Sorry I'm late

The 20-minute version

20 min read · the full book is 55 min

It's twenty to nine on a Tuesday morning. There's a team meeting at nine in the building across the park, and the walk takes fifteen minutes, so you have five in hand. You spend them answering an email that came in overnight. It's a short reply, but the person has asked two questions and the second needs a figure from a spreadsheet.

At ten to nine you press send and stand up. Your keys aren't in your coat. They're in yesterday's coat, upstairs. You fetch them, come down, remember the notebook. At four minutes to nine you're on the pavement, walking faster than is comfortable, and the park path is wet so you go the long way round. You take the stairs two at a time and arrive at eight minutes past, warm in the face, saying "Sorry, sorry" as you find the last chair, the one by the door. Someone is halfway through the first item. You nod as if you've followed it.

You're not a chaotic person. You do good work. But this is the third Tuesday in a row, and last Thursday you arrived at Dan and Ellie's at five to eight for half past seven, and the lasagne had been in the oven for a while.

That's the habit. Late by small amounts, nearly always, with a true reason each time and a sincere apology on arrival. You've been told you're bad with time and you've come to believe it. Privately you may also think punctual people make rather a fuss.

The habit makes sense from the inside, and I'll start there. Then I'll argue that it rests on a mistake about time that nearly everyone makes and you make more often than most. Your idea of how long things take is a best case. Experience doesn't correct it, because memory shortens the past and every late arrival gets filed as an exception. The practical part follows from that. You stop aiming at the place and start aiming at your own front door, and you set the time you leave from your own record and not from your hope.

Being on time isn't a moral quality, and punctual people aren't better people. Lateness costs you more than you've counted, most of the cost lands on people you like, and the cure is a change of arithmetic.

Three cautions. If you've had trouble organising your time all your life, along with losing things, being easily distracted and finding it hard to finish tasks, and it's affecting your work or relationships, those may be signs of ADHD, and the NHS advises seeing a GP. What follows may still help with leaving late, but it isn't a treatment for a condition.

If you've only recently become late to everything, having been reliable before, and you've also lost interest in things, sleep badly or feel low most of the day, treat that as a change in your health, and talk to a GP.

And if you're late because you're carrying more than one person can carry, small children and an elderly parent and two jobs, you don't have this habit. You have a load problem, and the answer to that is help, and a better estimate of the walk won't touch it.

One thing before reading on, and don't change anything yet. For the next week, whenever you're about to leave for something with a fixed start, write down the time you think you'll arrive. Then write down when you actually arrive. Keep the list.

Fifteen minutes on a good day

You're late in one direction only.

If your estimates were merely loose, you'd be early as often as late. You aren't. You're late by five minutes, ten, twenty, and almost never early by the same. And you've never missed a flight, an exam or the start of an interview. Keep that in mind, because it shows the ability is there.

There are good reasons behind the habit. Arriving early means sitting in a corridor doing nothing, and you have plenty to do. The last thing before you leave is usually a real thing, and getting it done feels efficient. Your estimate feels honest, because on a good day the walk does take fifteen minutes. Standing outside a restaurant at twenty-five past, wondering whether to go in, feels slightly foolish. And nothing much seems to happen when you're late. The meeting goes on, dinner is served, your friends are still your friends.

So the habit is trying to use your time fully and avoid the dead minutes of waiting. It does both badly, and the estimate is where it starts.

Fifteen minutes is the walk on a day when the keys are in the right coat, the path is dry, nobody stops you on the stairs and you walk briskly from pavement to pavement. It isn't the time from sitting at your desk to sitting in the meeting room, which includes finding things, the coat, the door, the stairs at both ends and the corridor. And it isn't an average. It's a best case.

Roger Buehler, Dale Griffin and Michael Ross named this the planning fallacy in 1994, the tendency to underestimate how long your own tasks will take. When they asked people to think aloud while predicting, nearly all the thinking was about the plan, step by step, with each step going well. Very little drew on how long similar things had taken before. I think that fits the walk exactly. When you say fifteen minutes you're running a short film of the walk, and films of walks don't include hunting for keys.

A later study by Ian Newby-Clark and colleagues asked people to imagine the pessimistic version too. It didn't change their predictions. They rated the version where things went wrong as less plausible than the one where they went right, even though three small things go wrong most days.

So the method you've been using, picturing the trip, can only produce a good-day number. Your arrivals lean late because you've been aiming at the best case.

Here's the puzzle, though. You've walked that route a hundred times. Why haven't you learned?

