Flattening removes the layer. The work moves somewhere less visible.

A layer of management can disappear from an org chart without a single approval, coaching conversation, or escalation disappearing from the work. Fast Company reported that Meta invited some employees in its Applied AI division to take manager roles again. The invitation was reportedly voluntary and limited; Meta declined to comment. The useful question for a business leader is where the management work went when the roles were removed.

Korn Ferry's Workforce 2025 survey of more than 15,000 professionals found that 41 percent said their organization had reduced management layers, while 43 percent said leaders were misaligned. Gallup reported that average span of control rose from 10.9 direct reports in 2024 to 12.1 in 2025. Those numbers do not prove that flatter structures caused the alignment problem, and larger teams can work when managers have the capability, the work is coordinated, and feedback is frequent. But they do show why payroll savings are the wrong headline measure on their own. If a layer comes out and the same approval, coaching, coordination, and escalation work still exists, the cost has moved rather than disappeared.

Gallup's own distribution matters here. The median manager still has five to six direct reports, where it sat in 2019, and two thirds of managers oversee fewer than ten people. The average rose because the largest teams grew. Gallup also reports that 97 percent of managers do individual contributor work, with a median 40 percent of time spent on it. That is a more useful operating measure than span alone because it gets closer to the real constraint. A manager can carry a wider team if the role is mostly management. A manager already spending much of the week doing individual contributor work is carrying two jobs.

Gallup found that engagement held steady across team sizes when managers spent less than 40 percent of their time on individual contributor work. Above that level, engagement declined as team size increased. This is an association, not a universal causal threshold, and Gallup's analysis precedes the latest rise in large teams. It is still a reason to measure the work before reorganizing people. Look at how many hours managers spend producing individual outputs, how long decisions wait for them, and how often coaching and escalation are postponed.

Before eliminating a layer, take two comparable teams and measure manager time for four weeks: individual production, coaching, approvals, escalations, and decisions waiting for an owner. Remove or reassign a narrow set of non-management tasks first, with the responsible business leader approving the change. Then check whether decision delays, missed coaching, or team outcomes improve. If capacity returns, redesigning reporting lines may be unnecessary. If it does not, the evidence points to the role, coordination model, or structure that still needs to change. That is a smaller test than a reorganization and gives the leader a result to measure before claiming a payroll saving.

The evidence does not say every flat structure is harmful. Some layers genuinely duplicate work. Gallup's time-allocation figures are self-reported, the Meta story covers one division and anonymous sources, and Korn Ferry sells organizational advisory services. A layer reduction earns its business case only when the new owners of the work can carry it without degrading decisions, support, or performance. Count the hours and the results, not just the boxes removed.