The manager's shattered day, according to Mintzberg

A food manufacturer, 450 people. The plant manager blocked Thursday morning to review the half-year capacity plan. At 8:12 a question came in about a held invoice. At 8:30 a supervisor walked in about an absence. At 9:05, a call from sales about an urgent order. By 11:40 she had read two pages of the plan.

At the end of the day the feeling was the usual one: I couldn't get any work done. Except she worked all day — she made fourteen decisions, unblocked three things and prevented a customer problem. What didn't match was the expectation of how the day should have gone.

01The idea: what the stopwatch showed #

Henry Mintzberg built his doctoral thesis on shadowing five chief executives, timing every activity. The result became The Nature of Managerial Work in 1973, and later the article The Manager's Job: Folklore and Fact, published in Harvard Business Review and reissued in 1990.

What he found flatly contradicted the classic description of the manager as someone who plans, organizes, coordinates and controls in a systematic way. In the field, the work was marked by brevity, variety and fragmentation: many short activities, frequently interrupted, with few long stretches of uninterrupted reflection.

Mintzberg also observed that these managers preferred verbal, current and informal information — a conversation, a phone call, a note — over consolidated reports. And that much of what they did was processing information to feed decisions only they could make, because only they held the whole context.

02Why this isn't a flaw to fix #

The naive reading of the research is 'the manager needs more discipline'. Mintzberg argues the opposite: fragmentation is a rational response to the job. Whoever coordinates has to stay available, because the decision blocking ten people is worth more than an uninterrupted hour of analysis.

What is a failure is the design around it. If every interruption forces a person to rebuild context from scratch — open three systems, hunt for the latest file version, ask someone where it stood — the cost per interruption multiplies. Gloria Mark and colleagues measured that return to interrupted work in 2008 and showed it isn't free: people do come back, but slower and under more pressure.

In 1973 interruptions arrived through the door and the phone. In 2026 they arrive through four channels at once, and the manager of a mid-size company coordinates teams using different tools. The fragmentation is the same; the cost of each fragment went up.

There's also a scale effect mid-size companies know well. In a 50-person organization, the manager keeps the state of everything in their head, and fragmentation is cheap: they already know where each topic stood. Past roughly 200 people, that no longer fits in anyone's memory — but the coordination habits stay the same, and what used to be efficient turns into a sequence of questions around the table.

Mintzberg returned to the theme in Managing (2009) with an observation that aged well: no amount of communication tooling removes the manager's need to exercise judgment on concrete cases. What tooling can do is lower the effort of reaching the case with the right information in hand — and that, not personal discipline, is where the real gain lives.

03The math of rebuilding context #

At the scale of that 450-person plant, a mid-size Brazilian company (figures in Brazilian reais), with 30 people in coordinating roles. Conservative assumptions, stated so you can redo them:

Want to see how this looks inside a real operation? Explore the platform.

The math: 22 switches × 4 minutes = 88 minutes a day; 30 managers × 1.47 h × 220 working days ≈ 9,700 hours a year, at a loaded hourly cost of R$ 75 that's roughly R$ 730 thousand. Cutting those 4 minutes to a minute and a half — a question of where information lives, not of personal discipline — gives back about R$ 450 thousand of managerial capacity a year.

04Where this breaks in practice #

1. The state of things lives in people's heads. When the only way to know where something stands is to ask, every answer costs two interruptions instead of one.
2. Information is filed by tool, not by topic. The manager has to remember which system that topic lives in before they can even look.
3. The status meeting becomes the context mechanism. Since nobody can see progress on their own, a weekly ritual is created to say out loud what a record would show in ten seconds.
4. Calendar blocking is treated as the fix. It helps analytical work, but it doesn't touch the root problem: the fourteen decisions are still needed and will pile up in the afternoon.

05What data-driven management answers #

If a manager's day is made of a few minutes at a time, the system they work in has to deliver context in seconds: what's open, with whom, since when, what's waiting on their decision. That isn't analytical sophistication — it's recording the work where the work happens.

Two readings connect directly. The more open fronts, the more context switches per day — the arithmetic is in why more projects deliver less. And much of that interruption is, in economic terms, coordination cost: the piece on Coase covers why it grows with nobody measuring it.

The usual caveat holds too: visibility of work exists to give context, not to track anyone's every minute. The difference between the two is a choice about the theory you hold about people. In the platform, this is what the tasks, projects and operations modules solve, with the per-team cut described in solutions.

The plant manager didn't have a bad day. She had the day the job has — in a system that was never designed for it.