"Switching only takes a second"

Glancing at a message takes a second. Reading it takes ten. It seems absurd to say that this could do any harm to a morning's work. You look, you deal with it or don't, and you turn back.

The cost isn't in the glance. It's in what happens when you turn back.

Erik Altmann, Gregory Trafton and David Hambrick measured this with some precision. People worked through a task with several steps that had to be done in order, a bit like following a procedure. Every so often they were interrupted briefly before carrying on. The interruptions were very short. Those averaging under three seconds doubled the rate at which people lost their place and did the wrong step next. Interruptions averaging about four and a half seconds tripled it. Their other kinds of error didn't increase, so the interruptions hadn't made people careless in general. They'd knocked out the one thing the task depended on, which was knowing where you were.

Writing a report depends on the same thing. At any moment you're holding a thread. You know what you've just said, what comes next, and where the paragraph is going. None of that is on the screen. You're carrying it in your head, and it took you several minutes to build.

A three-second glance at a message about parking permits is enough to drop it. When you look back at the report, the words are there and the thread isn't, and so you reread from the top of the section. That's the restart. On a complex task, in my estimation, it runs to several minutes each time. I've no measurement for your work, but you can time it yourself.

There's a second cost, which the organisational researcher Sophie Leroy named attention residue. Her finding, as it's generally described, is that when you move from one task to another, part of your attention stays behind with the first, especially if you left it unfinished. You're looking at the report and some of you is still composing the reply to the supplier. So for a while after each switch you're working on the new task with less attention than you'd normally have.

And the detours are longer than they seem. Gloria Mark, Victor Gonzalez and Justin Harris shadowed twenty-four office workers and logged what they did. Work was highly fragmented. More than half of the workers' stretches of activity were interrupted. Most interrupted work did get picked up again the same day, but on average more than two other activities came in between. So when you look away to deal with one thing, it's usually two or three things before you're back.

Put the pieces together for an ordinary morning. Each switch costs a restart of some minutes and a period of divided attention, and it risks a detour through two other tasks. If the sheet of paper from chapter 1 has forty marks on it, the arithmetic isn't subtle. A large share of the day went on getting back to where you'd been.