The Cost of Context Switching Is Not What You Think
The standard argument against context switching goes like this: every time you jump between tasks, your brain needs time to reload the previous context. Studies cite 20-minute recovery windows. The implication is that fragmented schedules are inefficient, and deep work blocks are the fix.
This is true but incomplete. The more damaging cost of context switching is not the reload time. It is what happens to the task you interrupted.
When you leave a problem mid-thought — pulled by a Slack ping, a meeting, a reflexive tab-switch — the working memory state you had built up dissipates. That state is often the most valuable thing you produced all day. It is not written down. It cannot be reconstructed from your notes. It was a specific configuration of held variables, half-formed connections, and active tension that the mind was about to resolve. When it’s gone, it’s gone. You don’t resume the task; you restart it.
This is why the 20-minute figure understates the damage. You can reload facts in 20 minutes. You cannot reload a near-insight.
The practical implication is not just “block your calendar.” It is to treat interruptions as carrying asymmetric cost: cheap to absorb in the short term, expensive to absorb at the specific moment when your thinking is actually getting somewhere. The dangerous interruptions are not the ones during shallow work. They are the ones that arrive exactly when you are about to figure something out.
Which is, of course, when the phone always rings.