The line at 7:40 p.m.
It is 7:40 on a Tuesday morning and the owner gets in before his assistant. A metalworking plant in Pilar, 380 people, one customer who accounts for 22% of revenue. He puts down the coffee and opens his phone: sixty-one unread messages since eleven the night before. The first is from the shipping supervisor and it comes with a photo: a loaded truck that never left.
By 7:52 he has called him. At 8:20 he is on with the big customer, who rings about a quality complaint and ends up talking about the truck. At 9:15 a salesperson walks in without knocking: a price needs approving, it is for today. The owner looks at the margin for twelve seconds and signs.
It is a quarter past nine and he has settled three things. None of them was on the sheet he wrote himself on Sunday night: the costing model, the interview with the candidate for plant manager, the plan for the second line.
The rest of the day looks like the beginning of it. At 7:40 p.m., driving out on the Panamericana, he says the line. He says it tired, not angry, almost the way you would report the weather:
"I got none of my own work done today. I spent the whole day putting out fires."
You have probably said it yourself this month. It is also one of the very few management claims that can be disproved with a stopwatch. Somebody went to the trouble.
Forty-four years resting on five people
It is worth starting with some humility. Until recently, almost everything management believed it knew about the day of a chief executive rested on five people: in 1973 Henry Mintzberg sat beside five general managers and wrote down what they did while they were doing it. His book replaced the executive who plans from a quiet desk with someone fragmented, interrupted, working by talking; and adding the two classic samples that came afterwards, the empirical base did not reach fifty people in forty-four years.
In 2017, and in its definitive version in the Journal of Political Economy in 2020, Oriana Bandiera, Andrea Prat, Stephen Hansen and Raffaella Sadun did what was missing: they logged the diaries of 1,114 CEOs across six countries —Brazil, France, Germany, India, the United Kingdom and the United States.
It is worth saying who these people are, because that is the first objection anyone running a mid-sized company raises: they are CEOs of manufacturers, mostly large ones, many with a CFO, a board and even a COO to share the load. You have none of that: you are the CFO, the sales lead and the referee. That should push emergencies up in a three-hundred-person company, not down —and even so, it is better measured than assumed.
The procedure is dull to describe, which is exactly why it matters. Two calls a day: in the morning the planned agenda was recorded; in the evening, what had actually happened, activity by activity, in fifteen-minute blocks. Only activities longer than fifteen minutes were counted: WhatsApp and the corridor are not in here.
The hinge is a single question: for every activity they asked whether it was a response to an emergency. The label was applied by the CEO himself —in 43% of cases— or by his assistant: it is still self-reporting. That is precisely the point. Even by his own standard, on the same day, before memory had edited anything, only a sliver of the day struck him as an emergency. The gap is not in the world: it is in the recollection.
Four percent
Four percent. That was the share of the time of 1,114 CEOs spent on activities defined as emergencies.
The number, translated into minutes: over a week of sixty-two and a half hours —which is the week the people who measure these calendars actually measure— 4% comes to two and a half hours. Less than half an hour a day. Half a management meeting. And it is half an hour against the time that got logged —activities longer than fifteen minutes— not against the whole day.
And there is a second number in the same research that almost nobody quotes and that lands harder: comparing the morning's planned agenda against the evening's executed one, less than 10% of what was planned got cancelled. The chief executive's calendar does not fall apart on him: it goes ahead as written.
The other measurement in the genre points the same way: 27 CEOs of large companies tracked around the clock for 13 weeks, close to 60,000 hours of data (Porter and Nohria, Harvard Business Review, 2018).
The 4% does not say that emergencies do not exist, or that the truck in Pilar was a fantasy: we already wrote about the chronic firefighter in Problems That Don't Come Back, where urgency mode was real because the lens was the operation. With the lens on a single calendar, the number says something harder to dodge: the source of the stress is the selection of the agenda, not the randomness of the world. The Tuesday really was exhausting; what is not true is that it was imposed on him.
