Partial vs sitewide manual action: how much of the site do we need to fix?

Scope describes where the reported action applies; it does not tell us the number of engineering tasks or editorial decisions needed to fix the cause. Google's original explanation of partial matches used a forum section as an example of a limited action. That historical article explains the terminology; use today's Search Console notice for your site's current scope.

We copy the affected scope as written before translating it into an internal task. Avoid reducing it to 'just three pages' because three examples are visible, or calling it sitewide because the traffic loss is large. If access covers only a URL-prefix property, record that limitation and ask the property owner to confirm the relevant coverage.

Our report walkthrough covers finding the notice. This guide begins once the team needs to turn that notice into a reliable audit boundary.

Do not treat example URLs as an exhaustive inventory. They illustrate the issue; our job is to identify the related pattern and its full footprint. Google's Manual Actions report guidance requires remediation across affected pages rather than a partial cleanup.

Start by comparing the examples: do they share a template, publication account, content supplier, link placement, or generation rule? Then search the site's own content and technical records for that relationship. A set of visible URLs is evidence; the connecting mechanism is often what makes the investigation scalable.

We maintain a reason for including each group in the audit. 'Uses the same promotional template' is testable. 'Traffic went down around the same time' is a reason to investigate performance, but it does not prove the same manual-action violation is present. Keeping those reasons separate helps avoid both under-cleanup and unnecessary removal. We ask the investigator to state what would justify stopping the expansion, such as an independently verified template boundary. Otherwise an audit can become unbounded without becoming more complete.

Yes, an action can have a limited section-level scope. We still check whether that section shares the offending mechanism with other parts of the site. A forum may occupy one folder while its profile pages or attachments live elsewhere. A publishing integration may populate several paths. The visible directory structure and the system creating the problem are not always identical.

For an illustrative audit, begin with the named forum path, identify the content types created by the same user accounts, and inspect whether the same spam pattern appears there. Expand the investigation based on findings, not on an assumption that every neighboring page is penalized. The reported action boundary and our investigative boundary can differ without contradicting each other.

Some policies also require attention to audience or location. Read the current notice and current policy text , especially for site reputation policy, instead of applying a blanket rule derived from a different action. A folder map alone is not enough to explain every enforcement context.

We use one row per affected group with a concrete owner: URL pattern, content source, observed violation, intended correction, deployment or edit reference, and verification result. Keep the original examples tagged so they remain easy to revisit, but add independently identified examples to test the reach of the correction.

For a shared-component problem, the inventory may center on templates and the page families that render them. For a content problem, it may center on an import batch or editorial category. For a link issue, the relevant evidence may be external placements rather than an internal directory. Use the unnatural-links guide or the appropriate policy guide to select the right unit of analysis.

Mark unknown areas explicitly. A spreadsheet full of green cells is unhelpful if the team quietly excluded an old CMS or subdomain. Write down what was included, why it was included, and any coverage limitations. The next review should challenge those boundaries before the request asserts that work is complete. We prefer an explicit unresolved group to an impressive cleanup percentage whose denominator quietly excludes difficult systems.

A performance change outside the named scope deserves investigation, but it does not redefine the notice. We compare affected and apparently unaffected groups with consistent date ranges and filters. Google's traffic-drop guidance covers technical problems, ranking changes, seasonality, and other causes that may overlap with the incident.

Record other changes made during remediation: removed pages, navigation edits, migrations, redirects, or analytics changes. Those events can alter what you are measuring. If the marketing team compares a smaller post-cleanup site with its entire former inventory, the total alone cannot tell us whether the remaining legitimate pages are recovering.

The submission should explain how the investigation moved from the notice to the complete affected pattern, which groups were corrected, and how the team checked the result. We do not promise that a partial action will disappear after a partial fix, or that a sitewide action requires deleting every page. The response follows the violation and evidence.

Our reconsideration guide covers the submission. If the decision is unsuccessful, the scope record makes the next investigation more focused. After clearance, use the post-revocation guide to assess visibility separately. For a complex inventory spanning several teams, our manual penalty support can help coordinate the work.