Async across 4 timezones: what we changed after a 9-hour gap broke our sprint

My product team is spread across Manila, Bengaluru, Lisbon and Toronto. On paper it's "follow the sun". In practice, in August, it was "follow the confusion".

What broke

Our Toronto engineer asked a question about an API change at 5pm her time. That's 5am in Manila, 2:30am in Bengaluru. The Bengaluru dev had the answer, but by the time he woke up the question was buried under 40 messages in a busy channel. The change waited 26 hours. The sprint slipped by two days. Nobody did anything wrong. The system did.

What we changed

1. A decision log. One doc. Every decision gets a line: date, what was decided, who owns it, link to the discussion. If it isn't in the log, it isn't decided.
2. End-of-day handoff notes with a fixed template:

   Done today: ...
   Open questions (with @owner): ...
   What I need by your morning: ...

   That last line is the one that saved us.
3. Questions go in threads named after the ticket, never in DMs and never loose in a busy channel.
4. One synchronous meeting a week, with a rotating time, so the pain is shared. We call it pain rotation. Every timezone gets the bad slot sometimes.
5. No meeting over 30 minutes without a written pre-read. People in bad timezones join to discuss, not to listen to a slide being read aloud.
6. Follow the sun for incidents, with a written handover at each shift change.

Results after 6 weeks

• Cycle time (ticket started → merged): 6.1 days → 4.3 days
• Meetings per person per week: 7 → 3
• "Wait, who decided that?" messages in Slack: I stopped counting, but a lot fewer

Heads up for this month

Europe moves its clocks back on the last Sunday of October and North America a week later. For that one week our rotation calendar is wrong. Every single year. Put a reminder now.

For individuals, Clément's video is a good list of personal remote habits. This post is more about the team side.

What's your team's worst timezone story? I collect them. For research. And comfort.