The decision

This chapter asks you to decide something. It comes late because at the beginning it would have sounded like the advice you've always been given, which is to leave earlier. By now you've seen why that advice never worked, which is that your idea of earlier was still built on a best case, and why it's right anyway. Your estimates are pictures, and pictures show things going well. Experience hasn't corrected them, because the past gets filed as a set of exceptions and stored short. The last thing always takes longer than its slot. Waiting is cheaper than the dash. And the cost of being a little late falls mostly on people you like, in ways they'll never mention.

Here's the rule I'd suggest. Put it 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.

It has five parts, and each answers a chapter.

I use the record. For regular trips I use the timed figure from an ordinary bad day, and for new ones I unpack the trip and add a cushion. When the estimate in my head says fifteen and the record says twenty-one, the record wins.

I aim at the door. I work back from the start time to the door time and I set the alarm for that. The start time stops being my target.

I don't start anything in the final ten minutes. Things are ready by the door before I do anything else. The last thing waits until the other side of the door.

I treat the door time as fixed. The same way I treat a flight.

I don't make it up on the way. I never drive differently because I'm late. If I'm going to be late, I say so before I arrive, with a time, and I say sorry once and briefly when I get there.

It helps to plan for the moments when the habit will try to restart. Peter Gollwitzer and Paschal Sheeran pooled ninety-four tests of plans in the form "when X happens, I'll do Y" and found that they make people more likely to follow through. A plan like this works because you've decided what to do before the moment comes. Three are worth writing down.

When the "get ready" alarm goes and I'm in the middle of something, I'll stop mid-sentence, put it down and pick up the bag.

When I catch the thought "there's time to do one more thing," I'll take that as the sign to leave now.

When I notice I'm going to be late, I'll send a message with a time before I do anything else, and then travel at the normal speed.

I'll be honest about what to expect, and this is my expectation and not a finding. For the first few weeks the record will feel pessimistic and the door time will feel absurdly early. You'll arrive early more often than you'd like, and some of those early minutes will be dull. You'll slip, and do the last thing, and be late, and that will be annoying because you'll know exactly why. Some people will notice you're on time and make a remark, and it'll be friendly, and you'll mind it slightly.

Against that, you should be on time to most things within a fortnight, because the arithmetic works from the first day you use it. The dash should mostly disappear from your week. You'll start to see the two minutes before things, and the people in them. The apology will become rare, and so it will mean something again. And somewhere around the second month, someone who's waited for you for years will notice that they haven't had to, and won't say anything, and you'll know.