What problem does it solve? When pageviews, sessions, or conversions suddenly spike or drop, teams waste time guessing at causes. This Skill provides a structured investigation workflow that confirms whether the change is real, rules out broken instrumentation, and isolates the dimension (channel, device, geography, or path) that explains most of the shift. ## Core Features & Use Cases - Signal confirmation: Queries the relevant web metric over a wide enough window to state the magnitude and timing of the anomaly in plain numbers. - Instrumentation checks: Correlates the change with deploys or SDK version changes to distinguish broken tracking from real user behavior shifts. - Dimension breakdown: Compares the anomalous window against baseline across referrer, device, geography, and top paths to find the dominant driver, with query hints in the bundled playbook. - Use Case: A user reports that signups dropped 40% overnight. The Skill confirms the drop, finds it is isolated to mobile Safari after a 1.2.0 release, and reports the likely cause plus one concrete next check. ## Quick Start Ask the agent to investigate why pageviews dropped sharply yesterday and identify the most likely cause.