What that measurement left out still needs saying, because it is the first thing on the reader's tongue: only what lasted more than fifteen minutes was counted. The two-minute phone call and the salesman putting his head round the door are not inside the 4%. What comes next measured exactly that, which is why the two numbers do not contradict each other: the 4% is what was an emergency; what follows is what feels like one.
So why does it feel like a fire?
If emergencies take up a sliver of the day, one question remains that no percentage answers: why is the feeling of that Tuesday so vivid? The owner in Pilar is not lying in the car. Some machinery is producing that feeling out of materials that are not emergencies.
The best documentation of that machinery does not come from a study of CEOs. Leslie Perlow spent nine months inside a team of software engineers and published the results in Administrative Science Quarterly under a title that is already a diagnosis: The Time Famine. Across 35 diaries kept by 12 engineers she measured every exchange with another person and asked them to classify it themselves. 95% happened without warning: interruptions.
96% was judged useful by the person on the receiving end.
Of that useful 96%, barely 10 points were urgent. The other 86 could have been scheduled for later with no consequence for anyone.
They are engineers, not CEOs: what travels from one case to the other is the mechanism, not the percentage. And it travels well. The mechanism has two pieces. The first is what Perlow calls a crisis mentality: in urgency mode, the distinction between the important and the immediate stops paying off and gets abandoned. The second explains why the pattern does not correct itself: a reward system built on individual heroics. Solving pays on the spot; preventing never pays, not even when it works.
We call that the epic of the firefighter. The best definition is not ours: an owner gave it, looking at the chart of his own week.
"Signing off a price in twelve seconds leaves me feeling like I run the company. A half-finished costing model leaves me nothing. Guess which one I do first."
The cost is not in the emergency sliver: it is in the shape left to the rest of the day. On Perlow's team, three out of every four blocks of uninterrupted individual work lasted an hour or less: there was no shortage of time alone, there was a shortage of unbroken time alone. And the chief executive's own work —assessing a person seriously, understanding why a margin moved two points— needs a long continuous run. A costing model does not advance in four fifteen-minute bursts separated by a truck, a complaint and a price: it starts over four times.
A calendar cut into scraps is not a busy calendar: it is a calendar without the one task that cannot be delegated.
The three jobs nobody ever wrote down
At this point —we have this conversation often, with the chart up on the screen— the same objection comes up, and it comes up well founded:
"All of that happens in Germany. Here the dollar moves on Thursday, the supplier does not deliver, the tax authority issues a ruling, the union stops the night shift."
Argentina's external noise is real, and this piece does not dispute it. What it disputes is that the noise explains the shape of your Tuesday. The two largest subsamples in the study are India, with 356 CEOs, and Brazil, with 282 —two economies that know a thing or two about devaluations. No country-level breakdown of emergencies is published, but the average holds up with Brazil and India inside it, not despite them.
What does explain the Tuesday is more specific and a good deal less heroic. In the mid-sized Argentine family firm, the person at the top is usually three things at once: the last stop for price approvals, the informal account manager for the three big customers and the referee of every conflict between managers. Each of those three hats generates, every single day, events that feel exactly like emergencies: they come with a rush, with someone waiting, and with a consequence for delay.
They are not fires. They are three jobs nobody ever wrote down and that you hold on top of your own. Nobody hijacked the rest of your day: you chose it, you scheduled it and you defended it. The difference matters: a fire cannot be planned and a job can. It gets defined, it gets staffed, it gets a limit and it gets handed over.
