Follow-the-sun support: what it takes to actually work
Three regions, eight hours each, no night shifts. That is the promise. The handover is where it succeeds or quietly fails.
The standard shape
A three-region rota gives roughly 24-hour coverage without anyone working nights:
| Region | Local hours | UTC window |
|---|---|---|
| Asia-Pacific | 09:00–17:00 Singapore | 01:00–09:00 |
| Europe | 09:00–17:00 London | 08:00–16:00 |
| Americas | 09:00–17:00 New York | 14:00–22:00 |
Note the gap: 22:00 to 01:00 UTC is uncovered. Most teams fill it with an on-call rotation rather than a fourth shift, because three hours does not justify a region.
The handover is the whole thing
Coverage is easy to draw and hard to run, and the failure is always the same: work restarts rather than continues. A handover note that actually works has four parts:
- Open items with owner and next action — not a list of ticket numbers.
- Anything time-sensitive in the next eight hours, stated in UTC.
- What was tried and ruled out, so the next shift does not repeat it.
- Customer commitments made, including anything promised by a specific time.
Why UTC, not local
Three regions writing in three local zones produces a timeline nobody can reconstruct. Pick UTC for everything written down and let each person convert for themselves. It is the one convention that costs each region equally.
The overlap window
Build in 30 minutes where both shifts are online. It sounds expensive and pays for itself the first time a live conversation prevents a repeated investigation. Europe–Americas overlaps naturally around 14:00–16:00 UTC; Asia–Europe around 08:00–09:00 UTC.
The clock-change trap
When the US and EU shift on different dates, a rota built on local start times develops a gap or an overlap for a few weeks. Anchoring the published rota to UTC and letting each region translate keeps the coverage continuous.
Two regions instead of three
Most companies cannot staff three regions, and a two-region rota is a legitimate structure — provided you are honest about the gap. Europe plus the US Americas covers roughly 08:00 to 22:00 UTC, which is fourteen hours. The remaining ten are on-call, and that needs to be said out loud rather than discovered during an incident.
| Structure | Covered | Gap |
|---|---|---|
| Europe only | 08:00–16:00 UTC | 16 hours |
| Europe + Americas | 08:00–22:00 UTC | 10 hours |
| Three regions | 01:00–22:00 UTC | 3 hours |
What "covered" should actually mean
Coverage is not the same as capacity. A single engineer awake in a region is coverage on paper and a bottleneck in practice. Define what each shift can handle alone and what triggers waking someone in another region — and write it down before you need it, not during.
Escalation across zones
- Define severity in customer terms, not internal ones, so the on-call person does not have to interpret.
- Name one escalation contact per region, with a documented phone route that does not depend on chat.
- Set a hard timer. If a Sev-1 is not moving in 30 minutes, the next region gets woken automatically.
- Review it quarterly, because people change roles and rotas quietly go stale.
The clock-change checklist
Four dates a year affect a US–EU rota: the two US changes and the two EU ones. Publish the rota in UTC, and on each of those dates confirm that the local shift start times still line up. Teams that skip this discover the gap when a ticket sits unassigned for an hour.
Measuring whether it works
Track time-to-first-response by shift and by handover boundary. If responses are consistently slower in the hour after a handover, the handover note is the problem — not the rota.
Filed under Operations · Tags: support, on-call, follow the sun, rota. Figures verified against the official definitions — see our editorial policy.
