Scheduling a meeting across time zones

A meeting with people in three countries, one hour that suits them all. Everyone sees the proposed hours in their own time zone, already converted.

Updated on October 29, 2026

Why that invitation saying 9:00 went wrong

An invitation went out for 9:00. Three people joined at 9:00, one joined at 10:00, and one had been awake since 3:00 in the morning for nothing. Nobody misread it: the message never said which 9:00 it meant.

That is the whole difficulty of scheduling a meeting across time zones. The question itself is simple: it is just asked in a unit that changes value depending on who reads it. Ask six colleagues in four countries when they are free and every answer arrives in a different currency, and you end up converting them yourself, late, in a spreadsheet, twice.

An availability poll takes the conversion out of the conversation. You propose the hours once; each participant reads them on their own clock and ticks what works. The arithmetic is done before anyone is asked anything.

Should you write the time zone into the poll yourself?

The helpful instinct is to label everything by hand: a poll called “Kick-off 9:00 CET”, or date labels reading “14:00 CET / 8:00 ET”. Do not. It is the one move that reliably breaks.

Every reader already sees the proposed hours converted into their own time zone: the browser’s by default, or another one picked from the settings menu. The zone being used is displayed next to the table, so nobody has to trust a guess, and changing it redraws the table immediately with the same votes in it. Hours show in 12h or 24h format according to the reader’s locale, which is one more thing you do not have to decide for anyone.

A zone typed into the title is worse than redundant. It is also:

  • Wrong twice a year. CET becomes CEST at the end of March, and ET becomes EDT two weeks earlier. A label frozen into text does not follow; the converted hour does.
  • Ambiguous for most of the world. “ET” covers two different offsets depending on the month, and means nothing at all to a reader in Bangalore or São Paulo.
  • Contradicted by the row underneath it. The title claims 9:00 CET, the table shows 13:30 to the reader in India. One of the two is wrong, and it is never the table.

Write the title in plain words, “Q4 planning, one hour”, and let the hours answer for themselves.

How wide is the overlap, really?

Before proposing any hour, work out the window. That is the real decision, and it takes five minutes with a world clock.

Lay the local working days over each other, say 9:00 to 18:00 for each group. Western Europe and the east coast of the United States intersect by about three hours, late in the European afternoon. Swap the United States for India and the window widens to four or five hours, early in the European morning instead. Ask for all three in the same hour and there is nothing left at all: Bangalore has gone home before Boston sits down.

  1. 1

    Count the zones, not the people

    Six colleagues in two countries is a two-zone problem. One person travelling makes it three, so ask where everyone will be that week, not where they live.

  2. 2

    Find the overlap once, in one place

    Keep only what intersects, and accept the number you get. Three hours across two continents is normal, and a third continent often leaves no hour at all; neither means you did it wrong.

  3. 3

    Offer three to five slots inside it

    Spread over two or three different days rather than stacked on one afternoon. A single busy Tuesday should not rule everybody out at once.

  4. 4

    Say how long the meeting is

    An hour inside a two-hour window leaves almost no room. Put the duration in the title so nobody votes for a slot they cannot actually hold.

Which poll type fits an international meeting?

You pick a type when you create the poll, and it changes what participants are asked. Across time zones there is only one sensible answer, but it is worth knowing why the other two are not it.

Date

Whole days, with the hour settled later. Across zones that only postpones the problem: a day is not a time, and the group still has to agree on one.

Time slot

Precise hours, several per day if you need them. This is the one you want: the hours are what get converted for every reader.

Period

Several consecutive days: a seminar, a trip, a holiday. Right when the group travels to meet, not for a one-hour call.

What the table looks like when the voters are in four countries

Nothing special, which is the point. Slots stay in calendar order, each voter is a column, and each answer is “I am available”, “I might be available” or “I’m not available”: a green tick, an amber question mark, a red cross.

Lea · ParisMia · LondonRaj · BangaloreSam · Boston
Tue 10 November9:00-10:00✓✓✓✕
Tue 10 November14:00-15:00✓✓✓✓
Wed 11 November17:00-18:00✓✓?✓
Three slots, four cities, one table: the hours shown are always the reader’s own.

Green marks a slot every single voter said yes to. Amber marks the next best one after it, or the leading one when nobody reached unanimity, which is the table’s way of saying “this is as close as we got”. A slot nobody said yes to is never highlighted, and with one voter nothing is.

Read that same board from Bangalore and the three labels say 13:30, 18:30 and 21:30. Same poll, same ticks, a different column of hours. Nobody converted a thing.

“Add to calendar” then sits on the highlighted rows, in three formats: iCalendar (.ics), Google Calendar and Outlook Web. The time zone travels with the event, so it lands at the right hour in every calendar that receives it, whatever the recipient’s own clock. If you asked participants for their email when you created the poll, the calendar event can invite the people who said yes. There is more on that in adding the chosen date to your calendar.

What daylight saving does to a slot you picked in October

Daylight saving is what catches out the people who did everything else right. Europe and North America do not change their clocks on the same date: there is a stretch of about two weeks in spring, and another in autumn, when the usual gap between Paris and Boston is one hour different from the rest of the year.

For a single meeting this is a non-event. The poll converts the real hour on the real date, whatever the offset happens to be that week. It bites when you turn one chosen slot into a standing meeting by hand: a Tuesday 15:00 in Paris that was 9:00 in Boston becomes 10:00 in Boston for a fortnight, and somebody quietly misses the first fifteen minutes.

  • Vote on the date it will happen, not on “the usual time”. A poll run in October answers for October, and answers it correctly.
  • Re-run the poll for the next quarter. There is no recurring poll, and that costs less than it sounds: a new link, the same four slots, two minutes. Availability has moved anyway.
  • Watch the shoulder weeks. Late March and late October in Europe, and the two weeks on either side of them in North America. If a series crosses one, say so in the message instead of hoping.
  • The southern hemisphere runs the other way. A colleague in Sydney or Santiago shifts their clock while Europe is not moving at all, so the gap closes instead of opening.

Whose evening is it going to be?

A two-hour window means somebody is taking the call with dinner going cold. Labelling the slot “18:00” and leaving it there does not change that; it only means the person it costs finds out on their own.

Say it in the title or in the message that carries the link: “17:00-19:00 Paris time, so late evening in Bangalore: sorry, and tell me if that is impossible.” People accept an awkward hour they were asked about far more readily than one they were handed. And “I might be available” on a late slot often means “I can, but not every week”, which is worth a direct question rather than a guess.

Read next