A manager walks the floor, sees heads down at desks, and feels the reassuring sense that things are under control. The feeling is real; its basis is not. What the manager has observed is attendance, posture, and the appearance of industry, none of which is the same as the work being good, or even being done. Presence is a proxy for effort, and the mandate is, at bottom, a demand that this proxy be restored. The question worth asking is why the proxy feels so necessary.
In 1985, the line did not collapse
Semco ran the definitive experiment on this instinct in 1985, when it let factory workers choose their own hours. Management's prediction was specific and confident: without a fixed schedule the assembly line would fall apart, because coordination on a factory floor surely required someone setting the clock. As Semler has told it, the collapse never came. Workers sorted themselves out, parents to mornings, single people to evenings, coordinating through the simple instruction to talk to the person next to you. The prediction of chaos said more about management's need to watch than about the workers' ability to work.
That is the pattern beneath most presence-as-proxy thinking. The belief that people must be seen to be productive is rarely tested, because the seeing is comforting and the test is threatening. If workers coordinate and produce without supervision, then a large part of what supervision was doing was reassuring the supervisor. The 1985 line held because the coordination lived in the workers, not in the manager's line of sight. What the manager lost was not control but the illusion of it.
When the proxy is all you have
The reason the proxy is so sticky is that the alternative is genuinely hard. Measuring output well requires knowing what good work looks like in each role, agreeing on it in advance, and checking it after the fact. Many managers cannot do this, and not always through fault. Gallup's global data show manager engagement itself falling, from 27% to 22% in recent years, against overall engagement of around 20%. A manager who is disengaged, overstretched, and unsure how to evaluate results will reach for the measure that requires none of those capacities. Presence asks nothing of the manager except that they look up.
This is why so many return-to-office arguments are, underneath, arguments from managerial anxiety rather than evidence. The randomized trials are instructive here too: at Trip.com, managers who had predicted that remote work would hurt productivity reversed their view by the end of the experiment, once they could see the actual results rather than an empty chair. Their initial objection was not grounded in outcomes; it was grounded in the loss of the proxy. Given real data, the objection dissolved.
Seeing is not knowing
The cost of confusing the two is that the proxy actively obscures the thing it stands for. A worker who is present but stuck looks the same as a worker who is present and productive; a worker who is absent but shipping excellent work looks, to the proxy, like a problem. Managing by presence means managing by the least informative signal available, and then feeling informed. It is the management equivalent of searching for keys under the streetlight because the light is better there.
The way out is not more surveillance but a different question. Semler's version, after he stopped counting hours, was to ask whether the pump had been delivered. It is a crude question and a better one than any headcount of the room, because it cannot be satisfied by mere attendance. A manager who can answer it does not need to see anyone at a desk. A manager who cannot answer it will not learn the answer by watching people sit. Presence tells you where a body is. It was never going to tell you what the mind attached to it produced, and no mandate can make it start.
Sources
- Bloom, Han & Liang (2024). Hybrid working from home improves retention without damaging performance (Nature). RCT: managers who predicted remote work would hurt productivity reversed their view once they saw outcomes; no measured performance loss under hybrid. https://www.nature.com/articles/s41586-024-07500-2
- Gallup (2025). State of the Global Workplace. Manager engagement fell from 27% to 22%; overall global engagement around 20%. https://www.gallup.com/workplace/659610/state-global-workplace.aspx