Manual actions & reconsideration

Understand the affected scope, fix the underlying violation, and submit a useful reconsideration request.

GSC UI required Advanced

A manual action means a human reviewer at Google determined that pages on a site do not comply with applicable Google Search spam or content policies. Some or all of the site can rank lower or be omitted from Search, usually without a browser warning.

The recovery sequence is: preserve the notice, determine its scope, understand the exact policy, remove the violating mechanism everywhere, correct the system that allowed it, validate the result in the contexts Google reviews, and submit a concise evidence-backed reconsideration request.

This guide reflects Google's documentation on July 18, 2026.

Manual action does not mean “algorithm penalty”

Google uses automated systems and, when needed, human review to detect spam. Only a human-issued action appears in the Manual Actions report.

Therefore:

a traffic decline with a green Manual Actions report is not evidence of a hidden manual action;
an algorithmic ranking change has no reconsideration-request path;
a visible manual action should be remediated according to its notice, not guessed from traffic patterns;
revocation removes the manual action but does not guarantee previous rankings or traffic.

Check the report directly and preserve its exact wording. Search Console also sends a message when a site is affected.

Manual Actions and Security Issues are different

Manual Actions	Security Issues
Human-detected issues, mostly attempts to manipulate Search or violations of applicable Search feature policies	Hacked content or behavior that can harm visitors or their devices
Can demote or omit pages without a visible warning	Can show Search warnings or a browser interstitial
Review is requested in the Manual Actions report	Review is requested in the Security Issues report

There is overlap: compromised sites can contain spam, and deceptive behavior can involve both quality and user safety. Review both reports and fix each through its own workflow.

Determine whether the scope is partial or sitewide

Expand every action in the report. Google displays affected URL patterns or Affects all pages.

A pattern such as https://example.com/real-estate/* means some or all pages under that path are affected; it does not guarantee that every matching URL was individually reviewed. Conversely, examples are not a complete repair list.

Build a scope table:

Field	Record
Action type	Exact report label
Scope	Listed patterns or all pages
Notice date	Search Console message and discovery date
Examples	URLs or templates identified in review
Responsible system	CMS, partner, content pipeline, account, link campaign, markup generator
Policy source	Current issue-specific Google documentation
Remediation	Page change plus systemic prevention
Validation	Crawl, render, link evidence, logs, editorial or governance controls

If multiple actions appear, all must be fixed. Google says fixing only some affected pages will not produce a partial return to Search.

Preserve evidence before making changes

Record:

property type and exact scope;
every action label, description, pattern, and message;
traffic and indexing baseline by affected directory or template;
responsible releases, vendors, partners, campaigns, or content systems;
representative violating and compliant examples;
review, ownership, and deployment history.

This evidence helps prove remediation and avoids a vague request. Do not retain personal data, credentials, or harmful material outside approved systems.

Match the remediation to the action

The Manual Actions report contains current instructions for each action. Use the issue's Learn more guidance rather than a generic checklist.

Common action families include:

Content and scaled abuse

thin content with little or no added value;
scraped, doorway, or thin affiliate pages;
scaled content abuse or other major spam problems;
site reputation abuse involving third-party content;
user-generated or third-party spam;
spammy free-host services.

Removing a few examples is not enough. Inventory the production mechanism, decide which pages can become genuinely useful and compliant, remove or exclude the rest as permitted by the issue-specific policy, and prevent the same system from recreating them.

For site reputation abuse, Google's current guidance offers options such as removing the content, rebuilding it as first-party content, moving it to a new domain without passing signals from the old host, or using noindex. Moving unchanged third-party content to another established site or merely into a subdomain or directory may not resolve the abuse and can be treated as circumvention.

unnatural links to the site;
unnatural links from the site.

For inbound-link actions, identify links created through paid, artificial, deceptive, or manipulative practices. Google instructs owners to make a good-faith removal effort first, then disavow policy-violating links that cannot be removed. Blindly disavowing every backlink or submitting a list without attempting removal can lead to rejection.

Preserve:

source URL and target;
why the link violates policy;
acquisition source or campaign;
outreach date and result;
removal, qualification, or disavow action;
the final disavow file and affected property.

