Removing a page from the results
The removals tool acts within a day but lasts about six months: it is a reprieve, not a removal. Three ways make it permanent, and only two of them also close off access. Robots.txt is not among them, and it can even prevent a removal from working.
An emergency and a decision, not to be confused
Making a page disappear from the results covers two different needs, and treating them as one is the source of most failures. There is the emergency — this page must not show today — and the decision — this page must never show again. The documentation answers both, with two distinct tools.
Requests made in the removals tool last for about six months; the tool takes a page out of the results within a day, but a lasting removal requires another action.
officialSource : Google Search Central, “Remove a page hosted on your site from Google”
Six months is neither short nor permanent, and that is exactly the trap. A team that handles the emergency in a day closes the file, and the page returns six months later, when nobody connects the two events any more.
The removals request and the permanent measure belong together rather than one after the other: the first holds the six months during which the second must be crawled to take effect.
hypothesis
That is our reading, and it follows from the previous one: the duration of the request only makes sense if it covers the time needed for the permanent measure to be crawled. The documentation does not prescribe that order, it gives both durations; we draw the practical consequence.
The permanent measures, and what each really closes
Three ways are described to block a page permanently: remove or update the content, password-protect the page, or add a noindex rule. The first is presented as the most secure, because it also holds for engines that might not respect that rule.
officialSource : Google Search Central, “Remove a page hosted on your site from Google”
The order proposed is not technical, it is logical. Removing the content makes it inaccessible to everyone; password-protecting the page makes it inaccessible to those without the right to see it; adding a noindex rule leaves it accessible and merely asks that it not be displayed.
The third is therefore the only one of the three that rests on a third party’s cooperation. That is enough for a page you would rather not see in a search engine, and not enough for information that should no longer circulate.
Confidential page published by mistake? Block it in robots.txt and Google will drop it at the next crawl.
The trap of two rules that cancel each other
The documentation explicitly says not to use robots.txt as a way to block a page from the results, and points to the limitations of that file.
officialSource : Google Search Central, “Remove a page hosted on your site from Google”
For a noindex rule to be effective, the page must not be blocked by robots.txt and must remain accessible to the crawler: without crawling the rule is never seen, and the page can still appear if other sites link to it.
officialSource : Google Search Central, “Block Search indexing with noindex”
Here is the combination that fails most often, and it fails silently. Disallowing the crawl of an address while placing a no-indexing rule on it amounts to writing an instruction behind a locked door. The crawler does not enter, does not read, and the address can keep appearing if other sites link to it.
Two details that sink successful removals
The documentation asks that every variation of the URL leading to the same content be handled, since different addresses can point to the same page.
officialSource : Google Search Central, “Remove a page hosted on your site from Google”
The example the documentation gives says it well: the same page can answer at a lowercase address, at the same one capitalised, and at a third carrying a parameter. Handling the first and forgetting the other two leaves the feeling of a job done.
The noindex rule is written either in a meta tag or in an HTTP response header named X-Robots-Tag, with the same effect; the header serves non-HTML resources such as a PDF or an image.
officialSource : Google Search Central, “Block Search indexing with noindex”
That answers the question that always comes next: a downloadable document cannot carry a meta tag, having no HTML head section. The response header settles the case, and it applies the same way to images and videos.
Filing the removal request and the permanent measure together
hypothesis
- Effort
- low
- Time to effect
- immediate
- Risk
- low
- Durability
- lasting
The effort is low: a request in the console and a line of configuration. The delay is the shortest on this site, the documentation stating less than a day for the temporary part. The risk deserves one notch more than the other cards, because a badly targeted rule removes pages you meant to keep. Durability depends on the measure chosen: good for content taken down, conditional for a rule that depends on regular crawling.
What does not work, or no longer works
Blocking in robots.txt to make a page disappear. The documentation says so explicitly, and that file is itself public.
Then adding a noindex without checking that the page remains crawlable. The two rules neutralise each other and the failure produces no signal at all.
Finally, treating the removal request as the end of the matter. It lasts about six months, long enough that nobody remembers the cause by the time the page reappears.
Sources
- Remove a page hosted on your site from Google — Google Search Centralaccessed on
- Block Search indexing with noindex — Google Search Centralaccessed on
Block revalidated on
Frequently asked questions
- How long does a removal request last?
- About six months, according to the documentation. That is a processing window, not a solution: past that point the page comes back if nothing else changed. The tool therefore buys time, during which the permanent measure has to be put in place and crawled.
- Is robots.txt enough to make a page disappear?
- No, and the documentation explicitly says not to use it for that. Worse, it produces the opposite effect when combined with a noindex rule: since the crawler is not allowed to fetch the page, it never reads the rule, and the address can stay visible if other sites link to it.
- What is the difference between removing the content and adding noindex?
- A noindex rule only prevents display in the results of an engine that respects it. Removing the content or password-protecting the page also closes off access, including for engines that ignore the rule and for people who know the address. The documentation presents removal as the most secure.
- How do you remove a PDF or an image?
- Through the X-Robots-Tag HTTP response header, which carries the same rule as a meta tag and applies to resources that are not HTML pages. The tag itself needs an HTML document to exist in, which rules out most downloadable files.
- Does each address need handling separately?
- Yes, and it is the most common oversight. Several addresses can lead to the same content, with or without parameters, with different capitalisation. The documentation asks that all variations be protected or removed, otherwise the content stays reachable through the one left behind.