It is worth going slowly through the costliest one, the price signature, because handing it over is not saying "from now on Martín signs". It is four lines on a single page. The limit: how much discount off list, for which product family and at what minimum margin the sales manager signs without asking anyone. A number, not a criterion; and if you do not know which number, look at the last twenty prices you signed yourself and find the point where you stopped hesitating. The exception: what forces a decision upwards —a new customer, payment terms outside the standard, anything below the minimum margin— and through which channel, with an answer committed to in hours and not "whenever I run into him in the corridor". The review: a fixed fifteen minutes a week, always the same day, over the complete list of signed prices and their realized margin, not over the three cases that caught your eye. And the day he gets it wrong, which is the line that decides whether the system survives: if you take the signature back, you did not delegate, you lent. What works is looking at it in the following week's review, correcting the limit if the limit was the error, and leaving the signature where it is. A manager who loses it on his first mistake learns exactly one thing, and it is not how to set prices.
The other two are handed over the same way, each with its own four lines: the big-accounts job needs a contact plan and a face-to-face introduction of the replacement; the referee job needs both sides to know in advance who decides what.
On meetings, even the numbers are smoke
At this point the reader will ask about meetings. Before going in, a warning: this is the patch of management with the worst data hygiene we have ever traced, and we apply to it the same discipline we applied to that other celebrated figure with no known parent, the 70% of change programmes that fail.
Eleven million meetings nobody ever counted. The most quoted figure in the world about meetings —"American workers attend some 11 million meetings a day"— reached serious conversation through Rogelberg, Scott and Kello (MIT Sloan Management Review, 2007), and from there it dropped into the decks. Now pull the thread backwards, which is what almost nobody does. The number was already in print in 1976, in How to Make Meetings Work!, by Michael Doyle and David Strauss: "Do you have any idea how many meetings are held each day in the United States alone? Over 11 million!". No footnote. No source. The most complete reconstruction of that genealogy —by Elise Keith, of Lucid Meetings— tracks the figure across more than fifty publications and ends where it had to end: the trail never reaches an original. And there is a newborn zombie: the "meeting-free days" study (MIT SMR, 2022, 76 companies) circulates as "+71% productivity", but the original table is headed percentage change in employee ratings — pulse surveys, not measured productivity. None of these figures appears here as data. They appear as an object: so that the next time someone puts them on the screen in your boardroom, you ask where they came from.
The figure that does hold up is this one. Perlow, Constance Hadley and Eunice Eun surveyed nearly 200 senior executives and asked them something finer than "do you have a lot of meetings?": whether their meetings were a good use of group time and of individual time, which are two different questions. 54% —the majority— answered that they waste both.
The easy reading of that 54% is "we should have fewer meetings". The correct one is that a bad meeting does double damage: it consumes the hour of everyone inside it and it chops up each of their calendars. The visible cost is the hour; the expensive cost is the block that no longer exists anywhere else in the day.
The proof that this comes apart
What is still missing is the evidence that this comes apart, which is what makes it a management problem and not a life sentence of character. Perlow installed quiet time —stretches of the day protected from interruption—: with it, the share of engineers rating their own productivity above average went from 41% to 65%, which is perception, and that has to be said. The hard number is that the team's product shipped on time, the second on-time launch in that division's history.
There is a second case, closer to home for this reader: a financial consultancy redesigned its approach to meetings and three months later was measuring large improvements, all of them perceptual. What produced them was not having fewer meetings: it was changing their shape. That is what Part 2 is about.
The four lines that defend the Tuesday
When the measurement goes up on the screen, the reaction is not silence: it is four defences —the day is not mine, the fires are real, the interruptions are important, it is different here— and every one of them is a partial truth used as a total conclusion.
| The line people say | What was measured | What still stands |
|---|---|---|
| "I don't run my calendar, my calendar runs me" | 4% of the time on emergencies; less than 10% of what is planned gets cancelled (Bandiera et al., 1,114 CEOs) | The day is long and exhausting. The authorship of the day is still yours. |
| "I spend my life putting out fires" | Crises at around 1% of the time, though 89% of CEOs devoted some time to them (Porter and Nohria, 27 CEOs, 13 weeks) | The fires exist and they do land on you. What they do not account for is the rest of the calendar. |
| "I get interrupted all day, and it is all important" | 96% judged useful; only 10% urgent; 86% postponable with no consequence (Perlow, ASQ 1999) | They are important, which is exactly why nobody stops them. None of them earns the right to happen now. |
| "It is different here: this is Argentina" | India contributes 356 CEOs and Brazil 282 of the 1,114; the average is computed with both economies inside it (no country breakdown published) | The external noise is real. The shape of your Tuesday has another explanation. |
The line, the measurement, and what still stands.