Do not disavow legitimate editorial links. A replacement upload overwrites the property's previous disavow list, and the tool does not support Domain properties.

For outbound-link actions, remove policy-violating links or qualify them so they do not pass ranking credit, using the attribute appropriate to the relationship. Also eliminate the commercial, editorial, template, or user-generated process that produced them.

Cloaking, redirects, and interface manipulation

cloaking or sneaky redirects;
sneaky mobile redirects;
cloaked images;
back-button hijacking;
hidden text or keyword stuffing;
AMP content mismatch.

Compare what Google retrieves with what users receive across device, referrer, geography, account, and rendering conditions. Check application code, redirect logic, tag managers, CMS extensions, edge rules, experiments, and compromised components.

Normal device adaptation or legitimate paywalls are not automatically cloaking. Intent, equivalence, policy-specific requirements, and what users can accomplish matter. Correct the actual behavior rather than making a one-off exception for Googlebot.

Structured data actions

Structured data must describe visible, relevant page content and follow both the general policies and the feature-specific documentation. Correct the templates or generators across the affected set, not only the sample markup.

Use a crawler and Rich Results Test or URL Inspection on representative pages. Passing syntax validation does not prove policy compliance: misleading, irrelevant, hidden, or inaccurate markup can be technically valid and still violate policy.

News and Discover policy actions

The report can include dangerous, deceptive, harassing, hateful, manipulated, misleading, explicit, violent, or other policy categories for News and Discover.

Follow the exact current policy and notice. Google may ask for evidence of changed editorial practice, such as updated editorial policies, relationship disclosure, oversight, or an editorial board with a history of improved practices. Deleting one item without changing the responsible editorial process may not be credible remediation.

Diagnose the system, not only the page

For every action, ask:

Who or what created the violation?
How many URLs or links can the mechanism affect?
When did it begin and is it still generating output?
Which controls failed—editorial, technical, vendor, moderation, access, or monitoring?
What stops recurrence after the examples are removed?

Possible systemic corrections include:

retiring a scaled page generator;
adding moderation and reputation controls for user content;
revising partner publishing and commercial review;
removing a link-acquisition practice;
correcting structured-data templates;
closing a compromised account or vulnerable component;
separating editorial and advertising decisions;
adding review gates, ownership, monitoring, and audit logs.

A reviewer needs evidence that the site is now compliant, not an explanation that a vendor or former employee was responsible.

Make affected pages reviewable

Google's general manual-action instructions say affected pages should be reachable: not unexpectedly behind login or paywall, blocked by robots.txt, or hidden from review through noindex. URL Inspection can test Google's access and rendering.

Follow issue-specific guidance where it intentionally differs. For example, Google documents crawlable noindex as one possible resolution for site-reputation-abuse content. In that case:

allow crawling so Google can see the directive;
do not block the URLs in robots.txt;
verify the noindex in the live response or rendered HTML;
document that this was the chosen issue-specific remediation.

For pages meant to remain indexed, ensure the corrected content and resources are accessible to Google. Do not create a temporary reviewer-only version that differs from what users receive.

Validate the complete remediation

Build a sample that covers:

every listed action and URL pattern;
each affected directory, template, content source, and partner;
high-traffic and recently created pages;
mobile and desktop behavior;
Google-rendered and user-visible content;
internal and external link changes;
structured data and rendered markup;
redirects, response codes, robots, and noindex where relevant;
content created after the fix;
recurrence controls and monitoring.

Use a full crawl or database/CMS export to quantify affected inventories. Search Console patterns and examples are not a complete list.

Do not submit while the generator, vendor feed, user-abuse path, or campaign can still recreate the violation.

Write a credible reconsideration request

Google says a good request:

explains the exact quality issue;
describes the steps taken to fix it;
documents the outcome.

Use a factual structure:

1. Acknowledge the action and scope

Name the report label and affected patterns. State what the investigation confirmed. If the domain was recently acquired, Google says to explain the acquisition after fixing the pre-existing issues.

2. Explain the root cause

Identify the page generator, editorial practice, partner relationship, link campaign, technical behavior, markup template, user-abuse path, or compromise. Avoid speculation and blame.

3. Detail removal and correction

Quantify what was removed, rewritten, qualified, redirected, noindexed under applicable guidance, or rebuilt. For link actions, summarize removal attempts and remaining disavowals. For editorial actions, identify policy and governance changes.

