“Ten is fine, and I will be in Denver that week.”
One clause in a friendly reply. Thursday at ten is still on the right calendar, still owned by the right account, still visible to everybody who needs to see it. And it’s now a different hour from the one you both thought you’d agreed.
The conversion itself is the reliable part of this subject, and has been for years. Nothing about it needs recalculating. What keeps the sentences right is a few habits: name cities instead of abbreviations, write the same moment in both clocks, say out loud what you’ve assumed about where somebody is. Those are set out in scheduling across time zones without the wrong hour. If what you want is a line of prose that reads one way only, read that one and come back.
This part is about the machinery rather than the wording. The lookup, the hold and the write that the last three parts built each pick up a second constraint when a zone is crossed. Each picks it up somewhere different. Two of those places are worth naming first, because between them they’re what the last part left standing.
The first is where each of you is physically standing. That’s a fact about a person. It changes without warning, and neither of you has it written down anywhere.
The second is a time zone setting. That’s a fact about an account rather than a fact about you. You may well hold three of them. They can disagree with each other, and each one has you in whatever city it was last told about. Denver is news to all three.
The window that is not on your grid
The lookup that finding a slot without the email chain argues for is one question with three parts in it: how long, inside what window, and under which of your standing rules. Ask it that way and the week gives one answer in one pass. Ask it by scanning and you get four minutes of work and a slightly different answer each time.
Across a zone the question grows a fourth part, and the fourth part is somebody else’s working day.
That’s a bigger addition than it sounds, and arithmetic has little to do with why. The first three parts are all things you can see. Your calendar draws your week, and your rules are yours. The fourth one sits off the grid. The hours a client in Boston is at a desk live outside your screen and always will. So you rebuild that constraint from scratch every time you need it, out of an office named in a signature and a memory of what the difference usually is.
Rebuilding it per request is what makes scheduling across zones feel expensive when the sums are free. It’s the same failure the whole series keeps finding. A judgment gets re-made from raw materials on every single message, when it could have been made once.
So make the other side a stored fact about the relationship rather than a calculation about the meeting. Two things, written down where your standing rules live.
The city they work in. The city, rather than the country, the office in the footer or the offset. Denver rather than mountain time. Boston rather than east coast.
The hours they keep in it. Nine to six, or eight to four, or the fact that Friday afternoon is gone. Most people will tell you if you ask once. Asking once is cheap, and it’s cheap the one time. Guessing is a weekly charge.
Store the city rather than the difference, because the difference belongs to the date as much as to the place. That holds domestically as well as across an ocean. Most of Arizona sits out daylight saving. NIST’s own local time guidance states that “Hawaii, American Samoa, Guam, Puerto Rico, the Virgin Islands and most of Arizona do not observe DST”, and notes the exception that “In Arizona, the Navajo Indian territories do observe DST” (checked September 6, 2026). A client in Phoenix is on the same clock as Los Angeles for most of the year, and an hour ahead of it in winter. Write down “Phoenix is two hours behind me” and you’ve written something that’s wrong for four months of every year. Write down “Phoenix” and everything you own stays true.
Then the part that surprises people, which is what the stored constraint does to the rules you already had.
Your standing rules were written for your own comfort in your own city, back when a second country was somebody else’s problem. Nothing before ten. No client calls on Friday. Wednesday afternoon belongs to the work. At home each of those costs you a slice of a large week, which is exactly why you can afford them.
Intersect them with somebody else’s day and the price changes. It changes differently for each relationship. A consultant on the east coast with a client in London has roughly four workable hours a day, nine to one, so a rule that starts the day at ten costs one hour out of four. The same consultant moves to Portland and the same rule costs everything. London’s working day ends at nine in the morning Pacific time, and the entire usable overlap is the hour or two before it. A mild preference on one coast is a decision to skip Europe on the other, and the person who made it remembers it as a preference.
This one is worth an actual review rather than a nod. Take the three or four rules you wrote down. Take the two or three relationships that cross a zone. Check each pair. Most people find one rule that’s quietly costing them a whole direction.
What the overlap will actually hold
Once the intersection is written down, it turns into something plainer than a fresh problem each week. It’s a band of hours, the same band most days, and it’s the one place a live conversation with that person can go.
Everything the first part says about choosing which of your free hours a given request deserves still applies. It just applies inside the band instead of across the week, and the band is small enough that the choosing gets real. The first conversation with a prospective client is supposed to get an hour you can walk into clean, with room behind it. When the only hours available are the two before your day properly starts, that hour is off the table. Pretend otherwise and the call that was meant to win the work gets taken at seven in the morning with a coffee in your hand.
Two consequences follow, and they’re decisions about the business rather than about the calendar.
A standing commitment can consume the band. A ninety-minute coaching session every Tuesday, placed in a two-hour overlap, is more than one meeting a week. It’s the whole of your capacity for that direction of the world, permanently, and the next client in the same region has nowhere to be put. At home this slips past you, because a week has plenty of Tuesdays. Before you agree to anything recurring inside a narrow band, find out what the band holds in total, and how much of it you’re signing away.
