Content refreshes: what to update and what to delete

Publishing is more satisfying than maintaining, which is why almost every site has a backlog of pages nobody has looked at since the day they went live.

It is also usually the worse investment. A page sitting at position twelve has already been crawled, indexed, understood and given a provisional verdict. Getting it to position six requires improving something specific. A brand-new page starts from nothing and needs months before it even receives a verdict.

The instinct to publish more is strong and often wrong. The problem is that "refresh your content" is advice without a method — which pages, changed how, and when is deletion the better answer? This is that method.

First, sort the library

Every page falls into one of five states. Sort before deciding anything, because the right action differs completely.

| State | Signal | Action | |---|---|---| | Winning | Ranking well, getting clicks | Leave alone. Check annually | | Close | Positions ~5–20, impressions, few clicks | Improve. Best return available | | Seen but ignored | Impressions, position 20–50 | Diagnose: usually intent or depth | | Invisible | Indexed, near-zero impressions for 6+ months | Merge or delete | | Duplicative | Two pages competing for one query | Merge |

Most of the available gain is in the second row, and most of the wasted effort goes into the fourth.

The pages worth improving

Close but not converting attention. A page with impressions and a position in the teens is being shown and not chosen. Two causes, in order of likelihood: the title and description do not earn the click, or the page answers the query less completely than the ones above it.

Fix the title first — it is fifteen minutes and it is measurable. Then compare the page honestly against the current top five. Not against your memory of them; open them today.

Pages that have drifted out of date. Prices, versions, screenshots, references to things that have changed, advice that was true three years ago. Decay is invisible to you and obvious to a reader.

Pages where the intent has moved. The results page for a query changes over time. A page written when a query wanted an explainer will slide when the query starts wanting a comparison. Re-read the results — the method is in how to read a search results page.

What a real refresh involves

Changing the date is not a refresh, and it fools nobody worth fooling.

A genuine one usually means:

• Answering the question more completely than the pages currently above it
• Adding what only you know — your own numbers, cases, constraints
• Removing what has stopped being true, which often means deleting more than you add
• Restructuring for the shape the query now wants — if every competitor leads with a table and you lead with three paragraphs of preamble, that is the finding
• Fixing the title and description to match what the page actually delivers
• Relinking it from newer pages, which is the step almost everyone skips

That last one matters more than its effort suggests. An older page linked from nothing you have published since is drifting; a couple of contextual links from newer work re-establishes it.

Refreshing without breaking what already worked

The under-discussed risk. A page ranking at position six is ranking for reasons, and some of those reasons are things you are about to remove without noticing.

Four rules that prevent the common damage.

Do not change the URL. By far the biggest self-inflicted loss. If the content is changing but the page is still about the same thing, the URL stays. A "tidier" slug is not worth the disruption.

Check what it currently ranks for before you edit. Open Search Console, filter to that page, and read the queries. Frequently a page earns most of its impressions from something you had forgotten was in it — a definition, a table, an aside answering a question you did not think was the point. Removing that section in the name of tightening is how a refresh loses traffic.

Add before you cut. If in doubt, expand the thin parts rather than deleting the parts you find boring. You can always trim next quarter with better information.

Change one page at a time when the page matters. If you refresh thirty pages in a week and traffic falls, you will not know which change did it. On a page producing real enquiries, make the change, wait a month, then look.

Keep a note of what you changed and when. Ten seconds. It is the difference between "traffic dropped in April, no idea why" and "traffic dropped after the April rewrite, put the comparison table back".

How to tell whether the refresh worked

Set the measurement up before you edit, or you will be guessing.

Record three things on the day: average position for the page's main query, total impressions over the previous 28 days, and total clicks over the same period. Then look again at four weeks and eight weeks.

