What a Working Arrangement Looks Like
A description to hold yours against. Most of it is a table that was written, a route that has two names on it, and a meeting that is half an hour long.
The same organisation, before and after one year
| Before | After | Change | - | |
|---|---|---|---|---|
| Populations with rules established | 2 of 7reference | 7 of 7△ Differs | differs△ Differs | not recorded✗ Not recorded |
| Annexes confirmed in last 12 months | 1reference | 6△ Differs | differs△ Differs | not recorded✗ Not recorded |
| Configuration checked against local rules | Neverreference | All sites△ Differs | differs△ Differs | not recorded✗ Not recorded |
| Median lag, rule change to configuration | 244 daysreference | 31 days△ Differs | differs△ Differs | not recorded✗ Not recorded |
| Comparative answer available | Noreference | One page△ Differs | differs△ Differs | not recorded✗ Not recorded |
Nothing in this table was bought. A register was written, two names were added per country, a quarterly half-hour was put on an existing agenda, and the configuration was exported and compared once.
It is easier to recognise a working arrangement than to specify one, so here is a description to hold yours against. Nothing in it needs a system and most of it is a document somebody wrote.
The workflow in “What a Working Arrangement Looks Like” becomes more dependable when scheduled work, actual time and later corrections can be distinguished. For teams exploring how to measure employee productivity, learn more on this page can provide operational context, while policy ownership, employee explanation and consequential decisions remain with accountable people.
The test throughout is one question: can this organisation say, today, which rules apply to a named person, and show where that was established.
For an independent reference relevant to “What a Working Arrangement Looks Like”, consult the WHO mental-health-at-work guidance. Use it to challenge assumptions about working time, privacy, recordkeeping and exception handling against the organisation’s real operating model.
The register
One row per population. Place of work, employing entity, governing law, sector regime, the rule set established as applying, and the source with a date.
Gaps marked as gaps. Confirmed annually by the local entity, owned centrally, kept somewhere that survives a system change, with dated versions.
The grid
Nine dimensions by country: what counts, the maximum and its period, averaging, daily rest, weekly rest, breaks, night work, opt-out availability, and records.
Each cell with a source and a date. Blanks visible. Sorted once into the three groups — what changes the rota, what changes the configuration, what changes only a document — with an owner for each list.
The documents
A short global policy stating principles, method and responsibilities, containing no figures and naming no mechanism that does not exist everywhere.
A local annex per country holding the figures, the agreements in force, who to ask and who confirms it. Adopted by each entity with a record of how, including any consultation that was required.
The configuration
Exported as a table, per site, and compared against the grid. Week boundary, night period, break rules, rounding, holiday calendar granularity, time zone and the categories counted, all set per country rather than inherited.
A new country tested against seven deliberate cases before go-live, signed off by somebody local.
The people in between
A maintained country-of-work field. An approval gate on working from another country. A trigger on travel volume. A list of assignments against their end dates, reviewed annually.
Agency and provider populations in the register with what has been obtained in writing, and the organisation's own rota discipline applied to them regardless.
Keeping it current
Two names per country — who sends changes and who receives them — published in the annex. Acknowledgement of everything sent. A quarterly half-hour on an existing agenda with four columns: received, decided, implemented, confirmed.
Four dates recorded per change, so the lag is measurable. An annual confirmation that sends each owner their own row and asks one question.
When it is tested
A one-page sheet per country: who handles an inspection, what is typically asked for, where each item lives, what must be available locally and in which language, and who to call.
A comparative page for the question about other countries, with each difference marked legal, agreed or historical. And one named place that comparative questions are routed to.
What it feels like to run
The measurable version is a register with no blanks and a lag in weeks rather than months. The experienced version is quieter: somebody asks which rules apply to a person and the answer takes a minute rather than a fortnight.
New countries are added with a known list of questions. Changes arrive at an address and are acknowledged. The comparative question has an answer that is the same whoever is asked.
None of that is visible from outside and all of it comes from a table, two names per country and half an hour a quarter.
Getting there
Six things, perhaps twelve days of work spread across a year. Write the register with gaps. Fill the grid for the two largest countries and leave the rest blank. Export the configuration and compare it. Split the policy into a global part and annexes. Put two names per country on a list. And put half an hour on a quarterly agenda.
None of it is a project and none of it needs a budget line. What it needs is one person who owns the question of which rules apply to whom — which remains the thing most organisations are missing, and the thing that makes all the rest possible.