, about 9 percent of global GDP. Correlation, not proof. But it's an odd coincidence to wave off.
Here's the part that should have been the headline instead of the layoff number. McKinsey found that relationships with management specifically account for 86 percent of what makes someone's workplace relationships satisfying at all, and workplace relationships account for 39 percent of overall job satisfaction. That's not a soft metric sitting next to the real ones. It's most of the actual mechanism by which a company's strategy reaches a person doing the work, translated into something they can act on, and by which that person's concerns travel back up before they calcify into a resignation letter. Cut the layer, and you haven't automated the coordination work away. You've just made someone find out the hard way what it was actually doing.
Korn Ferry's follow-up survey found 37 percent of employees felt directionless after a manager was cut, and 43 percent said their leaders weren't aligned with each other. Neither number should surprise anyone who's read the time-allocation data. The people best positioned to catch misalignment before it reached the frontline were the ones just laid off, on the theory that they weren't doing much of anything anyway.
Schaninger's book didn't argue against flattening in general, or claim every manager earns their keep. Some layers exist for reasons that have nothing to do with coordination and everything to do with how the org chart accumulated over a decade of expansions nobody cleaned up. The argument was narrower, and better for it: know what the layer is actually absorbing before you remove it, because the job description and the real job have not been the same thing for a while. Three years after the book came out, most companies still don't seem to know the difference. They just know the layer got cheaper to remove than it used to be.
Sources
- Span of Control: What's the Optimal Team Size for Managers? (Gallup, 2026 meta-analysis): gallup.com
- State of the Global Workplace: 2026 Report (Gallup): gallup.com
- Stop Wasting Your Most Precious Resource: Middle Managers (McKinsey & Company): mckinsey.com
- Power to the Middle: Why Managers Hold the Keys to the Future of Work, Schaninger, Hancock & Field (2023)
- The "pure manager" layoffs: how AI is reshaping middle management (Ramp Economics, reporting Korn Ferry, Bloomberg/Live Data Technologies, and Challenger Gray & Christmas data): ramp.com