“We are spread across four time zones” is the sentence every distributed manager reaches for, and it describes almost nothing. It does not tell you whether you can hold a team meeting. It does not tell you how long a blocked person waits. It does not tell you which of the standard practices you have been handed will work and which will quietly damage the team.
There is one number that tells you all three, and most managers have never worked it out. It is the number of hours in a normal weekday when everyone on your team is awake, working, and available at the same time. Your overlap window.
Everything below sorts by that number, because advice that is sensible at four hours of overlap becomes actively harmful at one.
Zone count and overlap are different things, and only one of them matters
Two teams, both honestly described as spread across four time zones.
The first is a manager in Chicago with engineers in New York, Denver and San Francisco. Everyone works something close to nine to five in their own city. The overlap runs from ten in the morning Pacific to five in the afternoon Eastern, which is four solid hours in which the whole team is present. This team can hold a meeting whenever it likes. It can also mistake itself for a distributed team when it is really a team in one long business day.
The second is a manager in Berlin with one person in São Paulo, one in Bangalore and one in Sydney. Also four zones from end to end, if you count generously. The actual overlap between the Bangalore morning and the São Paulo morning is under an hour, Sydney is asleep for most of it, and there is no moment in the week when all four are at their desks. Every practice that worked for the first team fails here, not because the second manager is worse, but because the resource the first team spends freely does not exist.
Two things follow, and both are easy to get wrong.
Overlap is an intersection, so it collapses as the team grows. Two people usually share plenty. Add a third in a new region and the shared window is the part all three have in common, which can be dramatically smaller than any pair. Hiring one person eleven hours away does not reduce your overlap a little. It can remove it.
Your own hours are part of the calculation and you probably have not counted them honestly. If you answer messages until nine in the evening, your personal overlap with the east of your team looks larger than the one your team can rely on. What matters is the window people can plan against, not the window you happen to be reachable in.
Work it out once, properly
Ask everyone for the hours they actually work in their own local time, not the hours in their contract, and not the hours you assume. Someone with a school run at three in the afternoon has a real gap in the middle of the day, and it will not appear on any calendar.
Convert every person’s day into one reference zone and find the span all of them share. That span, in hours, is the only input to the rest of this article. Write it down somewhere the team can see, because the second most common failure in distributed teams is a manager scheduling against a window that only they believe in.
Then read the section that matches your number and skip the other two.
Four hours or more: you can still manage live, and that is the trap
With four hours or more you can keep almost every habit you learned in an office. You can call a meeting the same day. You can catch someone before they go down the wrong path. Someone blocked at the start of the window gets unblocked inside it.
Two warnings.
You are one hire away from a different job. Teams in this bracket build practices that depend on being able to talk, then add one person in a distant region and keep the practices. That person becomes a permanent exception: briefed second-hand, present for the recording rather than the discussion, and gradually less consulted because consulting them is inconvenient. The moment you widen the spread, the practice has to change, not the person. The cheaper move is to treat the overlap as a hiring criterion before the offer rather than a discovery after it, which is one of the things our guide to hiring remotely puts on the scorecard.
Wide overlap hides bad delegation. When you can intervene at any point in the day, you never find out whether your handoffs were clear, because you always catch the ambiguity in time. That is a real cost, just not a visible one. It is worth deliberately handing something over in writing and not correcting it for a day, purely to see what your instructions actually contain.
One to three hours: the overlap is now the scarcest thing you own
This is where most distributed teams sit, and it is where generic remote advice does the most damage, because it assumes a resource you have very little of.
Treat those hours the way you would treat a shared meeting room that exists for ninety minutes a day. The question stops being “when shall we meet” and becomes “what is this window for, and what is not allowed to consume it”.
What earns the overlap: decisions that need argument, work where two people have to think at the same time, anything with an emotional dimension, and unblocking someone who is stuck now. Difficult conversations belong here too. Nobody should receive hard feedback in a message and then wait ten hours for the follow-up.
What does not earn it: status. Announcements. Anything one person could have written and three could have read. If a recurring meeting inside your overlap consists of people describing what they did, it is consuming the most expensive hour on the team to transmit information that a document transmits better. Our meeting cost calculator will price that meeting for you, and the number is usually enough to end the argument.
Protect the window from fragmentation, not just from meetings. Ninety minutes of overlap chopped into six interruptions is not ninety minutes of collaboration. Each context switch carries a refocus cost that nobody logs and everybody pays, and it lands hardest on exactly the shared hours you were trying to protect. The context switching cost calculator puts a figure on the aggregate, which tends to be larger than managers expect.
Batch your own questions. A manager who fires off queries as they occur will drain a narrow overlap on their own. Hold them, send them together, and be explicit about which ones need an answer today.
