Finding a meeting time that respects every time zone
Every distributed team has a meeting that somebody dreads — not for its content, but for its hour. It was set once, by whoever booked the room first, in their own morning, and it has been 21:00 or 06:00 for the same colleagues ever since. Finding a fair time is not a calendar feature. It is a small piece of arithmetic plus one honest admission: sometimes no good hour exists, and then the job is to share the bad ones deliberately.
Why "nine o'clock" is not a time
The naive proposal — "let's meet at nine" — fails because nine is not a moment, it is twenty-four different moments, one per time zone, spread across two calendar days. What the proposer means is nine for them, and the conversion work is silently outsourced to everyone else. Run the numbers for the classic four-city team. Nine in Shanghai is 09:00 at home, 09:00 in Singapore (the one easy pairing — the two cities share an offset), 01:00 in London, and 20:00 the previous evening in New York. Two colleagues are at work, one is asleep, one is at dinner with their family. Nobody chose that outcome; it fell out of a sentence that sounded reasonable in one city. The fix is not a better sentence. It is looking at the whole day at once — every hour, every city, side by side — which is what the time zone meeting planner draws: rows for cities, columns for the UTC hours of a date, each cell coloured by whether that instant is work, shoulder or night on the local clock.
The overlap maths is brutal and simple
A work window of 09:00–17:00 is eight hours wide. Two cities whose clocks differ by D hours share a work overlap of 8 − D hours, and when D reaches eight the overlap is exactly zero: one city's 17:00 is the other's 09:00, a handshake with no hour inside it. Differ by more than eight and there is no hour in which both are at work — not a bad hour, no hour. London and Shanghai differ by eight in winter: the overlap is the single point where London starts as Shanghai finishes, which is to say there is none. Add New York, five hours behind London and thirteen behind Shanghai, and the four-city intersection is empty by hours, not minutes. This is why experienced global teams stop hunting for the perfect slot surprisingly fast. The geometry forbids it. What remains is optimisation, not search: widen one side into the shoulder — the two hours either side of the window that people accept occasionally — and a twelve-hour spread becomes coverable at a price somebody specific pays. A good planner prices it honestly, ranking candidate hours by how many cities are fully at work, then by how many are at least in shoulder, instead of printing "no overlap found" and leaving the politics to you.
The DST traps: your slot moves twice a year
Suppose you find your compromise in January: 13:00 UTC, which is 21:00 in Shanghai (late, accepted), 13:00 in London, 08:00 in New York. In late March the United States springs forward and Europe does not — for three weeks New York is four hours behind London instead of five, and your 13:00 UTC becomes 09:00 for New York: better for them, invisible to you. Then Europe moves, the offset restores, and in late October the whole dance runs backwards with a one-week gap in the other direction. Meanwhile Sydney, if you have Sydney, has been moving the opposite way entirely, gaining and losing two hours against London across the year. Nothing about your meeting changed; the ground under it did, twice a year, on dates that differ by country and that nobody has memorised. The defences are mechanical. Store the meeting in a named time zone, never a fixed offset, so calendars move it correctly. Avoid anchoring a recurring slot in the changeover weeks if you can. And check the specific dates: the planner compares each zone's offset a week either side of the day you pick and flags the ones about to change — the warning exists because "it worked last month" is exactly how these failures arrive.
When there is no good hour: rotate, or go async
For spreads beyond twelve hours, the mature options are two. The first is deliberate rotation: alternate between two compromise slots — say 08:00 New York / 21:00 Shanghai one week, and the same cities' kinder hours the next — so the late night and the early start change owners on a published schedule. Rotation fails when it is accidental, drifting with whoever books; it works when it is a rota, as predictable as an on-call schedule. The second option is to question the meeting. A status update crosses twelve time zones perfectly well as a written note, and the hours a global team can share are scarce enough to reserve for the conversations that genuinely need synchrony: decisions, disagreements, design. Teams that move status async and rotate the rest usually discover their one shared hour got better — because it is finally being spent on the thing only that hour can do.
Start with the grid, not the invitation: open the meeting planner, put your cities in, and look at the actual shape of your day together. If a green window exists, book inside it and defend it. If it does not, pick the compromise hours the planner ranks highest, rotate them on a schedule everyone can see, and write the reasoning into the invitation — people accept a 21:00 meeting they understand far better than one that simply descended on them.
Frequently asked questions
What is the widest team spread that still has a real overlap?
With a 09:00–17:00 window, two cities can be at most eight hours apart and share a full hour; stretch one side into shoulder hours and you can cover about twelve. Beyond that — Shanghai to New York is twelve to thirteen — no honest overlap exists, and the answer is rotation or asynchrony, not a braver search. The meeting planner shows the boundary for your exact cities instead of leaving you to guess where it lies.
When exactly do the clock changes break recurring meetings?
In the weeks when regions change on different Sundays: the US springs forward on the second Sunday of March, Europe on the last Sunday; in autumn the US falls back a week after Europe. For two or three weeks each spring and autumn, transatlantic offsets sit an hour away from their usual value. The southern hemisphere moves on yet other dates, in the opposite direction. Set recurring meetings with a calendar that stores the zone, not the offset — and re-check the slot in those weeks.
Is a rotating meeting actually fairer?
Yes, measurably. A fixed bad slot charges the same people the same cost forever — the 21:00 attendee is tired at every single meeting. Rotating between two or three compromise hours spreads that cost, and attendance and sharpness recover on both sides. The unfair version is the accidental rotation, where the slot drifts because whoever schedules picks their own morning. Deliberate rotation, published in advance, is the version people accept.