Whose day
The Week You Can Change and the Week You Cannot
Most advice about time zones assumes you can change how people work. With clients and suppliers you cannot, and the remaining moves are different ones.
Who is working when, for the people on this page
The client's procurement office, Dubai
Thursday afternoon is their Friday
The port inspector, Santos
Monday is not a working day for him
two parties · three days of the seven are working days for all of them · Tue, Wed, Thu
Two of the parties these nine notes are about. Neither week was asked about in advance, and both were discovered by something failing to arrive.
When different working weeks make capacity hard to read, teams can the official product page before deciding whether a delay came from workload, handoffs, or a calendar mismatch.
Teams can compare this approach with the established guidance available from Oracle.
Five parties on one job, and who could be asked to move
Your team, Hamburg
Internal, and the only week anybody here can change
Your own contractor in Porto, on a retainer
Close enough to internal that a request would be heard
The client's procurement office, Dubai
Not askable: they are the ones paying
The fabrication shop, Taipei
Not askable: forty other customers on the same floor
The port inspector, Santos
Not askable at all, and not a commercial relationship
five parties · three days of the seven are working days for all of them · Tue, Wed, Thu
Of the five rows here, two could in principle be asked to shift a day and three could not. Every piece of advice written for distributed teams addresses the first two. One team's own job.
There is a large literature on working across time zones and almost all of it is addressed to people who can change something.
Adopt a core overlap. Write things down instead of meeting. Agree a convention for dates. Rotate the inconvenient call. Every one of those is sound and every one of them assumes an organisation that can issue an instruction and have it followed.
The assumption that does not hold
A client cannot be instructed. A supplier with forty other customers cannot be instructed. A port inspector, a customs broker, a notary, a certification body and a bank cannot be instructed by anybody at all, least of all by the party waiting on them.
With those parties the available moves are different. You can ask, once, and record the answer. You can observe what they actually do and plan around it. You can price the difference into what you agree. You cannot change their week.
What falls away when you cannot instruct
Several familiar recommendations stop being available, and it is worth naming them rather than quietly failing to apply them.
- A core overlap window. It requires both sides to adjust, and only one side here is adjusting.
- A shared tool with everybody's hours in it. The other party will not be setting up an account.
- A convention for how dates are written. You can follow one; you cannot impose one.
- A rota that moves the painful hour around. The painful hour will be yours.
What is left is narrower and more durable: find out their week, write it down, state your own instants unambiguously, and build the difference into the dates you agree rather than into your evenings.
The two populations
It helps to sort the parties on any piece of work into those two groups at the start, because the groups take completely different handling and mixing them produces advice that fits neither.
The first group is anybody whose week is in principle negotiable: your own team, a long-term contractor on a retainer, a subsidiary, an agency you are the largest customer of. With these you can propose, and a proposal that costs them little will usually be accepted.
The second group is everybody else. Here the only lever is the agreement, and the only moment the lever works is before it is signed.
The honest size of the first group
In the panel above the first group is two rows out of five, and that proportion is not unusual for physical work with a supply chain in it. For a software team with two vendors it may be the other way round.
The point is not the ratio. It is that the ratio is usually assumed rather than counted, and the assumption runs in the optimistic direction: people plan as though more of the chain is adjustable than actually is, and discover otherwise in the week something slips.
What to do with the second group
Three things, and they all happen early.
- Ask which days they work, before agreeing any date with them.
- Ask which of the next eight weeks contain a holiday on their side.
- State every date you agree as an instant with a zone attached.
None of those requires their cooperation beyond answering two questions, and the third requires nothing from them at all.
The thing that is actually in your control
Your own week is the one week on the page you can change, and the temptation is to change it — to start earlier, to answer in the evening, to take the call at six in the morning because that is when Taipei is awake.
That works and it is not free. It is a transfer of the problem onto whoever on your side has the fewest grounds to refuse, usually the most junior or the most conscientious person. A schedule that only holds because somebody is up at six is a schedule with an undeclared cost in it, and the cost is paid every week while the saving is counted once.
What this section is about
The six parts of this collection follow the shape of the problem rather than the shape of a working day. Whose days are working days at all; what a stated deadline turns out to mean on the other side; what it costs to wait and what it costs to be waited on; the hour, if there is one, where everybody is awake; and what to put in the agreement so that none of it is discovered afterwards.
The subject throughout is the party who never agreed to your arrangement. Your own distributed team is a different problem with a large and useful literature of its own, and almost none of it transfers.
A note on the figures in these fifty pages
Every panel in this collection is drawn from the record of one organisation's own work over about eighteen months: one buyer, a handful of suppliers, a client, a forwarder and the various offices and inspectors that sat between them.
None of it is a benchmark and none of it is research. A nine-day wait for a drawing is what happened on one occasion with one client, not a figure anybody should plan against. Where a panel shows a wait of fourteen days or an offset of seven hours, those are the actual numbers from the actual job, and the only claim made for them is that they happened.
They are included because the alternative — describing the shape of a problem without any numbers in it — produces advice that sounds reasonable and cannot be checked. A figure that can be disagreed with is more useful than a generality that cannot.