Buehler's team looked at that too. When people explained past misses, they blamed causes specific to that occasion and unlikely to recur, a bout of flu or a friend visiting. Each was true, and each let the person set that time aside as a special case, so the past taught nothing. When people were told to connect their predictions with past experience, the optimism went away. The information was there. It wasn't being consulted. Last Tuesday it was the keys, the week before your mother rang, the week before that the wet path. Each was a one-off, so the walk still takes fifteen minutes, and the one-offs, which happen every week, live in a separate file the estimate never opens.

Memory adds a second problem. Michael Roy, Nicholas Christenfeld and Craig McKenzie reviewed the evidence and argued that people also misremember how long things took, and remember them as shorter. The ten minutes of keys and coats didn't feel like part of the walk, so they aren't in your memory of it.

There's a useful detail in the 1994 studies. When people predicted how long someone else's task would take, they weren't optimistic, and they used the past. If a colleague said the walk took fifteen minutes and they'd leave at quarter to, you'd probably say, "Leave at twenty to, you know what you're like." You can already estimate. You just don't do it for yourself.

Now the other half of the habit, which I think does more damage. The last thing.

At twenty to nine you had five minutes, and leaving would have wasted them, so you answered an email. The last thing gets a slot exactly the size of the spare time, and then it takes longer than the slot. That's no accident. You picked it because it looked small, and small tasks are the ones whose steps you haven't pictured. Two questions, the second needing a figure, the figure in a spreadsheet. None of that was in the picture when you sat down.

Then something else happens while you're doing it. Richard Block, Peter Hancock and Dan Zakay pooled a hundred and seventeen experiments and found that when people knew they'd have to judge time and their attention was loaded with a demanding task, they judged less time to have passed than really had. You sit down for five minutes, look up feeling that five have gone, and it's ten to nine. The clock was in front of you the whole time. You were looking at the email.

The last thing also turns spare time, which feels like nothing, into a finished job, which feels like something. It puts off the small chore of coats and keys and the cold. And it has its successes. On the days the email really was two lines, you sent it, left, were on time, and it felt like skill.

Jeffrey Conte and Rick Jacobs studied a hundred and eighty-one train operators and found that those who preferred doing several things at once were somewhat more often late for work. It was a modest link, the kind you'd expect if the last thing is part of what makes people late.

So you have a best-case task, in a window already too small, under conditions that make time feel shorter than it is. That's the email you were sending at ten to nine.

Who does the waiting

On Thursday Dan and Ellie put the lasagne in at seven, expecting to serve it at eight. By ten to eight they were checking their phones and wondering whether to text you, and whether the other guests would like a drink while they waited. At five to eight you arrived and said sorry, and the evening started, and it was fine. The first half hour of their evening had been spent managing your absence.

I want to grant you something first. If you leave enough time to be sure of arriving on time on a bad day, you'll be early on the good ones. There's no way round that, and anyone who says you can be exactly on time every day is selling something. The real question is what the early minutes are worth and what the late ones cost, and I think the habit has priced both wrong.

On one side, ten minutes in a corridor or on a bench outside the dentist. On the other, the dash, the apology, the chair by the door and the first ten minutes of the meeting spent catching up. You've been treating the first as pure waste and the second as free. The ten early minutes are yours. You can read, think, look at the trees or send the email. The ten late minutes belong to nobody. You aren't working, because you're walking fast, and you aren't in the meeting, because you haven't arrived.

And an appointment is a small promise. Being early is the price of keeping it reliably, because time is uncertain and the only way to be sure of not being late is to aim early. If you refuse that price, the waiting doesn't vanish. It gets passed to someone else. Dan and Ellie waited twenty-five minutes. You didn't wait at all.

Some people say they need the pressure. With time to spare they dawdle, and the dash is what gets them moving. Buehler, Griffin and Heather MacDonald looked at people filing tax returns. Those expecting a refund, who wanted to be quick, predicted they'd finish sooner than the others, and they were more wrong. In a laboratory version, an incentive for finishing early made people's predictions more optimistic and made them look less at how long things had taken before. Wanting to be quick makes you believe you'll be quick. It doesn't make you quick. And the flight last summer had no pressure in it. You looked at when the gate closed, worked backwards, and left when the arithmetic said.

It's only five minutes, though. Look at the five minutes from the other side.

Eight people are in the room at nine. You arrive at eight past. Either they started without you and now have to go back over the first item, or seven people each lost eight minutes waiting. At the dentist, the person after you starts late, and the one after them. You don't see it, because you've gone.

Over months people quietly adjust. They tell you seven when they mean half past. They start without you as a matter of course and stop asking you to pick things up on the way. A colleague books the meeting for five past and doesn't say why. None of it is said to your face. It's a set of small discounts on your word, because your word about time has turned out to be worth less than theirs. You can be warm, funny and good at your job and still be the one everyone works around.

