Owner, trigger and scope
The content owner decides that the page goes and gives the reason. The SEO lead approves the response the address will give and signs off the tests. The web publisher runs the procedure: they control the content management system, the redirect rules and the sitemap, and they coordinate the channel owners. Each channel owner (paid media, email, social, partners, print) stops sending people to the address in their own channel and confirms when it is done.
The trigger is a live page being retired: what it describes no longer exists, the question it answers is no longer asked, or it cannot be made accurate at reasonable cost, and no other page answers the same question. Retiring is one of the decisions the content governance guide makes in a content audit, and this procedure is how it is carried out on the live site.
Neighbouring jobs that look similar belong elsewhere:
| Situation | Where it is handled |
|---|---|
| The page's content moves to a new address, or two pages merge | Changing a page URL SOP |
| Listings, jobs, events or offers that end by design | Expired listings, events and offers, as a rule built into the templates |
| A discontinued product with a direct successor | The category and product page guide redirects it, through the URL change SOP |
| A branch closes or moves | Moving or closing a location |
| Many pages go at once in a restructure or a new site | SEO for site migrations with the redirect map template |
An error on a live page starts with the live ad and page error response SOP; tag, author and date archives follow the index rules for archive pages; the SOP library routes the rest, from publishing a web page to error triage.
Before you start: inputs and access
| Input | Owner | Notes |
|---|---|---|
| The retire decision, its reason and its approver | Content owner | A sentence on why, from the inventory row, the product change or the review |
| What still points at the page | Web publisher, with the SEO lead | Internal links from a crawl, links from other sites, recent search clicks |
| The places outside the site that use the address | Channel owners | Ads, the product feed, email templates, automated journeys, profiles, partner pages, print |
| Access to the site | Web publisher | The content management system, the redirect rules and the sitemap |
| Search Console access | SEO lead | For the property that holds the page |
| The content inventory and the list of intended removals | Content owner | Where the decision, the response and the date are recorded |
The Chesterton's fence framework is the lens: remove nothing until someone knows why it is there and who relies on it. The status codeThe numeric code a web server returns with every response, saying whether the request succeeded, was redirected or failed and why. is the visible part of the job; most of the work is the checking before it.
The procedure
Confirm the retire decision, the reason and who approved it
Write down which page goes, why, and who decided. The content governance guide's test applies: retire a page when nobody asks its question now, when what it describes has gone, or when making it accurate would cost more than it is worth. A fixable page is corrected instead, as the updating and correcting guide advises, and thin pagesPages that offer little original value of their own, such as near-empty templates, near-duplicates or copies with only a word or place changed. that together would make one useful page are consolidated through the URL change SOP.
Check what still points at the page before anything changes
The guide to updating and correcting content sets the rule: if other sites link to the page, searchers still reach it or a near replacement exists, it redirects to whichever page now answers the question; if none of those holds, it is removed. Crawl the site for every Internal LinkA link from one page to another page on the same website. to the address, check a backlink tool and the server logs for other sites linking to it, and read its recent clicks in Search ConsoleGoogle's search performance and indexing tool.. The link reclamation tutorial asks for each linked address's destination to be decided before content is deleted. Then ask the content owner whether any live page now answers the same question.
Choose the response, or hand a page with a successor to the URL change procedure
If a true equivalent exists, this is a move: run the changing a page URL SOP, which writes a one-hop permanent redirectA permanent server redirect telling browsers and search engines a URL has moved for good., and stop here. Google's redirects documentation, as the expired pages guide reads it, takes a permanent redirect as signalling strongly that the target should become the canonical page, which is honest only when the target is the same thing. If readers relied on the page, keep a short notice at the address linking to the best current answer; a discontinued product follows the product page guide's table, which keeps the page briefly with a notice and alternatives. Otherwise the address answers not found or gone. Never leave an empty page answering success, and never redirect to the home page.
Find every place outside the site that sends people to the address
List every channel that uses the address the way the tutorial on withdrawing an asset from every channel lists every use of a claim: name the places, then confirm each one. Ask paid media, email, social, partner and print owners to search their own tools, including email templates, automated journeys, pinned posts, affiliate pages and QR codes on print. The web publisher holds the list and gives each line an owner.
Repoint or pause ads, feed items, emails and automations first
Clear the channels before the page changes, so nobody pays to send visitors to a dead address. The Google Ads website migration guide lists the places an address sits inside an account: the final URL of every ad, paused ones included, link assets, Performance Max asset groups and URL exclusions, tracking templates, the feed's product links and thank-you addresses in conversion triggers. Paused ads belong on the list because they are often re-enabled later. Give each ad a destination that matches its message, or pause it, and take the product out of the feed. Then change email templates and automated journeys, which keep sending people to a dead page long after everyone has forgotten it.