Effectively zero: management becomes written or it does not happen
If there is no hour when the team is together, the practices that carried you through the previous two brackets are gone, and no amount of goodwill replaces them. Three things have to exist in writing.
Decisions and their reasons. Not activity logs, which measure presence rather than progress and are gameable by exactly the people you should worry about. What you need is a short record of what was decided, by whom, and what it ruled out. The reason matters more than the decision, because the next person to hit the question needs to know whether the circumstances still hold.
Standing permission to proceed. The defining cost of zero overlap is that a blocked person waits a full day for a sentence. The fix is not faster messaging, it is deciding in advance who may unblock whom without asking, and what size of decision people are allowed to make alone. Most first-time managers set that bar far too low and then wonder why velocity died. Say the number out loud: below this threshold, decide it and tell me afterwards.
Handoffs with an owner. Work passed between regions needs a named person on both ends and an explicit statement of what state it is in. “Over to you” without a state is how a task sits untouched for two days while both people believe the other has it.
Meetings do not disappear at zero overlap, they just become rare and expensive. When you do call one, somebody is joining outside their working day, which brings us to the question most managers avoid.
Core hours, and the version of it that fails
Core hours are the standard answer: a declared band when everyone is expected to be reachable. It works, with one condition that is almost always broken.
The failing version is core hours declared in the manager’s own time zone and presented as neutral. It is not neutral. It is a schedule that costs the manager nothing and costs everyone else their evenings, and the team can see that even when nobody says it. Compliance follows for a while, then erodes, and the erosion gets read as an attitude problem.
The honest version does three things. It names the cost out loud, so that the people paying it know it was noticed. It keeps the band as narrow as the work genuinely requires, because every hour you claim beyond that is borrowed from someone’s life. And it puts the manager inside the inconvenience at least some of the time, which is the only signal that reliably distinguishes a shared constraint from a preference dressed as one.
Whose sleep pays for the meeting
At some spread, one recurring meeting cannot be comfortable for everyone. You have two options and both cost something.
Fix the inconvenience on one group and you get a simple calendar and a slowly accumulating grievance. The people taking the late call notice that the arrangement was decided by the person it suits. This is defensible only when the group is small and compensated for it in a way they would recognize as compensation.
Rotate it and the cost is spread but coordination gets harder, because a meeting that moves is a meeting people miss. Rotation also has to be visible to work: if nobody can see the pattern, they only remember the times it landed on them.
The rule I would apply: rotate anything recurring, fix anything one-off. A weekly call that always lands at ten at night for the same person is a resignation being written slowly. A single unavoidable session at an awkward hour is a normal cost of the job, provided you say so and provided it is genuinely single.
The one-on-one is the meeting you protect first
If your overlap only supports a small number of live conversations per week, one-on-ones go to the front of the queue, ahead of team meetings. A team meeting that becomes a document loses very little. A one-on-one that becomes a document stops being a one-on-one.
Two adjustments matter at distance. Keep the cadence absolutely fixed, because a moved one-on-one across time zones is usually a cancelled one and the person on the other end reads the cancellation as a ranking. And write the agenda down beforehand, both sides, so that the scarce live time goes on the difficult item rather than on working out what the difficult item is. Our guide to remote one-on-ones goes through the structure in detail, including what to do when the connection is bad and what to do when the silence is not technical.
The person who is hiding in the gap
It happens, and pretending otherwise makes you useless when it does. Somewhere in a wide overlap gap there is occasionally a person using the delay as cover: answers arrive at the edge of your day, questions get deflected into the next cycle, and progress is always about to be visible.
Resist the obvious response. Tightening the monitoring is the reflex, and it punishes the whole team for one person while producing evidence about availability rather than output. The diagnostic that actually works at any distance is whether the work is moving, and it does not require you to watch anyone: agree what done looks like before the work starts, then check the work against it on a fixed rhythm. Someone who is genuinely stuck shows up differently from someone who is stalling, and both show up quickly once the check is about output.
If it is stalling, treat it as an underperformance conversation and not as a remote-work problem. The distance changes the logistics of that conversation and none of its substance.
Do this in the next hour
Collect your team’s real working hours, convert them into one zone, and compute the overlap. One number.
Then look at your recurring calendar and ask what percentage of that window is currently spent on people describing work they have already done. For most teams in the one-to-three-hour bracket the answer is uncomfortable, and it is the cheapest thing on this page to fix.
Time zones are one of six dimensions that break at distance, and they are rarely the one doing the most damage on their own. If the overlap arithmetic here did not describe your problem, the audit in our guide to what actually breaks when you manage remotely covers the other five, including the visibility trap that generates more bad management than distance ever does.