Google Manual Actions: Types, How to Check, and Penalty Recovery
How to confirm whether you have a Google manual action or an algorithmic drop, what each manual action type means and how to fix it, how to write a reconsideration request that gets approved, and what Google actually says about recovery time.
A Google manual action is a penalty that a human reviewer at Google applies when pages on your site break Google's spam policies, and it demotes or removes those pages, or the whole site, from Google Search. It appears in Search Console's Manual actions report and is lifted only after you fix every affected page and Google approves a reconsideration request.
Most "penalty recovery" work goes wrong in the first hour, when someone sees traffic fall and starts disavowing links or deleting pages before confirming what happened. Many sudden drops are not penalties at all: a noindex tag shipped to production, a broken redirect map, a tracking tag that stopped firing, or a seasonal dip. Others are algorithmic: a core or spam update reassessed the site, and no reviewer was involved.
This guide shows you how to tell the cases apart in under 30 minutes, lists the manual action types Google documents with the cause and fix for each, walks through recovery and the reconsideration request, and sets out the timelines Google actually states, so you can brief a client or a manager without overpromising.
Manual action vs algorithmic drop
People say "Google penalty" for two different things. Google uses the word "penalty" for neither; it talks about manual actions and about ranking systems. The difference decides the whole recovery path.
| Manual action | Algorithmic drop | |
|---|---|---|
| Who decides | A human reviewer at Google | Google's automated ranking and spam systems |
| Cause | A violation of Google's spam policies | A core update, spam update or other system reassessing your pages against competitors |
| Notification | Manual actions report and a Search Console message | None |
| How it is lifted | Fix every affected page, then a reconsideration request | Improve the site and wait for the systems to reassess it |
| Reconsideration request | Yes | Not available |
If you are not sure which you have, the next section settles it in a few minutes.
How to check for a manual action in Search Console
Step 1: open the Manual actions report. In Search Console, go to Security & Manual Actions → Manual actions. If it says "No issues detected", you do not have a manual action on that property. If you do have one, the report names the action type, lists the affected URL patterns (or says it affects all pages) and lets you expand each action for details and a link to Google's guidance. Google also notifies you in the Search Console message center. Check every property you own for the site (domain property, www and non-www, HTTP and HTTPS), because the action is reported against the property that was reviewed.
Step 2: open the Security issues report, in the same menu. A hacked site, injected spam pages or malware warnings show up there, not under manual actions, and they can cause a sharp drop by themselves.
Step 3: line up the drop with Google's update history. Open the Google Search Status Dashboard, which lists confirmed ranking updates (core updates, spam updates and others) with start and end dates. Then open the Search Console Performance report, set the date range to 16 months, and compare the day clicks and impressions started falling with those dates. A drop that begins inside a confirmed rollout window, with no manual action, is very likely algorithmic.
Step 4: rule out self-inflicted causes. If the drop matches no update and there is no manual action, check deploy history for that date, the Page indexing report for a spike in "Excluded by 'noindex' tag" or "Blocked by robots.txt", and whether the fall is in Search Console data or only in analytics (which points to tracking, not rankings). Our ranking drop analysis guide covers this step in detail, and Google Search Console explains the reports used here.
Diagnosis table: symptom, likely cause, where to check
| Symptom | Likely cause | Where to check |
|---|---|---|
| Message in Search Console, pages or whole site gone for brand queries | Manual action (major spam problems, site reputation, thin content) | Security & Manual Actions → Manual actions |
| Gradual or sudden drop across many pages, starting during a confirmed update | Core update or other ranking system reassessment | Search Status Dashboard dates vs Performance report |
| Drop during a spam update, site relies on bought or exchanged links | Link spam neutralized by spam systems | Search Status Dashboard, Links report, your link-building records |
| Strange pages indexed in a language or topic you do not publish | Hacked site or injected spam | Security issues report, site: search, server logs |
| Overnight drop to near zero on one date, no message | Accidental noindex, robots.txt block, broken migration | Page indexing report, URL Inspection on key pages, deploy log |
| One section (e.g. a coupon or partner directory) collapses, rest of site fine | Site reputation or partial manual action | Manual actions report (affected URL patterns), section-level Performance filter |
| Analytics shows a drop but Search Console clicks are flat | Tracking or consent change, not rankings | Compare GSC clicks with GA4 organic sessions by day |
| Impressions steady, clicks down | SERP change: AI Overviews, more ads or features above you | Performance report CTR by query; look at the live results page |
Types of Google manual actions: cause and fix
Google's Manual actions report help page lists each action type with its own fix. The labels below follow that page. Read the exact wording of your action in the report: it is the checklist the reviewer will use.
