Traffic quality & anomalies

Select the population#

Reports default to human traffic. Switch to all to compare the broader population, or isolate known_bot, suspicious and unknown. Classification is evidence-based, not a guarantee that every automated or human visit is perfectly identified.

Investigate the evidence#

Open Quality to compare traffic classes, sources, devices, paths and outcomes. Use evidence trails to understand why an acquisition segment looks unusual. Avoid treating a spike or classification alone as proof of fraud.

Detect changes#

The anomaly tool evaluates its supported metrics; inspect its generated schema for the metric enum and limits. Anomaly alerts and weekly digests require configured email delivery and scheduled jobs. A disabled scheduler can leave notification settings enabled without sending anything.

A repeatable investigation#

1. Confirm installation did not change during the period.
2. Compare human and all traffic with identical dates.
3. Break down the changed segment by source and page.
4. Compare engagement, goals and payment evidence.
5. Export evidence for your own review without publishing visitor details.

API equivalents#

get_bot_traffic, get_evidence_trail, get_traffic_diagnostics and detect_anomalies.