How to run a meeting across eight time zones
Past about eight hours of spread there is no comfortable slot. The question stops being "when" and becomes "whose turn is it".
First, find the actual overlap
Write every participant's working day in one reference zone and look for the intersection. For a team spread across San Francisco, London and Singapore:
| City | 9am–5pm local, in UTC |
|---|---|
| San Francisco | 16:00–00:00 |
| London | 08:00–16:00 |
| Singapore | 01:00–09:00 |
The intersection is empty. London and Singapore share 08:00–09:00 UTC; London and San Francisco share 16:00 UTC exactly. All three at once does not exist inside normal hours.
Then decide what actually needs to be synchronous
In most teams the honest answer is very little. Status updates, reviews and decisions with a clear owner can all be written. What genuinely benefits from being live is disagreement, ambiguity and anything where tone matters.
Then rotate
If a live call is unavoidable, rotate which region takes the awkward slot. A fixed schedule where the same office always joins at 10pm produces attrition, not complaints — people simply leave, and the reason given at the exit interview is rarely "time zones".
- Month one — 16:00 UTC: comfortable in London, early in San Francisco, late in Singapore.
- Month two — 01:00 UTC: comfortable in Singapore, late in London, afternoon in San Francisco.
- Month three — 08:00 UTC: comfortable in London and Singapore, very early in San Francisco.
Record everything
With a wide spread, the recording is not a courtesy — it is the primary artefact. Whoever could not attend needs the decision and the reasoning, not a video they will not watch. Written summaries beat recordings for exactly this reason.
The three-city test
Most distributed teams reduce to a handful of shapes. Here is what each one allows:
| Shape | Spread | Realistic core hours |
|---|---|---|
| US east + US west | 3 hours | 12:00–16:00 Eastern |
| US east + Western Europe | 5 hours | 14:00–17:00 London |
| US west + Western Europe | 8 hours | 17:00–18:00 London |
| Europe + India | 4.5 hours | 13:00–17:30 London |
| US west + Europe + India | 12.5 hours | None inside normal hours |
When the overlap is one hour
Protect it. A single shared hour should hold the things that genuinely need to be live — disagreements, ambiguity, anything where a written thread would take four days. Status updates and reviews do not qualify and will eat the hour if you let them.
When there is no overlap
Then the team is really two teams that hand over, and it works better when you design it that way:
- Write a handover note at the end of each region's day, in a fixed format.
- Give each region ownership of whole workstreams rather than shared tasks.
- Hold the all-hands on a rotating slot and always record it.
- Escalate through an on-call rotation with a written definition of "urgent".
The cost nobody measures
Question latency. A question asked at 5pm in London and answered at 9am in San Francisco costs sixteen hours of elapsed time and appears nowhere in any report. Teams that get good at this write questions with enough context to be answered in one pass, which is a writing skill rather than a scheduling one.
Filed under Scheduling · Tags: remote work, meetings, overlap. Figures verified against the official definitions — see our editorial policy.