Content Decay: Why Old Posts Lose Traffic

Content does not usually fail loudly. The page that ranked well two years ago does not disappear in a week, it slides. A few positions in month three, a handful of clicks gone by month six, and by the following year it brings in a fraction of what it used to while everything on the site still works.

I have watched Australian businesses lose a large share of their organic traffic this way, without a single technical fault anywhere on the site. When somebody asks me to look at it, the answer is usually decay rather than damage.

This is written for the owner or marketer who has noticed a page, or a section of the site, quietly sliding, and wants to know whether it is decay or something else, and what to do about it. It covers what decay looks like in the data, the reports that show it early, why old posts lose visibility when nothing broke, a refresh that fits into an afternoon, and the point where merging or retiring a page is the better answer. This is about slow decline. It is not about sudden drops and it does not speculate about algorithm updates.

What content decay actually looks like

A slow slide, not a sudden drop

Content decay is the gradual loss of rankings, clicks and conversions on a page nobody changed. The page still loads, the content is still there, and the results get weaker month by month.

That pattern is what separates it from the other problems that look similar in a chart. A sharp fall across many pages at once is a different investigation, usually a technical fault or a change in how the site is being read. Decay affects one page or one cluster, and it moves slowly enough that nobody notices until the loss is large.

When I look at a declining page, the first question is not what changed on the site. It is what changed in the market around it.

Why the cause is usually several things

Decay is rarely one cause. It is usually three or four small ones arriving together: the query drifting towards a different intent, a competitor publishing something more current, a page that still describes a service the way it worked in 2022, and a set of internal links that have moved elsewhere.

That is why single fixes rarely work. Rewriting a title will not recover a page whose answer is out of date, and updating the facts will not recover a page that has lost the links that used to carry it.

Detecting decay before traffic collapses

The three reports that show it early

Three reports do most of the work. In Search Console, compare the last three months against the same three months a year earlier and sort by the biggest losses in clicks. In your analytics, run the same comparison on sessions to the pages that matter. Then, for the pages that appear in both, check the average position for the main query that page was written for.

Together those give you the shape of the problem. Losing position while demand holds steady is a competitiveness problem. Holding position while clicks fall usually means the results page has changed around you. Both dropping at once is often intent drift, where the page is answering a question that is no longer the one being asked.

Choosing a decline threshold you will act on

Set a threshold and write it down. Mine is a 20% fall in clicks over a rolling quarter on a page that matters, measured against the same quarter a year earlier, with the page's main query checked before anything else.

The threshold exists so the decision is made once, calmly, rather than argued about every time somebody notices a dip. Without it, the pages that are genuinely sliding sit in a list alongside every seasonal wobble on the site.

Why old posts lose visibility

Intent drift on the queries you once owned

The most common cause is not competition. It is that the query moved. A topic people used to research with an open question becomes one they expect a direct answer to, or one they compare options on, and a page still answering the old question stops being the best result even though nothing on it is wrong.

You can watch it happen. Per Ahrefs research on content decay, "For example, last year the SERPs for "How to start a blog" showed split intent: Reddit forum posts made up 34% of page one, and "How to" beginner guides accounted for 27%." Twelve months later the same query had moved again: "But by the following year, Reddit posts made up 60% of the SERPs, while "How to" beginner guides made up 22%."

Neither result is a penalty and nothing broke on either site. The audience changed what it wanted, and the pages matching the new expectation took the space.

Competition, links and freshness signals

Competition works on the same slow timeline. A competitor publishes a more complete page, earns links to it, and over a year or two takes the position you used to hold. It rarely looks dramatic in any single month, which is why it is so easy to miss.

Links decay the same way. The article that linked to yours was updated and dropped the link, the directory that listed you changed hands, the supplier page that referenced your guide was rewritten. Nothing on your page changed, and the signals pointing at it did.

Decay the page causes itself

Stale facts, dates, prices and screenshots

Some decay is self-inflicted, and this is the part you can fix this month. A page quoting a price from 2022, a screenshot of an interface that no longer looks like that, a tool described by a name it no longer uses, or a line about last year that was written in a different year.

Readers notice before search engines do. A page that feels out of date loses the reader in the first paragraph, and the engagement that follows tells the search engine the same thing.

Broken internal links and withdrawn offers

The other self-inflicted problem is the trail around the page. Internal links pointing to a discontinued service, a CTA for an offer that no longer exists, a download that returns an error, or a page that used to be linked from the navigation and no longer is.

I check every link on a declining page by hand before recommending anything else. It takes ten minutes, and it is where the fastest recoveries come from, because some pages are not decaying at all. They are simply pointing at things that no longer work.

A refresh process for one afternoon

What to update first on a declining page

Work in this order: the answer, the evidence, the details, then the packaging. First make sure the page answers the question people ask today, and answers it in the first screen rather than in paragraph nine. Then update the facts, the figures and the dates. Then correct the details, including prices, tool names and any feature that has changed. Only then look at the title and the description.