4. Prove the outcome

Provide representative corrected URLs, counts, crawl results, render checks, link evidence, deployment dates, and relevant logs or review records. Link to accessible supporting documentation only when it is necessary and appropriately permissioned.

5. Describe prevention

Explain the controls now preventing recurrence: moderation, publishing approval, vendor rules, technical validation, monitoring, ownership, training, or access changes.

Concise evidence is stronger than a long apology. Do not include secrets, personal data, exploit details, or unsupported promises.

Example request outline

Action: Site reputation abuse affecting /reviews/*.

Cause: A third-party publishing feed created 1,240 commercial pages without first-party editorial oversight.

Remediation: The feed was disabled on July 10. All 1,240 URLs were inventoried; 1,180 now return 410 and 60 first-party articles were independently rewritten under the new editorial process. Redirects to another established site were not used.

Validation: A July 15 crawl found no remaining feed template, indexable third-party pages, or internal links to removed URLs. Corrected examples: [representative URLs].

Prevention: Partner publishing now requires editorial ownership, policy review, template approval, and weekly inventory monitoring.

Use the real facts and issue-specific policy. Do not copy a generic template if it does not match the action.

Submit and wait for the decision

After every listed action is fixed, choose Request Review in the Manual Actions report. Search Console sends confirmation when the review is active and another message when it completes.

Google says most reconsideration reviews take several days or weeks. Link-related requests can take longer. Do not resubmit before the outstanding review has a final decision.

The absence of an instant reply is normal. Avoid making unrelated large changes during review unless required for users or security; they complicate the evidence baseline.

If the request is rejected

Rejection means the reviewer did not find complete, credible remediation. Reopen the evidence rather than resubmitting the same text.

Check for:

unsampled pages matching the affected pattern;
a generator or partner still publishing violations;
superficial rewrites with unchanged user value;
redirecting or relocating the same abusive content;
incomplete link outreach or an unsafe disavow list;
mobile, referrer, or geography-dependent behavior;
markup that passes syntax checks but remains misleading;
robots, authentication, rendering, or access preventing review;
multiple manual actions when only one was addressed;
recurrence after the initial cleanup.

Document additional work and submit only when validation supports a materially different outcome.

Monitor recovery after revocation

When the action is revoked, annotate the date and monitor:

the report and Search Console messages;
indexed status and Google-selected canonicals for representative URLs;
clicks and impressions for the affected scope;
new compliant content versus removed inventory;
crawl and response patterns;
moderation, link, vendor, or editorial-control metrics;
Security Issues if compromise contributed.

Revocation allows the site to be evaluated without that manual action. It does not restore deleted content, undo legitimate ranking changes, or guarantee return to a previous position. Recovery can depend on recrawling, reindexing, changed content and links, current competition, and Google's automated systems.

Define success as sustained policy compliance and healthy user value, not a one-day traffic rebound.

Automation boundary

The Manual Actions report and reconsideration submission require the Search Console UI. Search Analytics and URL Inspection APIs do not expose manual actions or review status. SearchConsole.ai cannot retrieve an action, submit a request, or guarantee revocation.

Automation can support remediation:

crawl affected patterns and validate responses, markup, robots, and canonicals;
inventory CMS, partner, user-generated, or structured-data output;
analyze performance for the affected directory;
track link-removal evidence and disavow candidates with human review;
inspect supplied URLs for Google's indexed and live technical state;
monitor recurrence after deployment.

Policy interpretation, page value, editorial governance, link intent, and the final request require accountable human judgment.

Common mistakes

Calling every unexplained ranking decline a manual action.
Treating displayed URL patterns as the complete repair inventory.
Fixing only example pages or only one of several actions.
Deleting outputs while leaving the generator or abuse path active.
Moving unchanged abusive content to another established host.
Blocking Google from seeing the remediation.
Assuming valid structured-data syntax means policy compliance.
Blindly disavowing all backlinks without removal effort.
Blaming an agency, partner, user, or prior owner without fixing controls.
Submitting a vague apology without quantities, examples, or validation.
Resubmitting while a review is pending.
Assuming revocation guarantees former rankings.