Notice what happens when someone's five minutes late for you. You check your phone. A small grievance starts. You're pleased when they arrive and you forgive them at once. That's exactly what happens to them.

At home the habit is quieter. "I'll be down in a minute." "I'll be home by six." "I'll just finish this and then we'll go." Each is a best-case estimate delivered as a promise to someone waiting on it, your partner in the car with the engine running or a child by the door in a coat. The minute passes and they've no idea whether it's now one more or twenty, because you don't know either. So they learn to translate. Five minutes means fifteen. "Home by six" means half past. When you say a time, they hear a mood. Children learn it early and say it plainly. "You always say a minute."

Say the real number. "Ten minutes" lets a partner make a cup of tea. "A minute" leaves them standing in the hall. If you don't know, say that. "Go on without me and I'll come when it's done" sounds worse and is much better, because it can be relied on.

And the apology. You mean it every time. But sorry for the same thing every week stops carrying information, and it asks the other person to say "Don't worry about it" for the fortieth time about something they do, a little, worry about. If it comes with a story, the keys or the phone call or the wet path, they've heard a different one each week and noticed that the outcome never changes.

So say sorry once, briefly, with no story, and change the subject. Better still is the message sent before you arrive. "Running ten minutes late, start without me" costs nothing and gives them what they most want, which is information. They can start, order, or put the lasagne in later.

Nobody who arrives at five to nine has to say anything.

Aim at the door

You've never missed a flight.

You were at the airport two hours early like everyone else, and it took no special agony. You've been on time for exams, interviews, probably your wedding. When the start can't move and missing it is serious, you're punctual. What you've been doing, without noticing, is deciding which appointments count as fixed. The flight counts. The Tuesday meeting doesn't, and Dan and Ellie don't.

That's a policy, and policies can be changed. Conte and Jacobs measured personality in their train operators, expecting conscientiousness to go with less lateness. It went with less absence and wasn't related to lateness at all in their sample. Robert Levine and Ara Norenzayan found the pace of life differed a great deal across thirty-one countries. If punctuality were built into people it wouldn't vary so much with where they live. It's learned, so it can be learned later.

Phones have made lateness cheap at the moment of leaving. "I'll text if I'm running behind" has become part of the plan, which means running behind has too. Six people meet at seven, one texts at ten to that they're late, the others ease off, and by quarter past four of the six are late. Some settings really don't mind, and a party that says eight means any time after half past. The ones that do mind are work, appointments, food in an oven, and a child waiting to be collected. There, people tolerate you because they like you, and they're spending goodwill on you every week.

"Something always comes up" is the most sincere objection. The phone rings, a child can't find a shoe, there's an accident on the ring road. I believe you about each of them. But has any of them come up for the first time this month? The individual event is unpredictable. That there'll be one is the most reliable thing about your mornings, and it's the best argument there is for a cushion. You don't need to know what it'll be. You add ten minutes for it.

Justin Kruger and Matt Evans titled their study "If you don't want to be late, enumerate". Listing a task's parts before estimating it reduces the planning fallacy, because each part brings its own small delays into view. "Leave for the meeting" is one item. "Find keys, find notebook, coat, lock door, stairs, park, wet path if raining, stairs, corridor" is nine, and each has taken time before. Daniel Kahneman and Amos Tversky described this as the difference between the inside view, picturing how it'll go, and the outside view, asking how things like it usually go. The inside view says fifteen minutes. The outside view counts your trips over the past year and says twenty-two.

So here's the method.

First, replace the estimate with a record. Take the week of predicted and actual arrivals, and you'll probably see that the actual times are later nearly every time and the gap is about the same size across quite different trips. Then time your regular trips for two weeks, from the moment you decide to leave until you're sitting where you're meant to be. Don't try to be quick, or the record will be as optimistic as the estimate. Time the things before leaving too, if you can bear it. The shower including your hair. A child into coat and shoes.

Take the longer figures, not the average. If the walk took fourteen, sixteen, twenty-one, seventeen and twenty-three minutes, the number you want is about twenty-one. That's the ordinary bad day, and ordinary bad days are ordinary. It'll feel absurdly pessimistic. Newby-Clark's finding says it will. For trips you haven't timed, list the parts, put a minute or two against each, and add a cushion of about ten minutes, nearer twenty where traffic, public transport or children are involved. Then use the number without argument. "It won't take that long today" is the picture talking.

Second, move the target from the place to the door.

