Why the $0.02/1M events claim was wrong (and what we learned)
Early on, our deck claimed a cost of $0.02 per million events. It was a nice number — and it was wrong. A million events on ClickHouse-grade infrastructure costs roughly $9–$15 to process, and pretending otherwise is exactly the kind of math that fails diligence and misleads customers.
What we got right is the structural point underneath: telemetry is infrastructure, and its cost should behave like infrastructure. Per-event processing on a tuned pipeline is dramatically cheaper than per-event pricing on a legacy analytics vendor, which is where the real savings come from.
We retired the $0.02 figure, replaced it with a measured target around $12 per million events at scale, and built an internal cost dashboard so the number is never aspirational again. When we publish a number, it should be one we can verify on our own infrastructure.
The lesson for us — and for any company pitching infrastructure — is that honesty about unit economics is a moat. Customers who understand your cost model trust your pricing; the ones who don't, walk away when they do the math.
More blog
We built a privacy-first analytics engine with zero cookies
Why the analytics industry got tracking wrong, and how session-storage hashing eliminates the need for consent banners.
The case for agent-readable analytics
Autonomous AI agents cannot scrape messy dashboards. They need deterministic HTTP JSON endpoints and structured metrics.
Turn every signal into your next decision.
Start with a lightweight tracker. Grow into agent-readable product intelligence.
No cookies. No third-party ad tracking. Your telemetry stays yours.