I’ll start with a data point: over the last year in our internal Grafana dashboards, the number of unique third-party ad server impressions we track has dropped by over 40%. The trendline isn't just dipping; it's heading toward zero.
This isn't about one platform failing. It's about the fundamental shift in how we handle ad delivery and measurement. The classic third-party ad server model—centralized, expensive, and often slow to report—feels increasingly misaligned with modern needs. Why?
First, walled gardens (Meta, Google, Amazon) now dominate so much of the spend, and they provide their own serving and reporting. Their data often lives in their own silos, making a third-party server feel like an extra, costly layer.
Second, the rise of server-to-server integrations and cloud data platforms means many teams now pipe campaign logs directly into their data warehouse (BigQuery, Snowflake, etc.). From there, they build dashboards in tools like Grafana for near-real-time performance and cost monitoring. The ad server's reporting UI becomes redundant.
And third, the focus has shifted. It's less about "serving an ad" and more about holistic observability: tying ad exposure to downstream events, managing frequency across channels in one view, and monitoring CAC in real-time. That requires a flexible, owned data stack, not a black-box vendor.
I'm curious if others are seeing this in their own monitoring setups. Are you still relying heavily on third-party ad server dashboards, or have you moved the core metrics into your own observability platforms? What are you using to stitch the data together?
- GG