  • Work back from the start. Meeting at nine, door to seat twenty-one minutes, so the door is twenty to nine. That's the only time that matters now.
  • Set an alarm for the door, and one ten minutes before it called "get ready". When the first goes, start packing the bag. When the second goes, leave. Don't finish the sentence.
  • Keys, bag, notebook and coat go in one place by the door before you sit down to anything else. In my observation the hunt for keys is the commonest last-minute delay, and a bowl by the door removes it. It's dull advice and it works.
  • In the ten minutes before the door, start nothing with a beginning, a middle and an end. No email, no call, no quick look at the news. You can't beat a best-case estimate and time that feels short by being careful. You can only beat them by not starting.
  • If you're ready early, leave early and spend the spare minutes on the other side of the door. Spare time at home turns into the last thing.
  • Never drive differently because you're late. A brisk walk saves perhaps a minute in fifteen, and in a car most of the time is lights and junctions. Pushing harder saves less than it feels like, and it puts other people on the road at risk to spare you an awkward entrance. Once the rule is fixed, the dash loses its only lever.
  • Treat the door time as fixed, the way you treat a flight. The flight leaves at eleven whether or not you've finished the email.

Sander Koole and Mascha van't Spijker found that people who stated where and when they'd do a task finished it more often, and their predictions came closer to what really happened. "I leave at twenty to nine from the front door with the bag that's already packed" is a plan. "I need to be there at nine" is a wish.

You'll arrive early, often, by five or ten minutes. Bring the book, or the email that used to be the last thing, and do it properly in the corridor with a coffee. Or sit on the bench and look at the street. At a friend's house, walk round the block. At work, the two minutes before a meeting, when the person you need is pouring a coffee, can be worth more than the meeting, and you've missed them for years because you were on the stairs.

The decision, in your own words if you like.

I aim to be at the door on time, not at the place, and I set that time from my record, not my hope.

Peter Gollwitzer and Paschal Sheeran pooled ninety-four tests of plans in the form "when X happens, I'll do Y" and found they make people more likely to follow through. Three are worth writing down.

  • When the "get ready" alarm goes and I'm mid-sentence, I'll stop and pick up the bag.
  • When I think "there's time to do one more thing", I'll take that as the sign to leave now.
  • When I realise I'll be late, I'll send a message with a time before anything else, then travel at normal speed.

Twelve minutes early

Expect the first few weeks to feel odd. The door time will seem absurdly early, and some of the spare minutes will be dull. Someone will say "You're early" in a tone you won't love.

You'll slip. You'll do the last thing and be late and know exactly why, which is useful, because it tells you which alarm you ignored. Set it again for tomorrow, and don't decide that the method doesn't work or that you're back where you started.

When you know you'll be late, send a message with a time at once, then travel normally. One short sorry on arrival, no story.

A few situations need their own word.

  • Small children make the record long. Time getting them out over two weeks like anything else, use the bad-day figure, add a bigger cushion, and get the children ready before yourself, because a child in a coat can wait and a child without shoes can't. With very small children you'll sometimes be late anyway, and a message with a time is all that's needed.
  • If you have more to do between six and nine than can be done between six and nine, a door time won't fix it. Something needs to leave the morning, or someone needs to take part of it. That conversation with a partner, employer or family is bigger than this, and worth having.
  • If a journey varies by half an hour from day to day, set the cushion for the ordinary bad day and accept that a closed road will sometimes make you late. Once a month because of a closed road isn't this habit. Every week because of the same "traffic" is.
  • When a meeting overruns into the next one, say at the start when you have to leave, and leave then.
  • If being early makes you anxious, start with two or three minutes and bring something to do.

If lateness has come with a lifetime of losing things and not finishing, go back to the first caution, because the door can still help but a GP can say what else might. And if your reliability has changed recently, along with low mood or poor sleep, that's your health. See a GP.

It's twenty to nine on a Tuesday, some months on. The record says door to seat is twenty-one minutes, the alarm went at half past, and you're on the pavement. The keys were in the bowl. The notebook went in the bag last night. The email from overnight is still in your inbox, and you'll answer it at ten to nine in the corridor, with time to check the figure properly.

The park path is wet again, and you take the long way round at a normal pace. On the far side you meet the person who chairs the meeting, who says, "Oh good, I wanted to ask you about the figures before we start," and you have four minutes to talk about it.

You're in the room at seven minutes to nine. You choose a chair, and it isn't the one by the door.

It wasn't like this at first. In the first fortnight you twice set the alarms and finished the sentence anyway. In the third week you arrived twelve minutes early and found the minutes surprisingly long. Then you started bringing a book. You've been late twice since. Once the road was closed, and once you did the last thing. Both times you sent a message the moment you knew.

Last Thursday you were at Dan and Ellie's at half past seven, having walked round the block once. Ellie said the lasagne had another twenty minutes and would you like a drink. It turned out the lasagne had always been timed for eight, and for years you'd been arriving just before the food and calling it late.

Pick your next appointment. Time the trip from door to seat, honestly, and leave by that figure and not the one in your head. Then see who's there when you arrive.