How to update old blog posts without replacing them

An old article is not a writing problem. It is a decision that has been put off. The page still answers a real question, the URL has accumulated whatever it has, and the only thing genuinely wrong with it is that some of what it says is no longer true. Replacing it with a new page is the most expensive answer to that problem.

The cheaper answer is to update the page you already have: same address, same history, a claim that has been brought back up to date. This is how to refresh outdated content without replacing it: what to change first, and which changes are not worth the risk.

Age is not a quality score

In our experience, “our blog needs refreshing” means somebody noticed a publication date rather than a defect. A four-year-old page that answers a stable question and answers it correctly is finished work. A page published last month that recommends a tool which no longer exists is not. The date is a fact about the page, not a judgement about it, and treating the two as the same thing produces refreshes that change wording and leave the wrong claim in place.

What the refresh is looking for is drift: a claim overtaken by a change in the world, a section made redundant by another page on the same site, a recommendation that has quietly become bad advice. Drift is what content decay means in practice, and it has no publication date. It appears whenever whatever the page described changed, which is why a page can be old and accurate, or recent and wrong.

A review interval is not a rewrite deadline

Schedules still earn their place, provided they are review schedules rather than rewrite schedules. Nielsen Norman Group’s content-strategy guidance, in its maintenance section, is where the interval comes from: it suggests you start with at least a 3-month interval for reviewing content and supporting collateral. That is a suggested starting cadence, not a rule about how often text must change and not a deadline. The same section asks you to determine the content’s fate, then decide whether to improve, replace or remove low-performing content; leaving a sound page unchanged stays our call. If every review ends in a rewrite, the interval is doing the work of a quota.

How to find the pages worth updating before you rewrite anything

Sort by consequence rather than by age. In practice three groups repay the work first. Pages that still receive visitors and now contain something false: those are the urgent ones, because the page is actively delivering the wrong answer. Pages that were written for a question the business no longer asks: these are merge candidates rather than update candidates, because they are not wrong, they are just no longer needed in that form. And pages whose ranking subject has changed, where the query people use has moved on and the page has not.

Everything else can wait. This is the sorting exercise a content audit that ends in decisions performs at scale; a refresh is the update column being worked, in whatever order that list produces.

Comparing two date ranges, calmly

Search Console’s performance report is the free evidence for this, described in Google’s own documentation as a way to see how your search traffic changes over time, where it’s coming from, and what search queries are most likely to show your site. Reading the same period this year and last year side by side is enough to tell you whether a page is holding, slipping or simply seasonal. It will not tell you why, and a line going down is not evidence that the text is bad. A dip is a prompt to open the page, not a diagnosis.

Update the claim, not the adjectives

The first edit is the one the page was written for: the central claim, restated so that it is true today. If the page argues that a certain approach is the cheapest way to do something, and the economics have changed, the argument changes — not the sentence length. Everything after that is subordinate to it.

After the claim, the specifics that decay: prices, dates, version numbers, named tools, job titles, links to documents that have moved, screenshots of interfaces that no longer look like that. Then the shape of the argument, if the shape is wrong: a page that buries its answer under three paragraphs of context was wrong when it was published, and a refresh is the cheapest moment to fix it. Only then the prose. Adjectives are the least valuable thing in an article and the first thing a nervous refresh touches, which is how an afternoon of updating content can leave the reader with exactly the same article.

Keep the URL: what a new address costs

Publishing the update under a new URL discards the one asset the page has: the address that has been shared, bookmarked, cited and indexed, none of which transfers on its own. If a new address is genuinely unavoidable (the subject has changed, the page is being split, the URL was a mistake to begin with), the move needs a permanent redirect from the old address to the new one, not a temporary one: Google’s documented choice is a permanent redirect, used when you are sure the change will not be reverted. When a page cannot be updated and cannot be merged, the fate of its address is the next decision, and it is a bigger one than it looks.

Re-source the claims, or the refresh is decoration

A refresh that leaves the sources untouched is a rewrite of the wrapper. If the page cites a study, open it and check that it still says what the page says it says. If it links to another site, check that the page is still there. If it makes a claim on the authority of an organisation, check that the organisation still makes it. This is where refreshes collapse: it is slower than rewriting sentences, and it occasionally means deleting a paragraph the whole page was built around.

Where the page’s subject is guidance rather than opinion, the refresh has a model to copy. GOV.UK’s publishing guidance says that guidance should usually be updated rather than withdrawn, and it treats withdrawing and unpublishing as separate operations with different consequences. That is GOV.UK’s house rule for its own publishing process, not a requirement imposed on anyone else’s site; what travels is the habit of preferring the update and being explicit about the alternative. An article with a named author and a stated last-reviewed date tells the reader something honest about how it is maintained, which a page of polished sentences with no provenance cannot.

What a refresh should never do

It should not be sold as a ranking technique. Google’s documentation on core updates is unusually direct about this: there’s no guarantee that changes you make to your website will result in noticeable impact in search results. The same document advises against quick fixes and recommends improving content in meaningful ways, which is the order a refresh follows anyway.

Asked how to update old blog posts for SEO, the honest answer is narrow. An accurate page is better than an inaccurate one. A page that answers its question in the first screen is better than one that does not. Neither statement comes with an outcome attached, and any agency promising that a refresh will recover a specific position is describing a result nobody controls. The work is still worth doing, because a page that is correct, clear and current is doing its job whether or not a graph moves.

A refresh you can finish in one sitting

Pick one article. Read it as a stranger. Write down the claim it makes, the specifics that have decayed, and the sources it leans on. If the claim is still true, change only what has decayed and stop; a page does not need to be rewritten to have been refreshed. If the claim is dead, the page needs a merge or a retirement rather than an edit, and that decision is worth making deliberately instead of by publishing something new beside it.

Senatore writes for companies whose published pages have to keep working after the launch: articles, guides and documentation that get maintained rather than abandoned. That is what the articles service covers, and if the question is whether a specific page should be updated or replaced, the five questions a brief has to answer are where the answer becomes obvious. Send Senatore the page you think has gone quiet, in an enquiry on the content-audit topic, and we will tell you which of the two it is.

---

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.

Articles that get maintained rather than abandoned · the five questions a brief has to answer

Back to the blog · More in SEO & content