In this presentation the speaker highlights a common mismatch between QoE dashboards that show all green indicators and users who experience terrible call quality. The core issue is that relying solely on server‑side metrics (packet loss, jitter) provides an incomplete view because it cannot detect client‑side problems such as maxed‑out CPU, Bluetooth headset disconnections, or Wi‑Fi to cellular handoffs. Conversely, collecting only basic client‑side metrics tells you that something failed but not where—whether the fault lies in the device, the network, the carrier, or the infrastructure. Without a detailed, layered breakdown, teams waste days chasing ghosts, reproducing issues that a proper client‑side diagnostic could pinpoint instantly. The speaker argues for dashboards that surface the specific layer that broke, enabling rapid, targeted fixes.
