15 MinutesWhy you always think you have more time than you do
Ask anyone how long a task will take and they will be wrong in the same direction. Not occasionally — reliably, and by a lot. It has a name: the planning fallacy, first described by Kahneman and Tversky in 1979.
The uncomfortable part
Knowing about it does not fix it. People who have been late on every previous version of a task still underestimate the next one. The reason is that we plan from the scenario rather than the history: we imagine the task going well, because imagining it going badly requires inventing specific obstacles, and we cannot invent the ones we have not met yet.
Experience does not correct it either, because we remember the duration of work and forget everything around it — the interruption, the file that would not open, the ten minutes finding where we had left off.
What actually helps
Use the outside view
Instead of asking "how long will this take", ask "how long did the last five take". Historical data beats estimation, and it is the one correction that reliably works — which is also why it requires having the data in the first place.
Estimate in quarter-hours, not hours
"About an hour" is a guess. "Four quarter-hours" forces you to say what happens in each. Smaller units expose the parts you were quietly assuming would take no time.
Track the whole day, not the task
Most time does not vanish inside tasks; it vanishes between them. Context switching, the gap after a meeting, the thirty seconds that becomes fifteen minutes. A task timer cannot see any of that, which is why task timers make people feel efficient while the day still disappears.
The week that changes itLog every quarter-hour for one week — honestly, including the parts you would rather not write down. Nearly everyone finds the same two things: the real working total is far lower than assumed, and the biggest single category is something they would never have listed.
Why logging beats estimating
An estimate is a prediction made by someone with an incentive to be optimistic. A log is a record made by someone with nothing to gain. After a week of records you stop estimating from imagination and start estimating from evidence, which is the only known cure.
The difficulty is that logging is tedious enough that almost nobody sustains it. That is a design problem rather than a discipline problem: if answering takes five seconds and happens where you already are, it survives a busy day. If it requires opening an app and choosing a category, it does not.