Titles and descriptions deserve attention even so, because they are frequently rewritten anyway. Per the Web Almanac 2024 SEO chapter, "One study showed that 71% of meta descriptions are rewritten for the first page of results." Treat them as an accurate description of the page rather than as text you fully control.

Log what you changed and when, in one line. A refresh with no record cannot be judged later, and the next person to look at the page will redo the same work.

What to deliberately leave alone

Leave the URL alone unless there is a strong reason to change it. Leave the parts of the page that are still accurate and still read well. And leave the publication date alone unless the substance has actually changed, because a new date over the same content is a cosmetic move that helps nobody.

Resist adding sections while you are in there. A page that grows every time somebody opens it becomes harder to read and no more useful, and the refresh stops being a maintenance job.

When to merge, redirect or retire a page

Merging two posts that split the same demand

Merge when two pages answer the same question for the same reader and neither has the weight to hold a position alone. Combining them gives the stronger page more to work with, and the redirect preserves whatever the weaker one earned.

Do not merge pages that serve different intents, even where the topics overlap. If one page is for somebody researching and the other is for somebody ready to buy, they are two pages doing two jobs, and collapsing them loses one of those jobs.

Retiring a page with no demand left

Retire a page when the demand has gone and nothing points at it. A discontinued product line, a service no longer offered, a guide for a tool the business stopped using. Check the backlinks first, then redirect to the nearest relevant page rather than letting it return a 404.

A page that is weak, rather than dead, is usually worth refreshing instead. Retirement is for demand that has left, not for performance that has slipped.

Decay and AI answers: the freshness question

Why assistants lean on current pages

Search assistants have added a freshness pressure that did not exist a few years ago. When an answer is assembled from sources, a page with current dates, current figures and a clear statement of what it covers has an advantage over one that leaves the reader guessing whether it is still accurate.

The difference shows in the sources that get picked. Per Ahrefs research on content decay, "Our own research shows that URLs cited by AI assistants are 25.7% "fresher" than organic SERP results on average." Fresher means more recently updated, and it is the one lever a page owner fully controls.

Making dates and facts verifiable

Make it easy for a reader, and for a system reading on their behalf, to tell when something was last checked. State the period a figure comes from, name the source, and keep the update log where it can be seen rather than buried in a content management system.

None of that guarantees a mention in an answer. It removes the most common reason a good page gets passed over, which is that nobody can tell whether it is still true.

Setting a baseline so the fix can be proven

Numbers to record before you edit

Before you touch a word, record four numbers: clicks and impressions for the main query over the last 28 days, average position, total organic entrances to the page, and the enquiries or conversions that page contributed. Export them, and keep them beside the refresh log.

Without a baseline, a refresh cannot be judged. Most disagreements about whether content work succeeded are really disagreements about what the page was doing before anybody started.

A quarterly decay routine

Then run the same routine every quarter and keep it small. Compare the last three months against the same three months a year earlier, pull the pages that fell past your threshold, check the ten worst by hand, and decide for each one whether it needs a refresh, a merge, a retirement or nothing at all.

That last option is deliberate. Some dips are seasonal and some are noise. The routine is worth running precisely because it lets you conclude that a page needs no work and move on.

The same discipline applies to which sources get picked for AI answers at all, and how AI answers choose their sources covers that side of it. When you want the quarterly pass run by somebody who does it every week, working with an SEO consultant is one route, and where the upkeep needs to sit inside a wider program, content marketing services cover it. I keep the working examples across the rest of the writing on this site.

Frequently asked questions

How do I tell content decay apart from a penalty?

Decay is gradual and usually affects one page or one topic cluster while the rest of the site holds. A penalty tends to be sharp, affects many pages at once, and shows up in Search Console. If the drop is sudden and across the whole site, that is a different problem from the slow decline covered above.

Is a drop in traffic always decay?

No. Seasonality, a competitor publishing something better, a change in what people search for, and a shorter or busier results page can all move a page down. Decay is when the page has genuinely become less able to answer the query as it is searched today.

How often should old posts be updated?

I would not touch them on a calendar. Review the pages that are declining, and the pages that carry facts which can go stale, such as pricing, legislation, tool features and years. A yearly sweep of your best performing content, plus a quarterly look at anything showing decline, is a workable rhythm.

Does refreshing a page mean changing the publication date?

Changing the date without changing anything else is a cosmetic move and it does nothing for the reader. If you update the substance, record what changed and when, and use a date that reflects the last meaningful update rather than a date that flatters the page.

Should we delete old posts that no longer get traffic?

Only after you check what they earn. A page with no traffic can still hold links, answer a question people ask you on the phone, or support a topic cluster. Retire it when the demand has gone and nothing points at it, and redirect it to the nearest relevant page rather than letting it return a 404.

What is the smallest useful decay review?

Open your analytics, compare the last three months against the same three months a year earlier, sort by the biggest traffic losses, and look at the top ten pages by hand. That single pass finds most of the decay that matters on a small site.