Questions for Monday
The first five can be answered today, from memory, without pulling a single piece of data. That is the point: if getting started requires building a dashboard, nobody gets started.
Five from memory, two with a stopwatch
- What was the longest block of uninterrupted work in your week —in minutes, not in intention? Which pending decision needs one longer than that?
- The list you wrote yourself on Sunday night: how much of it was still uncrossed on Friday? If the cancellation rate is high, the question is not what happened: it is who cancelled.
- Of your three hats —price signature, big accounts, referee between managers—: which one could stop being yours within ninety days, and what would have to exist on paper for that not to be reckless?
- Who gets the applause in your management meeting: the person who got the truck out at 10:40 p.m., or the person who made sure it never stopped? When did you last promote someone of the second kind?
- Of what interrupted you yesterday: what was useful and what was urgent? Of the useful and not urgent, how much of it could have waited until Thursday at 3 p.m.?
The two that follow are the long version, for whoever is game. The method has already been run over 1,114 cases and fits into two calls a day for five days: in the morning your assistant writes down what you say you are going to do; in the evening, what happened, activity by activity. No reconstructing from memory: recollection edits, and always in the same direction.
- On the fifth day, add it up: what percentage of your week ended up on paper, and what percentage slipped away without anyone being able to name it?
- Apply the label, activity by activity, with the researchers' own binary question: was this a response to an emergency? No "sort of". What is your number?
If it comes out at 4%, the good news is that your calendar is yours. If it comes out at 30%, either your company really is in crisis —in which case this is not the article you need— or your definition of an emergency is wider than the one those 1,114 used on the day they were asked.
The Tuesday in Pilar ended with a line that, after the measurement, has a less comfortable and far more useful translation: today I did, almost all day long, exactly what I decided to do. It is worse to hear and much better to own: the first version admits no action, and this one is a work agenda. That leaves the question this article does not answer: if you are the one choosing the day, what should you be choosing? Local instinct has its answer ready —fewer meetings, more shop floor, closer to the machines— and the same 1,114 diaries point somewhere else: one standard deviation of the behaviour index the authors build goes with 7% higher sales, and the end of that index associated with higher sales loads onto meetings that are planned, long and attended by several functions at once. Part 2 goes there.
Does your Tuesday look like the one in this piece?
Ninety days in, the owner's week looks different in three verifiable ways: the price signature has stopped being his and there is a written limit the sales manager uses every day; the referee between managers has a replacement and rules known in advance; and there is, every week, a long block on the calendar that survives Tuesdays. That is the outcome 32sur works towards, and it does not start with a workshop or a focus app: it starts with a reading of your real week —two working sessions, the first with your calendar for the last fifteen days on the table— to decide which of the three jobs to hand over first. If your Tuesday looks like the one in this piece, let's talk.
References
- Bandiera, O., Prat, A., Hansen, S. and Sadun, R., "CEO Behavior and Firm Performance", Journal of Political Economy, 128(4), 2020, pp. 1325-1369 — the backbone of this article: 1,114 manufacturing CEOs across six countries, 42,233 activities recorded in 225,721 fifteen-minute blocks; only 4% of the time on activities defined as emergencies, and less than 10% of what was planned cancelled. The survey recorded activities longer than fifteen minutes, was run through a daily call to the CEO (43% of cases) or to his assistant, and publishes no country-level breakdown of emergencies. India contributes 356 CEOs and Brazil 282. It is also the source of the behaviour index that closes this article: one standard deviation of the index is associated with 7% higher sales, controlling for labour, capital and the usual firm-level variables.