Clustering is worth more than it is at home. The first part makes the case that a new meeting is cheapest placed next to one that already exists, because a call dropped into the middle of an open morning takes the morning with it. Across zones the same saving comes back larger. Three early calls scattered through the week are three mornings that started before you did. The same three on one morning cost you one. The band is fixed anyway, so the choice of which day is nearly free. It’s the rare scheduling decision where being deliberate is free.
Plenty of readers get to stop here, and that’s a fine place to stop. If every client is within an hour or two of you, the band is the whole week, the intersection stays quiet, and the standing-rules work from the first part is the entire job. The rest of this piece is for the people whose Thursdays cross an ocean.
A hold that expires in the wrong morning
Three named hours with an empty calendar behind them are three hours pretending to be scarce. That’s why letting a client book the hour you meant to give them puts a tentative event on each time you name, and gives it an expiry at the moment it’s placed rather than at the moment somebody answers.
Both halves of that get harder across a gap, and they’re easy to miss, because neither one produces an error.
Start with the sentence. “I am keeping these open until Thursday morning” is a clean, useful clause at home and an ambiguous one at distance. Their Thursday morning and yours are two different objects. If they’re eight hours ahead, theirs has been and gone before yours starts. Whichever way the gap runs, one of you thinks the offer has expired while the other thinks it’s live, and whoever guesses wrong writes the awkward second message. The fix is the one that works everywhere else in this subject. Put the deadline in their clock, in the same sentence you put the times in.
Then the more expensive half, which is that the window is shorter than the number of hours in it suggests.
A two-day hold buys somebody nine hours away far less than two days of thinking time. Your Tuesday afternoon lands in their night. A message sent at four in the afternoon reaches a person who’ll next look at it tomorrow, on top of everything else that arrived overnight. What you actually gave them was one working morning. Count the offer in their business mornings rather than in hours, and the three-day answer to a two-day offer reads as prompt.
The same asymmetry raises the price of every round trip. The reply that changes a number is the one to watch for anywhere, because a counter-offer arrives dressed as an acceptance often enough to be worth checking every time. At home, spotting it and answering costs you an hour. Across nine zones it costs a day, and two rounds of it cost most of a week. That’s the strongest practical argument there is for handing over the choosing entirely. A shared page does the conversion at the moment somebody looks, so the exchange that would have taken three days is over as they look. The judgment about when a page is the right instrument and when it reads badly is the second part’s, and it holds here too. What changes is the weight on one side of it.
There’s a smaller trap in the release. A hold that comes off automatically comes off at a moment defined by a clock, and the clock belongs to whichever account the release is running in. Midnight is a time in a place. Which brings us to the setting that’s been sitting underneath all of this.
Which account thinks it is where
Each mailbox holds a time zone, and it holds it as a property of the account rather than a property of you.
Google’s own help documentation names two separate controls in a calendar’s settings. One is the “Primary time zone”, identified by searching for a city or country. The other is a “Display secondary time zone”, which puts a second column beside your week (checked September 6, 2026). Microsoft’s equivalent page names a “Display my calendar in time zone” control, and records a limit alongside it: “In the latest versions of Outlook for Microsoft 365, you can display up to three separate time zones in the Calendar” (checked September 6, 2026).
Two different things are on offer there, and they’re worth keeping apart, because one is free and one costs you something.
The extra zones are a display. Adding your main client’s city as a second column costs nothing, changes nothing, and puts their morning next to yours while you’re choosing an hour. If you take one action from this piece, take that one.
The primary zone does more. For the account you’re in, it’s what the grid is drawn in, and it’s what a new event gets created against. It decides what “ten o’clock” means when you write it.
Now put that next to the previous part. Keep a practice account, a personal account and a seat on a client’s system, and you’re holding three of these settings. Each was set at a different moment for a different reason, and any one of them could be pointing at a city you left. A booking page carries its own besides. Google’s documented settings for an appointment schedule include a time zone among the things fixed at the moment the schedule is built (checked September 6, 2026). A page you published two years ago is publishing hours in whatever zone was true two years ago.
That’s the setting the last part warned about, and here’s what it actually costs. The account that writes the event supplies the zone the event is anchored to. Choose the account and you’ve chosen the clock, silently, in the same act.
Travel is where the mismatch becomes visible, and it goes wrong in two shapes.
In the first, the account stays where it was and you move. Your grid keeps drawing your home week, which is internally consistent and therefore convincing. You spend two weeks offering hours in a clock that belongs to a city you left behind. This is the more common of the two, and the tell is that everything you propose is wrong by exactly the same amount.
In the second, you change the setting to where you are. That redraws every meeting you have, and anchors everything you create afterward to a city you’re about to leave. The right tool for a trip isn’t the account setting. It’s the event. The same Google page documents a time zone attached to an individual event rather than to the calendar. For a meeting that spans a move, it documents an option it calls separate start and end time zones (checked September 6, 2026).
The ten second check catches both. Look at what the event form says before you save anything. It names a zone. If that’s a different zone from the one you’re thinking in, everything downstream of it is about to be wrong, in a way that stays quiet for a week.
