Publish green / yellow / red bands - retire ASAP
Distributed teams break when "quick reply" means different clocks. One person's lunch is another's midnight. Without explicit response expectations, people either stay half-online forever or get branded unresponsive for sleeping.
Build norms around overlap, urgency labels, and written defaults - not perpetual availability. GetToWork bands: Green (FYI, optional reply), Yellow (normal request, one local business day), Red (production down, page path only). Overlap hours are where synchronous chat is fair; outside overlap, green and yellow stay async. Overlap and handoff design live in timezone collaboration without chaos. This article owns the reply windows themselves.
Green / yellow / red response bands
GetToWork rules of thumb for internal team traffic. Adjust to your industry SLAs; keep the bands named so "urgent" cannot mean everything.
| Band | Meaning | Expectation | Channel |
|---|---|---|---|
| Green | FYI, status, non-blocking | Reply optional; acknowledge in next standup note or within two local business days | Async channel / ticket |
| Yellow | Normal request; someone is waiting but not stuck on production | Reply or park with owner within one local business day of the recipient | Ticket or named request channel |
| Red | Production down, customer-blocking outage, safety | Page / phone / on-call path only - not a red emoji in Slack | Incident path |
Overlap hours: named windows where synchronous chat and calls are fair game. Outside overlap, green and yellow stay async. Red uses the page path regardless of local clock for the person on rotation.
No expectation outside posted hours unless you are on-call. Quiet overnight is health, not disengagement, if work moves in writing.
Publish the bands where newcomers look. Repeat them when someone panics about a quiet thread.
Example team charter
Example team spans UTC−8, UTC, and UTC+2. They protect mornings for deep work and keep a 90-minute overlap for decisions. Always-on culture is rejected (available all day is a bad system).
| Type | Band | Expectation |
|---|---|---|
| Status / FYI | Green | Reply optional |
| Request (normal) | Yellow | Reply or park with owner within 1 local business day |
| Request (blocks someone) | Yellow, escalate if idle past overlap | Reply within next overlap or hand to on-call |
| Incident | Red | Page path only |
Kickoff / SOW snippet you can paste:
I respond to non-urgent (yellow) messages within one business day during CORE_HOURS in TIMEZONE. Green FYIs need no reply. For production blockers use ESCALATION_PATH (red). Overlap for live conversation is OVERLAP_WINDOW UTC. Thumbs-up means seen, not "I will do this today" - I will write a commitment with a time when I take work.
Bad vs good messages
| Bad | Good |
|---|---|
| "Need this ASAP" at 22:00 their time | "Need decision by Thursday 16:00 UTC; options A/B" |
| "Just ping me whenever" | "Overlap 14:00-15:30 UTC; otherwise async" |
| Silence with no park note | "Got it - reply tomorrow after the deploy" |
| Urgent emoji on preferences | Red reserved for true blockers |
| Thumbs-up as fake commitment | "Ack - detail by tomorrow 10:00 CET" |
Worked example: overnight blocker thread
Example: Engineer in UTC+2 posts a blocker at 17:00 local. Teammate in UTC−5 sees it at 10:00 their time - already deep in another task. Thread feels "ignored" though yellow said next business day.
Fix: blocker template with impact, attempted fixes, and "need by" in UTC once. Ack with "seen; will respond by 14:00 UTC" if resolution takes longer. The band stays yellow unless production is down - then red via page.
Managers and individuals
Managers: measure parked blockers, clear owners, and finished outcomes - not green dots. The person who answers at midnight must not become the standard; everyone else will look slow.
Individuals: post your hours, use status text, batch replies in windows. If your role is truly on-call, separate that rotation from normal async so your brain knows which mode you are in. Customer-facing support may owe a one-hour ack while engineering owes internal teams one business day - name both so engineers are not judged by the support clock.
Freelancers and clients
Clients import urgency from their local culture. Your morning deep block is their "end of day panic." Without a written response window in the kickoff, every client assumes you are their timezone.
State business hours, yellow ack time, and red blocker path in the SOW. Model async status on the rhythm you can keep (how to write async status people read). Example: client stops marking every email urgent after you consistently reply next morning with quality - and explain the bands once in writing.
Failure modes
Implicit heroics. Midnight replies become the norm.
Fake urgency. Everything tagged red until nothing is. Audit monthly: demote false reds.
Meeting substitution. Unclear async norms turn every question into a call across unfair hours. Spend overlap on decisions that fail in writing; rotate who takes early or late overlap (timezone collaboration without chaos).
Tooling defaults. Slack pushes notify now. Turn channel notifications off except on-call; keep urgent path separate. If the tool rewards instant reaction, norms must be louder than the tool.
Crunch forever. Name crunch weeks: "This week is incident mode." Return to default with "back to async bands Monday." Permanent crunch means the norm was never real.
Review quarterly: are windows still accurate? Is red abused? Is anyone burning out on midnight replies? Someone owns the review or the wiki is a graveyard.
Worked band labels in a real thread
Bad thread: "Need this ASAP" with a red circle emoji at 21:40 recipient-local. Recipient answers at midnight. Next week everyone expects midnight.
Good thread: "Yellow: need API field decision by Thursday 16:00 UTC. Options A/B in doc. If this becomes customer-blocking, I'll open a red incident - it is not one yet." Recipient replies next morning inside yellow: "Ack - picking A; PR by Wed 12:00 UTC."
Train distinguish ack emoji from commitment. Thumbs-up means seen. Commitment needs a verb and a time. New hires get the band table on day one with a buddy who corrects false reds for the first month - otherwise they copy the noisiest responder and reset culture toward ping-first.
Hybrid hire example: new PM in US Central joins EU-heavy eng and schedules "quick syncs" at 08:00 their time (late evening for two engineers). Fix in onboarding: overlap map, green/yellow/red bands, "live time only inside overlap unless red."
Holidays and local law belong in the same charter: response windows respect local holidays without treating them as surprise. Publish a team holiday calendar or shared time-off visibility. Night coverage for follow-the-sun is a rotation with handoffs - not infinite async availability for whoever drew the short straw.
Read receipts increase pressure to reply instantly across zones. Prefer cultural or technical defaults that match yellow (next business day) rather than training anxiety. Healthy signal: blockers parked with dates. Unhealthy: threads ending in "???" without an owner.
Operating rule: publish green / yellow / red reply bands with UTC need-bys, keep live chat inside green overlap, and treat red as a page path - not a vibe. Overlap math that makes the bands fair: timezone collaboration without chaos.