- Bandiera, O., Hansen, S., Prat, A. and Sadun, R., "CEO Behavior and Firm Performance", NBER Working Paper 23248, 2017 — the working-paper version of the above, freely available, with the same figures; useful for the reader who wants the original data without a paywall.
- Porter, M. E. and Nohria, N., "How CEOs Manage Time", Harvard Business Review, 96(4), July-August 2018 — 27 CEOs of large companies tracked 24/7 for 13 weeks (close to 60,000 hours): 62.5 hours a week on average, work on 79% of weekends and on 70% of vacation days; crises reported at around 1% of the time, with 89% of CEOs devoting some time to them. The original article is paywalled: the figures are quoted as reproduced by independent coverage that agrees with itself, not from a direct reading of the full text.
- Perlow, L. A., "The Time Famine: Toward a Sociology of Work Time", Administrative Science Quarterly, 44(1), 1999, pp. 57-81 — the mechanism behind the illusion: across the interactive activities of 12 engineers (35 diaries), 96% judged useful, 10% urgent, 86% postponable and 95% occurring without warning; 75% of blocks of uninterrupted individual work lasted an hour or less; the crisis mentality and the reward system built on individual heroics; and the quiet time experiment (self-assessed above-average productivity: 59% in the first phase with quiet time, 41% without it, 65% in the third), with the PEARL launch delivered on time.
- Perlow, L. A., Hadley, C. N. and Eun, E., "Stop the Meeting Madness", Harvard Business Review, July-August 2017 — survey of nearly 200 senior executives: only 17% consider their meetings a good use of both group and individual time, and 54% say they waste both; plus the meeting redesign at a financial and regulatory consultancy, with perceptual improvements at three months (+42% team collaboration, +32% psychological safety, +28% performance, and satisfaction with work-life balance rising from 62% to 92%).
- Rogelberg, S. G., Scott, C. and Kello, J., "The Science and Fiction of Meetings", MIT Sloan Management Review, 48(2), 2007 — the article that returned the world's most quoted meeting figures to academic circulation: the "11 million meetings a day", the "nearly 23 hours a week" of senior managers, and the doubling of the average executive's meetings between the sixties and the eighties.
- Doyle, M. and Strauss, D., How to Make Meetings Work!, 1976, p. 4 — the oldest known appearance in print of the "over 11 million meetings a day", with no source at the foot: the point where the trail goes cold.
- Keith, E., "55 Million: A Fresh Look at the Number, Effectiveness, and Cost of Meetings in the U.S.", The Lucid Meetings Blog, 2015 (updated since) — the reconstruction of the lineage of the 11 million across more than fifty publications, with no original found; and the alternative calculation, built from the bottom up with employment and occupation data from the Bureau of Labor Statistics: between 36 and 56 million meetings a day, USD 1.4 trillion in total annual cost of meeting time, and estimated waste of between USD 70 and 283 billion.
- Laker, B., Pereira, V., Budhwar, P. and Malik, A., "The Surprising Impact of Meeting-Free Days", MIT Sloan Management Review, 18 January 2022 — 76 companies with more than 1,000 employees and operations in more than 50 countries; the table is headed "Percentage Change in Employee Ratings After Introduction of Meeting-Free Days", that is, changes in pulse surveys, not in measured productivity. A case study in how a zombie number is born.
- Mintzberg, H., The Nature of Managerial Work, Harper & Row, 1973 — five CEOs observed live: for forty-four years, the core of the empirical base of what management believed it knew about a manager's day. The other two classic samples in the genre are Kotter (fifteen general managers) and Porter and Nohria's work with twenty-seven CEOs.
- Rogelberg, S. G., Kreamer, L. M. and Gray, J., "Thirty Years of Meeting Science: Lessons Learned and the Road Ahead", Annual Review of Organizational Psychology and Organizational Behavior, 13, 2026, pp. 415-442 — the current academic review of the meetings field, recommended in place of brochure figures for anyone who wants to take the topic seriously.