| Manual action | What triggered it | Core fix |
|---|---|---|
| Unnatural links to your site | Paid links, link schemes, large-scale exchanges, PBNs pointing at you | Remove what you can, disavow the rest, document it |
| Unnatural links from your site | Selling links or outbound link schemes | Remove the links or add rel="sponsored" / rel="nofollow" |
| Thin content with little or no added value | Scraped, auto-generated, doorway or thin affiliate pages | Delete, merge or substantially rewrite the affected pages |
| Major spam problems | Aggressive or repeated spam techniques, such as scaled content abuse or cloaking (many SEO guides still call this "pure spam") | Bring the whole site into line with the spam policies; show examples of removed bad content and added good content |
| User-generated spam | Spam posts in forums, comments or profiles | Clean up, add moderation and anti-spam controls |
| Site abused with third-party spam | Spam left by visitors or third parties in forums, guestbooks or internal search pages | Remove the spam, block spam terms, stop spam pages from being created or indexed |
| Spammy free host | A large share of sites on a hosting service are spam | For the host: remove spam accounts, tighten sign-up |
| Cloaking and/or sneaky redirects | Different content to Googlebot than to users, or deceptive redirects | Serve the same content to both; remove the redirects |
| Cloaked images | Images shown differently in Google image results than on your site | Show users and Google the same images |
| Sneaky mobile redirects | Mobile users sent to a different page than Google sees, often by an ad script | Find and remove the redirecting script or ad code |
| Hidden text and/or keyword stuffing | Text hidden from users, or pages stuffed with repeated terms | Remove the hidden text, rewrite stuffed copy |
| Structured data issue | Markup that misrepresents the page or breaks the structured data guidelines | Fix or remove the markup so it matches visible content |
| Site reputation policy | Third-party pages published to exploit the host site's ranking signals | Remove the content, noindex it, or move it off the site |
| AMP content mismatch | The AMP page and its canonical page show different content | Make both versions show the same content |
| Back button hijacking | Code that interferes with browser history so users cannot go back | Remove the code that causes it |
| News and Discover policy violations | Content that breaks Google News or Discover content policies | Remove or fix the content; this affects News and Discover features |
Scope matters. Google's help page says the affected URL patterns can cover part of a site or all of it, and that fixing only some affected pages will not earn a partial return to search results. For third-party spam and site reputation actions, only the pages with that content are affected.
Manual action recovery, step by step
1. Export the evidence. Screenshot the manual action, note whether it affects all pages or only some URL patterns, and save the sample URLs. Those samples are examples, not a complete list.
2. Find every instance, not just the samples. Crawl the whole site and filter for the pattern. For thin content, list every page under the same template or folder. For user-generated spam, search the database for outbound links and spam keywords. For unnatural links, export links from the Search Console Links report plus at least one third-party backlink tool, because Search Console shows a sample. A full SEO audit is the most reliable way to find the whole pattern.
3. Fix at the source. Remove the plugin that injects hidden links, close the open registration that lets spammers create profiles, stop the link-building contract. A reviewer who sees the cause still active will reject the request.
4. Fix the pages. Deleted pages should return 404 or 410. Merged pages should 301 to the page that now covers the topic. Rewritten pages need to be useful on their own, not lightly edited.
5. Verify. Use URL Inspection on a sample of fixed pages and request indexing so Google sees the current version. For cloaking, use the live test's rendered HTML to confirm Googlebot sees what users see.
6. Submit a reconsideration request. In the Manual actions report, select Request Review once every affected page is fixed. You submit one request per property, covering all of its actions.
7. Wait for the reply. Google emails you to confirm it received the request and again when the review is complete, and the status appears in Search Console. The action is either revoked or kept with a note. If it is kept, read the note, fix what remains and submit again. Google asks you not to resubmit while a request is still pending.
