An invalid-traffic spike is a signal to investigate, not a verdict

A traffic-quality report suddenly flags more activity as invalid. The timing overlaps a new acquisition campaign, but that coincidence does not identify the cause.

Invalid traffic can include automated activity and interactions that should not contribute to legitimate advertising counts. It is not a synonym for every unusual visit, and a low conversion rate does not prove fraud.

Google’s definition of invalid traffic provides examples for its advertising products. A particular measurement provider may classify and filter traffic differently, so start with the alert’s own definition.

Establish what changed

Record the affected property, reporting period, metric, and source. Determine whether the change is in raw activity, filtered activity, estimated earnings, finalized earnings, or an enforcement status.

These are not interchangeable. A deduction can relate to earlier delivery, while a current traffic graph describes a different interval. Check the dates before connecting the two.

Compare both counts and rates. If invalid activity rises from 100 to 200 while total activity rises from 1,000 to 10,000, the absolute count increased but the share fell. That still deserves attention, but the description should be accurate.

Segment without inventing an explanation

Review acquisition source, landing page, device, geography, timing, and placement where the data permits. Look for a concentrated change rather than immediately treating the whole property as affected.

A sudden burst from one source is a lead for investigation. It is not enough to accuse the source of deliberate manipulation. Shared networks, testing, implementation errors, and changing classification can complicate interpretation.

Keep logs and exports in approved storage. Redact identifiers before sharing examples outside the team.

Inspect your own setup

Check whether an ad overlaps controls, whether a page encourages accidental interaction, and whether internal testing touched live ads. Review recent layout and traffic-source changes.

Do not click your own live advertisements to test whether they work. Use the platform’s approved preview and diagnostic methods.

If a paid acquisition source cannot explain where its traffic comes from, escalate that gap to the owner. Pause or limit the affected source when the evidence and existing authority justify it, while preserving the information needed for review.

Google’s invalid-traffic FAQ describes its monitoring at a high level. It does not expose every detection rule, so a local analytics report cannot fully reproduce the platform’s assessment.

Measure the response

After a correction, compare equivalent periods and allow for the provider’s reporting delay. Record what was changed and whether the affected segment improved.

Avoid claiming that all remaining traffic is valid because a visible alert disappears. Equally, do not keep changing unrelated settings when the original affected interval has not finished processing.

The incident-response guide provides a structure for ownership and evidence. A useful resolution explains the observed issue, the action taken, and what remains unknown. It does not require a confident story unsupported by the available data.