The clock a standing meeting is nailed to
A repeating meeting holds still for whoever created it and moves for everybody else, twice a year. The collection piece works through why the difference you memorized stops being true for a week or three, and what to send before the changeover weekend. That piece stands as it is.
What belongs to this part is narrower, and it happens earlier. Which clock did the meeting get nailed to in the first place, and who decided?
Nobody decided. The account that created it supplied the anchor. That account was chosen for other reasons entirely, and most of them were good ones: which address the invitation should come from, where the reschedule request needs to land. So a standing client call is usually anchored to whichever of your accounts you happened to be in on the afternoon you set it up.
That matters most when the drift lands on somebody who signed up for something else. If your client is in a place that keeps one clock all year, and most of the world does, then a meeting anchored to your city moves in their day twice a year while their own day stays put. They absorb it quietly, and it happens out of your sight.
There’s a move available here, and it’s the one people skip, because it looks like giving something up. Anchor the event to their city instead of yours, using the per-event setting above. The meeting then holds still for the client and drifts for you. That’s the right way round, for two reasons. You’re the one who knows the change is coming. You’re also the one who can see it, on a grid with their column on it.
This is also the one part of a badly created meeting that stays repairable. The previous part is blunt about the limits of fixing the account. A meeting can’t be walked across from one organization to another, and a video meeting can’t be handed over at all. The only remedy for the wrong account is a cancellation followed by a fresh invitation. The zone is friendlier. It sits on the event, you own the event, and changing it is an edit rather than a restart.
Where there is no good hour to find
Four limits, and the first one is a genuine answer rather than a shortfall.
Sometimes the band is empty. Past a certain gap, the best hour on offer is a less inconvenient one. Past a slightly larger gap, even that goes. A lookup works with the hours that exist. What’s needed then is a decision about the shape of the relationship, and the honest options are a written answer that arrives while the other person sleeps, taking turns, or a recording. Who should absorb the bad hour when there is one, and how to say so, belongs to the piece on getting the hour right rather than to this one.
Some offsets fall between the hours. India runs on a half hour, Newfoundland on a half hour the other way, and Nepal on three quarters of one. A band stored in round numbers rounds silently, and the direction it rounds decides whether you’re offering somebody an hour that ends after their day does.
Somebody moves and you still hold last month’s city. What you stored is wrong now, and it looks exactly as right as it did last month. The one thing that catches it is stating your assumption in the message and letting them correct it for free.
Their systems are anchored too. When the meeting comes through the client’s booking tool, their assistant or their tenant, the zone was set inside an account that sits out of your sight. You’re downstream of a decision somebody else made silently, which is the same problem this piece describes, seen from the wrong end.
Naming the hour in their clock
Point is a full email client running on the Gmail or Microsoft 365 mailbox you already have, so the thread you’re answering and the live calendar are one surface. The payoff for this subject is narrow, and it’s about when the check happens. You check while you’re still choosing an hour, rather than after the invitation has left.
An event’s time zone is a label you can tap. The picker behind it takes a city, an abbreviation or a plain offset, so Denver, MDT and minus six all reach the same answer. The event redraws in that clock while you’re looking at it. You see what two in the afternoon here is for the person you’re about to offer it to, before the offer exists. Dismiss the view and the event goes back into your own time, with your settings where they were. It’s a peek rather than a change of mind.
The rest of it runs where the request arrived. Ask for a meeting in plain words. Your reconciled calendars are read, and real times go into the reply. Point carries the correspondence from there, and understands the answers in whatever words people used. Once a time is agreed and you approve it, a proper event lands on the calendar you actually use, everyone on it. Where a shared page is the better instrument across a wide gap, and the section on holds argues that it often is, the page is drawn from the same reconciled week rather than from a second opinion about your Thursday.
Scheduling has an autonomy setting of its own, held apart from the rest of your mail, and it opens on review. There’s a concrete reason to leave it there for a while in this particular case. The first two or three proposals are how you find out whether the hour reaching a client is the hour you meant, and that’s worth watching before anything goes out unobserved. Further up, Point runs the exchange to the limit you set, and each position on the dial is written out. Actions are recorded in order, and undo works back along that record, as far as the edge of your own systems and no further. The rest of what changes across a mailbox is on the benefits page. Which half of this piece matters most depends on who this is built for. A coaching practice holding standing sessions abroad is mostly dealing with anchors, while a consultancy arranging one-off calls does the intersection fresh every time.
The band is written down. The hold expires in a morning both of you recognize. The invitation left from the right account, anchored to the right city. The event on the client’s screen in Denver reads in Denver’s clock.
Then, on the Monday, the client asks for it half an hour later.
Half an hour is the smallest change in scheduling, and the one everybody makes on reflex. You have the week open, the block is right there, and moving it is a drag and a release. At a distance, that half hour comes off the end of their afternoon, or goes onto the front of a morning that held one hour of overlap. On a wide enough gap it’s a different day. The move that costs nothing on your grid is the move to check at the far end, and the hard part of moving a meeting is somewhere else anyway.