The Wellness Tracker
Notes on habits and routines that I have actually kept, and several I have not.
12 March 2026
I did not notice this happening. That is the whole point of keeping the sheet, and it is also slightly humbling — I would have told you with confidence that my bedtime was unchanged since the autumn.
| Month | Median lights-out | Median wake | Time in bed |
|---|---|---|---|
| October | 22:48 | 06:35 | 7 h 47 m |
| November | 22:56 | 06:35 | 7 h 39 m |
| December | 23:11 | 06:40 | 7 h 29 m |
| January | 23:22 | 06:35 | 7 h 13 m |
| February | 23:31 | 06:34 | 7 h 03 m |
Roughly ten minutes a month, in the same direction, five months running. The wake time is nailed to the floor by a fixed start to the working day, so every minute of drift comes straight out of time in bed.
I keep a one-word note for what I was doing in the half hour before lights-out, which turns out to be the only column that explains anything.
| Last thing before bed | Share of nights, autumn | Share of nights, February |
|---|---|---|
| Reading, paper | 44% | 19% |
| Screen | 31% | 58% |
| Tidying or admin | 18% | 16% |
| Straight to bed | 7% | 7% |
The substitution is nearly one for one. Reading a book has a natural stopping point at the end of a chapter; the screen does not have one at all, and on the nights it went past midnight there was never a moment where I decided to stay up.
I am wary of the obvious conclusion here. The screen might be the cause, or it might be that the nights I am too tired to read are the nights I reach for a screen, in which case the arrow points the other way. The sheet cannot tell me which.
The book goes on the pillow in the morning. That is it — not a rule about screens, not a curfew, just an object in the way. I have tried curfews four times and kept one of them for more than a fortnight; I have had better luck with changes that require a decision at a moment when I am not tired.
Three weeks in, the median is back to 23:12, which is most of December undone and none of November. I will write it up again in May.
One person, no controls, self-reported times rounded to five minutes. Sleep is a subject where individual variation is enormous and where I have no expertise at all.