Content pruning: should you delete old blog posts?

Deleting an article feels like a decision. It is the last step of a decision nobody took. The page has been quiet for a year, nobody wants to defend it, and something that had earned a position is removed to settle an argument about effort.

Deletion is not an outcome of its own: it is what is left when a page is retired and its address is allowed to fail. This piece is about the disposition half of content pruning: which fate a page has earned, and what its address does next.

Four fates for a page, and a separate decision about its address

Keep the page on its current address. Update it on the same address. Merge it into a page that already answers the question. Retire it, which means stopping the maintenance. Those four are the fates used across this cluster, and the address is a second decision: it stays, it permanently redirects to the nearest surviving page, or it is left to fail when nothing relevant survives. Pruning gets described as the deletion exercise, which is why it wastes so much work. Deleting an article is retiring a page and letting the address fail, the only combination that leaves nothing in place.

They are not equally easy to undo: a merge can be undone by publishing again, and a retirement can be reversed while a copy of the page exists somewhere you control, whether that is your CMS version history, a repository or an archive. Keeping a copy is what makes the harder fates arguable rather than final.

Google’s own position: deletion is a last resort

You do not have to take that on trust. Google’s documentation on its core updates sets out the order itself: deleting content is a last resort, and only to be considered if you think the content can’t be salvaged, and the bullets above it recommend improving content in meaningful ways rather than quick fixes. The same bullet carries the qualification this piece depends on: “if you’re considering deleting entire sections of your site, that’s likely a sign those sections were created for search engines first, and not people. If that’s the case for your site, then deleting the unhelpful content can help the good content on your site perform better.” So the page that calls deletion a last resort also says when removal is right: where the section exists for search engines rather than readers.

The editorial reading

The editorial version is simpler. A page that answers a question somebody still asks is an asset, and a 2023 publication date is a fact about your calendar rather than about the page. Deleting it destroys the answer and the address at once, which is why deletion should require the most argument in its favour.

The three questions that decide a page’s fate

Work through them in this order. Does the page still answer a question somebody asks now, and can its claim be made true? The free traffic evidence matters more than the publication date. If both answers are yes, the fate is keep when the claim is already true and update when an edit would make it true; an update is not a deletion argument with extra steps. If either answer is no, does another page on the site genuinely answer that question well? If it does, the fate is merge: move what the survivor would have said anyway into it and point the old address at the survivor. If nothing else answers it, there is nowhere to merge and nothing to redirect to, so the fate is keep or update while the claim can be made true, and retirement without an irrelevant redirect when it cannot. A redirect to an unrelated page pretends the address still has a destination.

Retiring a page and letting its address fail is right only when nothing survives to answer the question and the claim cannot be salvaged. This is where a content audit that ends in decisions actually lets pages leave: a retire column that gets read properly shrinks before it is worked.

What the address does next: keep, redirect or fail

Keeping a page means keeping its address, the cheapest route, with no statement to anybody. Merging moves the content into the survivor, and the old address then points at it. Retiring with a relevant survivor points the address at the nearest page answering the same question, so the route keeps working while the maintenance stops; without one there is nothing to point at, and the page stays unless the claim cannot be salvaged. Google’s guidance on the mechanism is unambiguous: we recommend that you use a permanent server-side redirect whenever possible, and it adds that a permanent redirect is the right choice when you are sure it will not be reverted. A temporary redirect says the move is provisional and invites the world to keep the old address; if you are not sure the change will hold, that is information about the decision.

404 vs 410: what each status tells the world

The HTTP standard defines both precisely, and the difference between “not here” and “gone” matters. A 410 (Gone) says that access to the resource is no longer available at the origin server and that the condition is likely to be permanent; a 404 claims nothing about the future, which is why the standard prefers 410 when the server knows the condition is permanent and treats 404 as the right answer when it cannot know. Google’s own status-code documentation adds the operational half: If a URL was previously used but is now returning 4xx status code, Google systems will stop using the URL over time, and in Google Search a URL that is already indexed and returns a 4xx status code is removed from the index.

The homepage is not a redirect target

The easy shortcut is to point every removed article at the home page. It satisfies the letter of the instruction and fails the purpose: the reader arrives somewhere that does not answer their question, the surviving page gains nothing relevant, and the retirement becomes a deletion with better manners. Point the address at the nearest surviving page on the same subject, even when it is short; if none exists, that absence is the answer rather than a gap: the fate is keep or update while the claim can be made true, and retirement with no relevant redirect when it cannot.

Where a page is retired because it is dangerous rather than redundant (a price that no longer applies, a service that has been withdrawn), the mechanism changes. Google’s documentation on removing information notes that requests made in the Removals tool last for about 6 months, and that keeping a page out of the index for good means changing the page itself: removing the content, password-protecting it or adding a noindex tag. A six-month measure is not a retirement.

Merging in practice: one page, one claim

A merge is not a staple. Joining two arguments produces a longer page that answers both questions badly. The surviving page should be able to state one claim in one sentence; if it needs two sentences joined by “and”, the merge has not been thought through. Move only what the survivor would have said anyway, in the survivor’s voice, and keep the removed page’s address on the survivor by redirect. Where the page is merely wrong rather than redundant, the merge is unnecessary: the update path should be tried first, because it costs less and keeps more.

Write down what you changed, before you change it

Every retirement and every deletion produces a record: the address, the fate, the destination it points at, the date, and the person who decided. It is the only defence against the same page being deleted twice by two different people, against a redirect target changed by somebody who does not know why, and against the library being rebuilt from memory. Nobody enjoys being the person who says a piece of work is finished; that reluctance is the subject of the decision nobody wants to make.

Senatore writes and maintains page libraries for companies that publish under their own name, which means making these calls rather than listing them. The SEO and content service includes the written note on what to refresh and what to retire. Send Senatore the one page you are least sure about, with its address, in an enquiry on the content-audit topic: the free read covers one existing page, and seeing how the three questions land on your own page is the fastest way to judge the rest. No outcome is guaranteed in either direction: these are judged calls, and what a search engine does with a page afterwards is not something anyone can promise.

---

Written by Senatore. This is Senatore’s own writing about craft; it contains no client claims, and every figure it quotes is linked to its source.

The written note on what to refresh and what to retire · the decision nobody wants to make

Back to the blog · More in SEO & content