The short answer
Track actual referred visits separately from publication-wide traffic estimates. Record the article or referring domain, reporting period and the measurement source. A placement's presence on a site does not establish how many readers clicked through or what they did after arriving.
A worked example
A fictional article sits on a publication with a large estimated audience but sends only a small observed number of visits. The team logs those visits honestly and checks the destination's usefulness. It does not multiply the publication estimate by an assumed click rate and report the result as measured referral traffic.
A practical checklist
- Define the referral source and observation window before reporting. Confirm which analytics fields or campaign identifiers are actually available in the deployed implementation.
- Review landing-page behavior and relevant downstream actions. Keep measurement privacy-safe and do not infer identity from a referring domain.
- Report observed visits and actions with the attribution limits. Compare campaigns only when the scope and periods are sufficiently similar.
What to avoid
Avoid assuming that every direct visit after publication came from the article or that every referred visit was a qualified lead. Some sources and journeys will remain unobserved. State those gaps instead of inventing a complete attribution story.
A useful follow-up
Can estimated publication traffic stand in for referred visits?
No. The estimate describes its stated scope; referral visits require separate observed measurement.