How to write a reconsideration request
Google's Manual actions report help page asks for three things: explain the exact quality issue on your site, describe the steps you took to fix it, and document the outcome. A request that only says "we fixed everything, please review" rarely succeeds. A good request contains:
The issue, in Google's words. Name the action exactly as the report shows it, the affected sections, and how the problem came about.
Every fix, with numbers. How many pages were removed, merged or rewritten, how many links were removed or disavowed, and what code or plugin was taken out.
Evidence. A link to a shared spreadsheet with the full URL lists, before and after examples, and for link actions, records of removal requests: date, linking domain, contact used and outcome.
Prevention. What changed in your process: who now reviews templates, how user content is moderated, which agency contract ended.
Ownership. Take responsibility. A contractor's work is still your site, and blaming one does not help the review.
A structure that works:
Issue
Our site received a "Thin content with little or no added value" action
affecting /deals/ and /city/ pages. These were auto-generated location
pages with near-identical copy, created between 2024 and 2025.
What we did
- Removed 3,412 auto-generated /city/ pages; they now return 410.
- Merged 186 /deals/ pages into 24 category pages with original
comparisons and pricing tables; old URLs 301 to the new pages.
- Disabled the page generator in our CMS and added an editorial
review step before any new template is published.
Evidence
- List of removed and merged URLs: [link to shared spreadsheet]
- Before/after examples: [3 sample URLs]
Going forward
Content owner: [name, role]. All new templates are reviewed against
Google's spam policies before launch.Keep it factual and short, and link to the full URL list rather than pasting thousands of lines. The numbers in the example are illustrative; use your own.
Link cleanup and the disavow tool
Google's "Disavow links to your site" help page is clear that most sites do not need to disavow anything. Its systems already ignore most spammy links. Use the disavow tool only when both of these are true: you have a significant number of spammy, artificial or low-quality links pointing to your site, and those links have caused a manual action or are likely to cause one (for example, you or a previous agency bought links).
Do not disavow because a tool gave a domain a low "toxicity" score, because of random scraper links you did not create, or as a reaction to a core update drop. Disavowing good links removes value you were getting. For links you did not build, see our guide to negative SEO.
The cleanup order: classify every linking domain (natural, paid, exchanged, network, scraper), request removal for the paid and network links where you have a contact, record the attempt, then disavow what is left. For a manual action, removal attempts matter, because the reviewer wants to see the links actually addressed, not just disavowed.
The disavow file is a plain text file, UTF-8 or 7-bit ASCII, one entry per line. Use domain: to disavow a whole domain, a full URL to disavow one page, and # for comments:
# Paid links from 2024 campaign, removal requested 2026-09-02, no reply
domain:cheap-seo-links.example
domain:pbn-network-one.example
# Single pages on otherwise legitimate sites
https://blog.example.org/sponsored-post-2024-best-tools/
https://news.example.net/partner/12345Google's limits are 2 MB and 100,000 lines per file. Upload it in the disavow links tool for the matching property. Each upload replaces the previous file, so always edit and re-upload your master list rather than uploading only new entries. The tool takes effect as Google recrawls the linking pages, which takes time. Our disavowing backlinks guide covers the classification step in more detail.
How long manual action recovery takes
Give stakeholders ranges, not dates. Google publishes only a few statements about timing, and this table sticks to them.
| Situation | What Google says |
|---|---|
| Reconsideration review | The Manual actions report help page says most reviews can take several days or weeks, and link-related requests can take longer |
| While a request is pending | Do not resubmit before you receive a final decision |
| After a manual action is revoked | No fixed time is given. Pages become eligible again as Google recrawls them, but rankings return only as far as the remaining site deserves; traffic built on removed links or pages does not come back |
| Core update drop | Google's guidance on core updates says some improvement can show between updates, but larger changes may not appear until a later core update |
| Spam update drop | Google says it can take months for its systems to recognize the improvements |
Anyone who promises a fixed recovery date is guessing. The review time is Google's; the quality of the fix is yours.
Recovering from an algorithmic drop (core, helpful content, spam updates)
There is nothing to "lift" after a core update. No reconsideration request exists for algorithmic changes, and Google has said core update drops often do not mean anything is wrong: other pages were judged more useful. Since March 2024 the signals once called the helpful content system are part of Google's core ranking systems, so treat helpful-content-era drops the same way.