Apply the response at the address
The crawling and indexing guide gives the standard for a permanent removal: answer not found or gone and take out the internal links, and never redirect to the home page. The expired pages guide adds that, by its reading of Google's crawl budget guide, not found or gone is the correct response for a page removed for good and a strong signal that the address need not be crawled again. For a notice page, follow the same guide's advice on keeping an ended page useful: the notice and its date at the top, a link to the current answer or alternatives, no buy or apply buttons, and no markup still describing the item as available.
Remove internal links and take the address out of the sitemap the same day
Fix every link the site controls at source. The internal linking guide lists links to deleted pages among the faults any crawl turns up; repoint or remove them in navigation, footers, templates, related-content blocks and body copy. Take the address out of the sitemapA file listing the URLs a site wants search engines to discover and crawl.: the expired pages guide asks for removed addresses to leave it the day they go, and a notice page kept out of search leaves it too.
Test the old address and the pages that used to link to it
Request the old address and confirm it answers with the chosen response; for a notice page, the SEO page checker reads its head and directives quickly. Rule out the causes in the soft not-found diagnosis: an error page answering success, a stripped-down product page, a redirect to an irrelevant page. Crawl the pages that used to link to the address and click through from an ad preview and a test email. A not-found event, built as the tutorial on tracking errors and failed actions describes, shows which links visitors are still following.
Record the removal in the inventory and the list of intended removals
The content inventory template holds the decision and its reason, the destination for retired pages (the surviving address, or "removed", with the date the change went live) and a change record with who approved it. Add the address to the list of intended removals with its date, which the not-found diagnosis asks for so that its entries can be marked as expected without a second investigation. If the page carried campaign traffic or a conversion trigger, add an entry to the measurement change log.
Read the not-found report and close the removal
At the next reading of the Page Indexing report, find the address. The not-found diagnosis sorts entries into groups, among them expected (deliberately removed, with no links left), leaking (the site's own links or sitemap still name it) and losing (other sites or past searchers still arrive). An expected entry closes the removal. A leaking entry goes back to the links step. A losing entry is decided again, as the link reclamation tutorial sets out: a destination for the address or a request to the linking site. A soft not-foundA page that returns a success status code but shows an error, empty or not-found message. goes to the Search Console error triage SOP.
A model railway shop retires discontinued kit pages (illustrative)
Nine kit pages and a layout guide retired in one pass
Plimtrax Model Railways, an invented model railway shop, learned that its supplier had stopped making 9 kits, and its content review flagged a layout guide written for a track system the shop no longer sells. The check found that 1 kit had a direct successor, so its page went to the URL change procedure. The layout guide had 23 links from clubs and forums and about 140 search clicks a month, so it kept a short notice linking to the current guide. The other 8 kit pages had no outside links and fewer than 10 clicks each in the past quarter; they now answer gone. Before anything changed, paid media repointed 2 shopping ads and removed 8 feed items, and the email owner replaced a kit link in the welcome email. The web publisher removed 31 internal links, updated the sitemap the same afternoon and added the addresses to the list of intended removals. When the not-found report next listed them, the SEO lead marked all 8 as expected in a few minutes.
QA checks
Before the removal is closed, the SEO lead confirms each check. A failed check is fixed where it started and tested again, never hidden behind a redirect.
- The old address gives the chosen response: a notice page, not found or gone, never success on an empty page or a redirect to the home page.
- No internal link, navigation item, related-content block or sitemap entry still names the address.
- No ad, feed item, email template, automated journey, profile or partner page still sends people to the address.
- Links from other sites and recent search clicks were checked, and a page with a true equivalent went to the URL change procedure instead.
- The content inventory and the list of intended removals show the address, the date, the reason, the response and who approved it.
Hand-offs and records
Paid media receives the destination or pause decision for every ad and feed item. The email, social and partner owners receive the address with a request to confirm it is gone from their channel. The content owner receives the updated inventory row, so the next audit does not reopen the same decision. The SEO lead keeps the address on watch until the not-found report shows it as expected.
The lasting record is a dated row in the content inventory and the list of intended removals naming the address, the reason, the response chosen, who decided and who approved it, the channels cleared and the test results, with a measurement change log entry when the page carried campaign traffic or tracking.
Common mistakes
- Deleting before checking. A page with links from other sites or steady search visits deserves a destination or a notice; find out who relies on it first.
- Redirecting everything to the home page. It commonly counts as a soft not-found and helps nobody.
- Leaving an empty page answering success. A page stripped of its content is a common cause of soft errors.
- Removing the page but not the links. Internal links and sitemap entries keep sending visitors and crawlers to a dead end.
- Forgetting paused ads, the feed and the welcome email. They outlive everyone's memory of the page.
- Treating a move as a removal. If another page answers the same question, redirect to it through the URL change procedure.
- Keeping no record. An unexplained not-found entry is investigated again every time the report is read.
