A rank tracker can report a sharp fall while organic visits remain stable. That disagreement deserves a comparison of observations before it becomes a request to rewrite pages. November 20 coverage of “poisoned” Google result pages offers a concrete reason: a plausible-looking automated result set may contain substituted sites.
Nozzle’s research and DataForSEO’s account describe the collection problem from the supplier side. These are vendor findings, not a Google-confirmed explanation of every ranking fluctuation. DataForSEO reported a repair option in a November 16 update. Neither a vendor announcement nor stable total traffic proves that your particular alert is false.
Compare observations with the same dimensions
Keep the original tracker snapshot, collection time, query, country, target location, device and result URL. Include the surrounding result set, not just your position. If familiar competitors vanish together and unrelated sites appear, that is useful evidence for the supplier. It is not sufficient on its own to identify the cause.
Take a small manual sample using the closest available location and device conditions. Record when and how it was checked. Compare the actual result URLs rather than a broad domain label. A current manual search and a snapshot collected yesterday are not simultaneous observations, so disagreement between them can have several explanations. Preserve that limitation in the ticket.
Use Search Console to answer a different question
Google defines average position as an aggregate of the topmost result from the site for the relevant grouping. It is not a replacement measurement of the exact result page captured by a tracker. Filter to the affected query or landing page, country, device and comparable dates before interpreting clicks and impressions alongside it.
Then inspect the commercial scope. Stable sitewide visits can conceal a loss for an important product query, offset by growth elsewhere. Conversely, a tracker alert alone cannot establish lost enquiries or sales. Separate the supplier’s collection evidence, the site’s search observations and downstream business outcomes; do not force them into one supposedly equivalent ranking number.
Ask the supplier how the data was collected and repaired
Send the saved observation and ask whether the provider detected a questionable result set, which collection route produced it and whether the displayed history was reprocessed. If a repair feature is offered, ask which dates, markets and result types it covers. Find out how a corrected record is distinguished from the original so an automated client report does not mix both silently.
Using a second dashboard is not automatically an independent check: ask whether the tools rely on the same upstream data supplier. Agree who owns the investigation and what evidence will resolve the ticket. A useful answer explains the affected observations and remaining limits; a general assurance that the tool is now better does not validate a specific historical position.
Choose the action that the combined evidence supports
Use three explicit states. A measurement investigation is appropriate when the collection record is questionable and matched checks do not reproduce the claimed change. An SEO incident is appropriate when comparable observations support a real loss or independently confirm a broken page. Keep an unresolved state when samples, dates or query coverage are insufficient. The states are our decision framework, not official Google classifications.
For each state, name the next evidence required and the owner. Correct the affected report when a provider supplies a verified revision; investigate the site when the evidence supports a site problem. Do not overwrite uncertain history with invented positions or close an incident simply because an aggregate traffic line looks healthy. The durable skill is knowing which observation supports which action.
Sources
- Search Engine Land: Poisoned Google SERPs
- Nozzle: Poisoned Google SERPs research
- DataForSEO: Poisoned SERPs and provider response
- Google Search Console Help: Performance report