What the outcomes mean:

• Impressions up, position up. It worked. Leave it alone now.
• Impressions flat, clicks up. The title or description did the work, not the content. Useful to know — it means the same fix is available cheaply on other pages.
• Impressions up, clicks flat. You are being shown more and chosen no more often. The content improved and the title did not.
• Everything down after eight weeks. Something the page was ranking on has gone. Compare against the old version and put back what you removed.
• Nothing moves at all. The page's problem was never the content — it is competitive position, or the query is not one you can win. Stop refreshing it.

That last outcome is the most common on pages people refresh repeatedly, and recognising it saves the third and fourth attempt.

When deleting is the right answer

This is where most advice stops short, because deletion feels like admitting waste.

Delete or merge when a page:

• has had near-zero impressions for six months or more, and you cannot name the query it should win
• duplicates another page's intent — one of them has to go, and it should be the weaker
• was written for a query that no longer has meaningful demand
• is so thin that improving it means rewriting it entirely — in which case write the new one and redirect
• exists only because someone once said to publish weekly

The reasoning is not that pages are penalised for existing. It is simpler: a library where a third of the pages are dead is harder for you to maintain, harder to link coherently, and dilutes the internal linking that would otherwise concentrate on pages that can win.

Google's guidance on helpful content is explicit that producing content mainly to attract search traffic is not the goal. A page nobody has ever wanted is that page, and keeping it does not improve with time.

Delete, redirect or merge?

| Situation | Do this | |---|---| | Thin page, no links, no traffic, nothing similar | Delete. Let it return 404 | | Thin page, but it has inbound links | Redirect to the closest genuinely relevant page | | Two pages competing for one query | Merge the best of both into the stronger URL, redirect the other | | Outdated but the topic still has demand | Rewrite in place, keep the URL | | Seasonal or event page, now past | Keep if it recurs; redirect if it does not |

Do not redirect everything to the homepage. A redirect to something irrelevant is a worse experience than a 404, and it is treated accordingly. If there is no relevant destination, let it 404 honestly.

The one thing to check before deleting anything

Look at whether the page has inbound links from other sites. A forgotten page with three genuine external links is not worthless — it is an asset with the wrong content on it. Rewrite it or redirect it; do not throw the links away.

Similarly, check what links to it internally. Deleting a page that four others reference leaves four broken links, which is a self-inflicted version of the problem you were trying to fix.

What to expect afterwards

Honest expectations, because pruning makes people nervous:

A short dip is normal. Removing pages removes their impressions, however small. The total looks worse before it looks better.

Improvements to "close" pages move fastest. Weeks rather than months, because the page is already understood.

Deleting genuinely dead pages changes almost nothing measurable, which is the point — they were producing nothing. The gain is in maintainability and in link concentration, not a traffic spike.

Do not do it all at once. Prune in batches, so that if something unexpected happens you know which change caused it.

A schedule that actually gets done

Quarterly, an hour:

1. List pages with impressions but positions 5–20. Pick the three with the most impressions. Improve those.
2. List pages with near-zero impressions for six months. Decide: merge, redirect, or delete.
3. Check the three pages that matter most for factual decay.
4. Add two internal links from recent work to older pages that deserve them.

That is four actions, and it outperforms publishing four more articles on most established sites.

When publishing is still the right answer

Refreshing beats publishing only once you have covered a subject. If you have eight articles on a subject that has forty real questions in it, the gap is coverage, and no amount of polishing eight pages fills it. The test is in how many articles before SEO starts working — the answer is about coverage, not count.

And if the honest reading is that the queries you need are not winnable from where you stand, neither refreshing nor publishing solves that. Judging whether you can realistically win a keyword is the check, and the rankings board shows what the alternative costs.

Common questions

How often should I refresh? When something has changed or a page is close and stuck. Not on a calendar.

Does updating the date help? Not on its own. Changing the content is what matters; the date should follow real changes.

Will deleting pages hurt my rankings? Deleting pages that rank will. Deleting pages that do not rank, and never did, will not.

Should I redirect or 404 a deleted page? Redirect if there is a genuinely relevant destination or the page has links. Otherwise 404 is the honest answer.