8 tools · field notebook
Nine homes · forty-five days · 225 sittings
What ends a sitting
Nine people working from home logged every stretch they spent at a place of work: where it was, how long it lasted, and what ended it. Most of the time nothing in the room ended it at all.
50notes on where the work actually happened
6parts, from what a sitting is to what a household can change
225sittings logged across nine homes and forty-five working days
131of them ended because the work ended, not because the place stopped working
Nothing here is a survey and nothing is a benchmark. Nine households kept a log for one working week each, and every figure on this site comes from those logs. Nobody is identified — not a person, not a city, not an occupation. The record says nothing about health, comfort or how anybody felt: it says how long people stayed and what moved them, which is a narrower question and the only one a log of this kind can answer.
Six parts, from what a sitting is to what a household can change
Every note belongs to one of them. The line at the foot of each card is what that part came to on the record counted here, stated as a count rather than as a conclusion.
A sitting
What a sitting is
A stretch at one place, ended by something. Why the record counts stretches rather than hours, and what that choice leaves out.
225 sittings, median 48 minutes
Endings
What ended them
Nine kinds of ending, counted the same way, against the far larger number of sittings that simply ran out of work.
94 endings, 131 bare
The places
The thirty-one places
Nine homes held thirty-one places that were worked at. Seven had somewhere set up for the purpose; in three of those it was the place people stayed longest.
31 places, 3 of 7 desks
The hour
The same place, different hour
A window seat lasts sixty-two minutes before noon and thirty-one after. The place did not change.
62 against 31 minutes
What is fixed
What is fixed
Orientation, glazing, heating and the position of sockets were settled before anybody worked there, and they account for most of the endings.
42 of 94 endings
What to change
What to change
What nine logs can and cannot support, including the advice this record gives no grounds for at all.
Three things, and four refusals
Nine homes, counted three ways
Three panels built from the same nine weeks: a sample of days drawn sitting by sitting along the working day; every ending counted; and the length of a sitting by the hour it began. Nothing here is a survey and nothing is a benchmark. Nine people kept a log for a week each, nobody is identified, and the only claim made for the record is that it was kept.
Five of the forty-five days, sitting by sitting, with the mark at the end of each bar saying what ended it
Light 3Temperature 2Wanted 2Noise 1Surface 1Elsewhere 1Seat 1Power 1Unsaid 1
Twenty-seven sittings across five days. Median of the days drawn here 55 minutes. Fourteen of them end bare, because the work ended rather than the place.
One line was outside the 07:00–19:00 window the axis covers. One line names an ending the panel has no mark for. One line overlaps a sitting already drawn for that day, and nobody sits in two places at once.
Five days of the forty-five, chosen to show the range of lengths rather than the typical day. Three lines the panel refuses are described beneath it.
Every ending in the record, counted the same way
Nine named endings covering ninety-four sittings. One hundred and thirty-one more ended with the work rather than with the place, which is two hundred and twenty-five in all. The largest single ending is light, at 24.
The longest bar is not an ending at all. It is the count of sittings that finished because the work finished, and it is larger than every genuine ending put together.
How long a sitting lasted, by the hour it began
Six bands of the working day, built on two hundred and twenty-five sittings. Longest in 07–09 at 62 minutes, shortest in 13–15 at 31. A difference of 31 minutes across the same places and the same people.
Same homes, same places, same people. The only thing that differs between the first band and the fourth is the hour.
Two hundred and twenty-five stretches of work, logged by nine people in nine homes over one working week each. Where each one happened, how long it lasted, and what ended it.
The question the logs were built around is narrow. Not whether working from home is good, not how anybody felt about it, and not what they should have bought. Just: how long did somewhere hold, and what made the person get up.
The answer that was not expected
Most of the time, nothing did. A record of rooms and interruptions becomes more useful when it is paired with actual project hours, so a team can review the platform in context before defining a limited, transparent pilot.
A hundred and thirty-one of the two hundred and twenty-five sittings ended because the work ended — a meeting was due, a task finished, it was lunch, it was the end of the day. The room had nothing to do with it. Those are the bars on this page that simply stop, with no mark at the end. For broader research beyond this small household record, consult work published by The Royal Society.