1. Find what lost. In the Performance report, compare the four weeks after the update with the four weeks before, by page. Export it and sort by click loss. Usually a few templates or sections account for most of the drop.
2. Compare against what now ranks. For the top 20 lost queries, look at the pages that moved above you. Note what they have that yours lacks: first-hand testing, original data, a direct answer early in the page, clearer authorship, fewer ads, more current information.
3. Run a content quality audit. For each affected page, use the questions from Google's guidance on creating helpful, reliable, people-first content. Does the page offer original information or analysis? Would a reader leave satisfied, or need to search again? Is it clear who wrote it and why they are qualified? Was it written mainly to rank? Score each page keep, improve, merge or remove.
4. Act on the whole site, not five pages. Google describes site-wide signals as part of the assessment, so a large mass of unhelpful pages can weigh on the good ones. Remove or noindex pages that serve no searcher, merge overlapping articles (see avoiding duplicate content), and substantially improve the pages that matter.
5. For spam updates, the cause is usually a practice: link schemes, scaled low-value content, expired domain abuse. Stop the practice. If links were neutralized, the ranking value they gave is gone, and disavowing will not bring it back. Recovery means earning rankings without them.
Our core updates recovery guide goes deeper on the audit, and Quality and E-E-A-T covers the standards the audit measures against.
How to avoid Google manual actions
Every manual action maps to a section of Google's "Spam policies for Google web search", part of Google Search Essentials. Read them once, then build these habits:
Do not buy or trade links. Mark paid, sponsored and affiliate links with rel="sponsored", and ask any agency for a list of every link it builds. Earn links instead; our link building guide covers how.
Review templates before they scale. Any feature that creates hundreds of pages (city pages, tag pages, programmatic comparisons) needs an editor to check that each page is useful on its own.
Moderate user content. Use moderation queues, spam filters and rel="ugc" on user links, and keep internal search result pages out of the index.
Vet third-party content. Coupon sections, sponsored directories and partner content hosted on your domain are where site reputation actions start. Ask whether the content would exist without your domain's ranking signals.
Keep markup honest. Structured data must describe what is visible on the page.
Audit ad and script code. Sneaky mobile redirects and back button hijacking often come from third-party ad or tracking scripts, not your own code.
Patch and monitor. Keep the CMS and plugins updated and check the Security issues report, because hacked pages can lead to spam problems you did not create.
When to hire help for penalty recovery
You can handle many manual actions in-house: a structured data issue, a batch of user-generated spam, or a cloaking problem caused by one script. Consider outside help when the action is "Unnatural links to your site" with thousands of linking domains, when a site-wide action hits a site that earns most of its revenue from search, when a request has already been rejected once, or when nobody on the team knows what a previous agency did.
A good penalty recovery specialist or agency will start by confirming whether you have a manual action at all, show you the full list of affected pages or links, explain each fix in plain language, and write the reconsideration request with you. Be wary of anyone who guarantees a recovery date, disavows links based only on a tool's toxicity score, or suggests moving to a new domain. AI tools can speed up link classification and drafting, but a person still has to check the evidence and stand behind the request.
Penalties and AI answers
Google's AI Overviews and AI Mode draw on Google's own index. A page removed for major spam problems, or a section removed for site reputation abuse, cannot be cited there either. ChatGPT search, Perplexity and Copilot use their own crawlers and search providers, so a Google manual action does not remove you from them directly. But the pages that trigger thin content actions are the same pages AI engines skip: they rarely contain a specific, quotable fact. After a recovery, make sure AI crawlers can reach the rebuilt pages (see AI search technical optimization), then check which of your pages these engines cite for your core questions, using our AI visibility tracker. The AI visibility guide covers what earns mentions outside classic results.
Mistakes to avoid
Disavowing in a panic. A core update drop is not a link problem, and a broad disavow can remove links that were helping.
Fixing only the sample URLs. The samples in the report are examples. Reviewers check for the pattern across the site.
Submitting a vague request. "We fixed the issues" without what, how and evidence is a common reason requests are rejected.
Redirecting penalized pages to good ones. 301-ing thin or spammy pages into your best pages moves the problem; delete them instead.
Moving to a new domain to escape. Redirecting a penalized domain can carry the problem along, and starting fresh throws away every legitimate signal you had.
Blaming a penalty for a tracking break. Always confirm the drop in Search Console clicks before changing anything on the site.