Deletion is a fact about a server. De-indexing is a fact about a card in a search engine. Australians collapse the two because both make a URL “go away”, and because the companies that sell reputation work have an interest in keeping the words interchangeable. They are not interchangeable. A publisher can delete a post and Google will keep printing the title for weeks. Google can de-index a URL and the Age will still have the article. If you remember one distinction from this piece, make it that one.
The complaint is always the same. Someone has used the CMS trash button, or a solicitor’s letter that actually worked. They search their name in an incognito window from Sydney. The result is still there, often with a stale snippet. They write to “Google support” as if an editor has refused a correction. Nobody has refused. The crawler has not been given a signal it trusts, or it has been given three that contradict each other. The rest of the search and archives hub is about copies that survive a clean takedown. This piece is about the live URL, and the five ways people try to tell Google it no longer counts.
What Google is actually looking at
Google does not sit inside your WordPress install. It requests a URL, reads the HTTP status, reads the HTML, and compares that visit with the last one. The Google Search removal guide covers the forms you file when you do not control the site. When you do control it, the machinery is cruder and more reliable than a form. Five instruments get used, usually in the wrong order.
404 Not Found. The server says the resource is not here. That is a soft promise. Sites return 404s during outages; Google has learned to wait. A true HTTP 404 will drop the URL, but not in an afternoon. A soft 404 — the article is gone, the server returns 200, and a theme prints “page not found” — is worse than leaving the article up. The crawler still has a live document. It just has a less useful one.
410 Gone. The server says the resource is not here and will not be back. This is the status the documentation has favoured for years when you mean it. We have watched 410s clear a results page faster than 404s on the same host, particularly where the sitemap has also dropped the path. It is not a spell. It is a clearer sentence.
noindex. A meta robots tag, or an X-Robots-Tag header, that tells crawlers not to keep the URL in the index. The page can still be fetched by anyone with the link. Use it when the URL must stay reachable — a login wall, a legal archive — but not in Search. Do not combine it with a robots.txt block. The crawler may never see the tag.
robots.txt Disallow. This is the most expensive mistake in the file. robots.txt asks well-behaved crawlers to skip a path. It is not a deletion order and it is not a de-index order. Google can keep listing a blocked URL as a skeleton result because it saw the address on another page. Blocking the crawl also blinds Google to your noindex. People who “robots.txt the scandal” often freeze the result in place.
The removal form. There are several, and they are not interchangeable. Search Console’s temporary removals tool hides a URL on a property you have verified, for about six months, while you finish the 410. The outdated-content tool asks Google to recrawl a snippet that no longer matches. The personal-information and legal products, mapped in the removal guide, hide a URL you do not control. None of them delete the file. All of them fail when the live response is still a 200 with the old text.
Why the blue link is still there
When a result survives a “successful” cleanup, the cause is usually one of these, and often two.
The page still returns 200. A plugin rebuilt a public preview. A CDN is serving a stale node. A redirect was pointed at the homepage, which tells Google the old URL became the homepage and can transfer the old title into sitelinks. We treat a 301 of a damaging URL onto a live section as a new publication, not as a deletion.
The sitemap still lists the path. Leaving a deleted article in sitemap.xml is an instruction to come back. Internal “related posts”, a tag archive, an author page — each is a recrawl invitation. You do not have to hunt every backlink on the internet. You do have to clean the ones on the same host.
The removal was query-specific. A personal-information decision can stop a URL appearing for a name search and still leave it standing for a more precise query, or for an overseas index. Search is not one drawer.
Or the leftover is not the live result at all. It is a cache, a sitelink, or a Knowledge Panel still drawing on a paragraph you have not touched. That is a different job: cached copies, sitelinks and Knowledge Panels.
A working order that does not fight itself
If you control the host, use this sequence.
- Capture the public result and the live URL, including any cache URL, in an incognito window from Australia.
- Decide whether the page must remain reachable. If it must, apply noindex and keep it out of the sitemap. Do not Disallow it.
- If it must not remain reachable, delete the content and serve a hard 410 on that exact path. Do not 301 it to the homepage. Do not serve a branded error page that returns 200.
- Remove the URL from every sitemap you submit in Search Console. Request a recrawl. Purge the CDN and confirm the status code from two networks.
- Only then use Search Console’s temporary removal if the result is still being clicked this week. It is a bandage. The 410 is the work.
- If you do not control the host, stop. You are in the removal-form business, not the status-code business.
Site owners who skip to the hide and leave a 200 in place are paying for a six-month bandage.
What a de-index is not
A de-indexed URL can still be opened by anyone who kept the link, or who finds it in an email or a PDF. It can still be archived, summarised, or sitting in Bing. Treating a clean Google result as the end of the matter is how people discover, six months later, that a recruiter used a different engine. It is also not a finding that the article was false. Google’s personal-information products hide contact details. They do not adjudicate a news story. Do not write to a complainant as if Google has “cleared” them. It has not.
The Australian overlay
Nothing in the status-code chapter is unique to Australia. Two things around it are. First, the result set is regional. A URL de-indexed for google.com.au can still appear on google.com for a searcher who is not in this country, or who is on a VPN. If the harm is an overseas recruiter, say so in any legal cover note and attach both screenshots. Second, a court order that tells a publisher to take a page down is not, by itself, an order that tells Google to de-index the URL. You still file the legal removal, with the order and the exact paths, as the Search removal guide describes. Solicitors who serve the publisher and stop there leave the card in place. If an eSafety notice has landed and the host has complied, your job at Google is freshness: 410, sitemap, outdated-content.
When each instrument fails
404 fails when it is soft, when a CDN disagrees, or when the CMS resurrects a slug. 410 fails when a platform front-end always returns 200. noindex fails when robots.txt or a conflicting header overwrites the head. robots.txt fails because it was never a de-index tool. The form fails when the page is still live and the complaint is “I do not like this article.” In that last case the work moves to the publisher, to a concerns notice, to the platform that hosts the thread. De-indexing is not a substitute for a source takedown. It is the step that stops the corpse appearing on the first page of a name search after the source has been dealt with.
Agencies that promise to “wipe the URL from the internet” by filing a Search Console hide are selling a bandage as surgery. There is no shame in paying someone to check status codes. There is shame in believing the receipt is the deletion.
If you want the desk to look at a general question about signals and leftovers, use contact. Do not send the court file or the leak. We publish method. We do not run your case.