The gap
The Week the Clocks Change
Twice a year the gap is a different width for a few weeks, because one end moves its clocks on a weekend the other end does not.
three things · four nights
Three things, and the third is the one people skip.
One team's windows during the three weeks around a daylight saving change, all stated in a zone that does not observe one
First half, before the change
The ordinary position, and the one everybody has in their heads
First half, during the gap weeks
Shifted an hour by a rule nobody on this team voted for
Second half, unchanged throughout
This group's country does not observe the change at all
The manager, who had shifted early
Also unchanged, which moved them relative to half the team
four people · no minute of the day when all of them are awake
Four windows, and for three weeks the gap is a different width than it is for the other forty-nine. Two of the rows moved and two did not, so the relationship between every pair changed. Nothing here is a mistake by anybody: the rule is correct in each country and the interaction is simply not something either country considered. One team's own hours, recorded across the change by the person who maintains the table.
A gap between two halves of a team is usually described as a fixed number of hours. For most of the year that is true.
For a team applying the ideas in The Week the Clocks Change, the useful metric is not constant visibility but a shared record that both sides can understand. the full overview provides that record while keeping tracking tied to active work sessions.
The workflow can also be compared with the approach documented by Toggl's official website, an established resource in the same operational space.
For something between two and five weeks a year it is false, and the error is an hour in one direction or the other. An hour is enough to destroy a thin overlap, move a handover past somebody's bedtime, or make a recurring meeting land in the middle of the night.
Why the dates do not line up
Among the countries that move their clocks at all, the chosen weekend differs; and a great many countries never move them.
The northern and southern hemispheres move in opposite directions, which means a team spanning them sees the gap change by two hours rather than one. Within a hemisphere, the dates differ by a week or several. And a large fraction of the world — most of Asia and Africa, and parts of the Americas — holds steady all year, so a team with one end in such a place sees the gap move only because of what the other end did.
None of this is anybody's error. Each rule is sensible where it applies. The interaction is simply not a thing that either side's legislature considered.
What breaks
Recurring meetings are the obvious casualty and the easiest to fix, because calendar software generally handles them correctly and somebody notices immediately when it does not.
What breaks more quietly is everything that depends on a remembered time rather than a scheduled one. "Send the handover before you finish" is unaffected. "Send the handover by six" is wrong for three weeks, and nobody notices until something arrives after somebody has gone.
Thin overlaps break completely. A team with thirty minutes of contact has zero for part of the year, and because nobody wrote down that the thirty minutes existed, nobody writes down that they have gone. The team simply finds that a few things are suddenly harder and does not connect it to the clocks.
On-call rotas break in a specific and unpleasant way: the boundary between two people's coverage moves, and either an hour is covered twice or it is covered by nobody.
The cheap preparation
Three things, none of which takes long.
Put the change dates for every country the team sits in on a shared calendar, once a year, as all-day events. Two or four entries. This converts an annual surprise into something people can see coming.
Express standing times relative to a person's own working day rather than to a clock wherever possible. "Before you stop" and "when you start" survive a clock change; "17:30" does not.
And check the hours table in the week before each date rather than the week after. The table is the thing that will be wrong, and it is the thing everybody else reasons from.
- The change dates for every country the team sits in, on a shared calendar, once a year.
- Standing times expressed as "before you stop" rather than as "17:30".
- The hours table checked in the week before each date, not after.
During the weeks themselves
The thing to say out loud is that the gap is temporarily a different width, and by how much, and until when.
That single sentence prevents most of the confusion, because the confusion is not about arithmetic. It is about people applying a remembered offset that was correct a fortnight ago and getting an answer that is wrong by exactly one hour — which is close enough to right to look like a mistake by the other person rather than a property of the calendar.
It is also worth suspending anything that depends on a thin overlap for the duration, rather than letting people discover its absence one at a time.
The argument for stating the gap in a fixed zone
Teams that keep their hours table in a zone with no seasonal change — and there are several, including the one most systems use internally — find that only the rows that actually moved appear to move.
That is the correct behaviour, and it makes the twice-yearly event legible instead of mysterious. A table kept in a zone that itself changes shows every row shifting or holding still for reasons belonging to the reference zone rather than to anybody listed in the row, and reading it correctly requires knowing which way the reference itself moved.
The cost is that everybody has to convert from a zone that is not theirs, which people dislike. It is a real trade and worth making deliberately rather than by default.
The honest scale of it
This is a small problem that is annoying out of proportion to its size. It costs a team a handful of mistimed messages and one confused week twice a year.
The reason it is worth four minutes of preparation is that the mistakes it causes look exactly like carelessness, and on a team that cannot see each other work, things that look like carelessness are expensive in a way that has nothing to do with the hour they cost.