Traffic Analysis

Choose user, session or event acquisition deliberately. Record property, dates, timezone, filters, channel grouping and attribution dimension. A first-user source and a session source answer different questions. Compare absolute counts and shares by source/medium, country, device, landing page and downstream outcome when available.

For external campaign links, define consistent utm_source, utm_medium, utm_campaign and optional content or term. Test redirect preservation. Do not tag internal links, which can overwrite acquisition context. Check the current GA4 channel rules instead of assuming a proposed medium always maps to a chosen channel.

GA4 (direct) / (none) means no clear referral source. Typed URLs, bookmarks, offline documents, lost tags, redirects, ad blockers and app handoffs can contribute. A deep landing URL merits investigation but cannot identify its origin. See Google's direct-traffic explanation. Avoid unsupported fixed percentages for dark traffic by app.

For organic search, separate branded and non-branded demand where query data permits, while acknowledging hidden query rows. Brand and non-brand can differ in intent and conversion, so avoid interpreting combined growth as generic SEO improvement. For a spike, preserve original landing URL, UTM, referrer where present, click IDs and time distribution. Check campaigns, social distribution, tracking changes, redirects, internal or bot traffic and data freshness. Segment known internal activity and investigate automation using behavioral evidence; GA4 alone cannot certify that every remaining session is human. A UTM describes a tagged link, not the environment where every user clicked it. Label observations separately from source hypotheses.

Open Analytics' traffic-anomaly-investigation compares the last two observed dates per source/medium. It does not run statistical anomaly detection or multi-touch attribution. Deliver a scoped channel table, measurement gaps, hypotheses and tests that could distinguish them. Use analytics-tracking for event implementation.