What your focus sessions actually add up to

Start with the hours. Most people have two or three hours in the day when their sessions actually finish, and those hours are usually earlier or later than they assume. Once you can see yours, defend them: no meetings, no errands, nothing that can be moved to a weaker part of the day.

Four to five focus sessions is a strong day, not a weak one. Ericsson's studies of expert performers found that even violinists at the top of their conservatory rarely sustained more than about four hours of deliberate practice a day, spread across the morning and afternoon with rest between. If your daily totals sit near that, the constraint is real rather than a discipline problem.

The estimated versus actual ratio is the most useful line on this page. Treat it as a personal multiplier. If your completed tasks came in at 150 percent of estimate, a task that feels like four sessions is a six-session task, and planning it as six is what stops the week from sliding.

Watch the 14-day chart for shape rather than height. Long flat stretches broken by one tall day usually mean the work is being pushed to deadlines, which costs more than a steady three sessions a day would.

A session is filed under the hour it ended, in this browser's local time, so a twenty five minute block begun at 16:50 counts towards 17:00. Only hours that contain at least one session appear, up to twelve of them, ranked by how many sessions ended there and then by minutes. The weekday panel is the opposite view: every day is listed whether or not you worked, and it totals the whole log rather than a recent window.

The four figures at the top are narrower than the charts. Today means the current calendar date on this device, and the week runs from Monday to Sunday around it, which is why the weekly totals fall back to zero on a Monday morning rather than rolling over the last seven days. Tasks done this week counts the ones whose completion date lands in that same window.

Time by project reads the project name off the task each session was linked to and takes the eight projects with the most minutes. Sessions with no linked task, or a task with no project name, are filed under No project. A project outside the top eight is left off the chart, though its minutes still count in the totals at the top. The fourteen day strip keeps empty days in place on purpose: the gaps are the shape worth looking at, and the list under the chart gives each date with its exact total.

The estimated versus actual table only considers tasks that are finished and carry an estimate above zero, takes the twelve with the most logged sessions, and works out the percentage from the totals of those twelve alone. A task you estimated and abandoned is invisible here, and so is one you finished without ever setting an estimate, which flatters the ratio in both directions.

Sessions counted against a task are the ones you ran with that task selected on the timer. Work you did on it at a whiteboard, in a meeting or on a laptop with the timer closed does not appear, so the actual column is a floor rather than a true cost. Read the ratio as the drift between what you planned to sit down and do and what you actually sat down and did.

Give it a dozen finished tasks before you trust it. Two tasks that ran long will push a small sample well past a hundred percent, and a single tiny task that came in early will pull it back, neither of which tells you anything about the next estimate you write.