Ninety-four ended because the place stopped working, and that is the smaller number. It is smaller than the count of sittings that ended by themselves, and it is the number this site spends most of its pages on, because it is the one that has structure in it.
What was ending them
The sun, mostly. Twenty-four endings are the light changing — going, or arriving as glare on a surface. Eighteen are temperature. Together those two are forty-two of the ninety-four, and both are consequences of how a building faces and how it holds heat, settled before anybody in these logs worked there.
Fourteen are somebody else wanting the place, and thirteen of those fourteen happened at a table the household also eats at.
The seat accounts for six.
Six, across forty-five person-days of logging in nine dwellings. The object that most of the discussion about working from home is about was what made somebody get up six times. That figure is on this page rather than buried further in, because a record that puts its inconvenient number on page nine is arranging its own evidence.
The largest thing in the data is the clock
A sitting that begins between seven and nine runs a median of sixty-two minutes. One that begins between one and three runs thirty-one.
Exactly twice, in all nine dwellings, by ratios between 1.4 and 2.2. Nothing else measured here — the type of dwelling, the type of place, whether there was a desk — produces a difference of anything like that size.
What follows from it is that a place is not a thing with a single answer attached. The same window seat held somebody for sixty-eight minutes in the morning and thirty-three in the afternoon, five days running, with no furniture having moved. Any assessment of somewhere to work that does not say when is describing something that does not exist in the form described.
What this record will not tell you
Four things, and the list is longer than the findings.
It says nothing about health, comfort or how anybody felt. There is no field for it anywhere, and a log of durations cannot be made to produce one.
It makes no recommendation about furniture. A log of endings sees objects that make people move and is blind to objects that make people stay, and six seat endings is consistent with seating mattering a great deal.
It does not show that having more places helps. That is the most intuitive thing a study like this could conclude, and the two-place dwellings in this record produced both the highest and the lowest rate of failure in it.
And it does not apply to anybody else. Nine dwellings, chosen by no procedure that would make them a sample of anything.
Why publish it anyway
Because nine logs kept properly are worth more than a survey of nine hundred people remembering their week, and because the method is published in full so that somebody with a better sheet can go and find out whether any of this holds.
The three fixes that would improve a repeat are named on the method page. The questions this record cannot close are listed in the order somebody would attack them.
What these nine households produced is a small, narrow, internally consistent account of where work happened in their homes, and an answer to a question that turned out to be different from the one everybody asks.
01 / 06225 sittings, median 48 minutes
What a sitting is, and why it is the unit
A stretch at one place, ended by something. Why the record counts stretches rather than hours, and what that choice leaves out.
02 / 0694 endings, 131 bare
The nine endings, and the dash
Nine kinds of ending, counted the same way, against the far larger number of sittings that simply ran out of work.
03 / 0631 places, 3 of 7 desks
Where the work happened
Nine homes held thirty-one places that were worked at. Seven had somewhere set up for the purpose; in three of those it was the place people stayed longest.
04 / 0662 against 31 minutes
What the clock does
A window seat lasts sixty-two minutes before noon and thirty-one after. The place did not change.
05 / 0642 of 94 endings
The constraints nobody chose
Orientation, glazing, heating and the position of sockets were settled before anybody worked there, and they account for most of the endings.
06 / 06Three things, and four refusals
What follows, and what does not
What nine logs can and cannot support, including the advice this record gives no grounds for at all.
Practical tool guides for work at home
Three different comparison formats for time, collaboration and the wider work-from-home stack.
12 platforms · comparison board
Remote collaboration platforms
Tasks, documents, meetings and handoffs.16 tools · workspace lab
A measurable home-work stack
Operational evidence across a distributed workflow.The whole record
Two hundred and twenty-five sittings, and what ended each one
Nine homes, forty-five working days, one closed list of nine endings and a dash that turned out to be the commonest entry of all. The method is published in full, and so are the questions it cannot answer.