fix(case-studies): correct Pomerado relay numbers
The Pomerado numbers merged in #123 came from a message store that only goes back to October 14. Pomerado started on September 22, so the page undercounted and called the wrong days the busiest.
This uses PostHog send events (message created, DM, thread reply, group DM) for both Pomerado workspaces, September 22 to November 16, Pacific time:
| Was | Now | |
|---|---|---|
| Messages | 30,000+ | 38,000+ (38,557) |
| Typical week | 5,000-6,500, up to 98 agents | 4,000-6,600, up to 97 agents sending |
| Busiest day | 1,313 (Oct 21) | 1,753 (Sep 24) |
| Busiest hour | 188 | 201 |
| Agents connected | 86 first week, 650+ by launch | 390 since Oct 8; 144 in launch week |
| Launch-week paragraph | 4,893 msgs, then 5,247 / 6,551 / 5,068 | 4,302 msgs, then 5,611 / 6,601 / 4,803 |
Agent counts come from presence heartbeats, which only start October 8, so there's no count for their first two weeks. The old "650+" counted agent records rather than connected agents and is dropped.
vitest run lib/test/case-studies.test.ts passes.
🤖 Generated with Claude Code
![Autofix with [code]smith](https://pr-comments-assets.blacksmith.sh/codesmith/autofix-with-codesmith-dark-3x.png)
Open in [code]smith · Need help on this PR? Tag @codesmith-bot with what you need.
Summary by cubic
Corrects the Pomerado case study numbers, which previously came from a message store that only went back to October 14 and missed Pomerado's first three weeks on the relay.
- Replaces all figures with PostHog send events from both workspaces, September 22 to November 16 (Pacific time).
- Updates the total to 38,000+ messages, and corrects the busiest day (September 24), busiest hour, and weekly ranges.
- Drops the old "650+ agents" figure, which counted agent records rather than connected agents; presence data only begins October 8.