Your managers are the least engaged people in the room

Gallup's own State of the Global Workplace page reports manager engagement at 22% in 2025, down five points from 27% in 2024, with data collected January through December 2025. Global employee engagement over the same period was 20%, down one point from 21%. Two things are worth pulling out. First, managers are barely ahead of the people they manage, which is historically unusual, because managers have generally reported higher engagement than individual contributors. Second, the manager figure is moving faster and in the wrong direction. A one-point move in the overall population alongside a five-point move among managers is not a general malaise, it is a specific population under specific pressure. I am citing Gallup directly here rather than secondary coverage, because the secondary write-ups of this report disagree with each other and some report a nine-point drop against a 2022 baseline that Gallup's own summary does not use.

Because facilitation is not a task you can complete while absent. Most of a manager's work tolerates a bad week. Reports still get written, approvals still get granted, the inbox still empties, and nobody downstream can tell much difference between a manager who cares and one who is running on procedure. Running a room is the exception. The whole function is attention: noticing who has not spoken, hearing the objection underneath the question, deciding whether a silence is thinking or avoidance, choosing not to fill it. Every one of those is a live judgement that requires presence in the moment. A disengaged manager does not run a bad meeting in an obvious way. They run a meeting that proceeds correctly through the agenda and surfaces nothing, and everybody leaves vaguely unsatisfied without being able to name why.

Partly, and this is where the usual response goes wrong. The reflex when manager performance drops is to schedule leadership development, which means adding four hours to the calendar of the person whose calendar is the problem. Gallup's own data supports training as a lever, reporting substantially better management performance where managers received best-practice training rather than none. So the training works. The implementation is what fails, because it is typically delivered as an additional obligation to somebody already carrying more than they can hold, by a vendor who will not be there on Tuesday when it has to be applied. The useful question is not whether managers need training. It is whether the thing you are adding replaces something or stacks on top of it, and in my experience it almost always stacks.

The span and the sandwich. Managers absorbed flatter structures after successive reorganisations, which means more direct reports each, and they absorbed the coordination cost of hybrid work, which is real and mostly invisible. On top of that they are now the implementation layer for AI adoption, expected to work out what their team should automate while also reassuring that team about job security, with little guidance and no additional time. And structurally the role has always been squeezed: accountable for outcomes they do not control, carrying downward pressure from above and upward frustration from below, with the authority of neither. None of that is new. What is new is the volume, and the five-point drop is what it looks like when volume outruns the role's capacity to absorb it.

Four things, roughly in order of how much they cost. Cut meetings the manager is required to attend but does not run, which is usually the largest recoverable block on their calendar and costs nothing but a decision. Give them one skill rather than a program, because a manager who can hold a four-second silence and ask one good follow-up question will outperform one who has been through a leadership curriculum and absorbed none of it. Make the span visible, because a manager with fourteen direct reports is not underperforming at one-to-ones, they are arithmetically unable to do them, and no amount of time management fixes a number. And stop measuring their meetings by attendance, because that metric rewards scheduling rather than outcomes and a disengaged manager will hit it comfortably.

No, and the data does not support that reading. Gallup's 2026 figures show engagement differences by work arrangement that are real but modest, and the manager decline runs across arrangements rather than concentrating in remote teams. Treating this as a return-to-office argument is a category error that happens to be convenient for people who already wanted that argument. What is genuinely harder when distributed is the informal repair work: the two-minute corridor conversation that resolves a misunderstanding before it becomes a grievance, the read on somebody's mood that prompts a check-in. Those have to become deliberate when the team is spread out, which means they land on the calendar, which means they land on the manager. So distribution does not cause the problem. It converts a set of free interactions into scheduled ones and charges the manager for them.

It is the highest-leverage skill in the role and the least likely to be taught. A manager spends a large share of their week in rooms they are responsible for, and the difference between a room that surfaces the real problem and one that does not is almost entirely facilitation technique: how the question is framed, whether the quietest person is given a route in, whether disagreement is made cheap or expensive, whether the manager can tolerate not speaking. These are learnable and they are specific. Most leadership development treats them as personality, teaches communication in the abstract, and leaves the manager with a model rather than a move. The practical reason to invest here rather than elsewhere is compounding: a manager who runs better rooms gets better information, and better information makes every other decision they make that quarter slightly less wrong.