Content decay in 2026: how to audit, refresh, or retire aging posts before they drag your rankings and AI citations down

Most content strategy is aimed at the next post. Almost none of it is aimed at the last fifty. That’s backwards — the majority of a mature site’s organic traffic comes from posts published months or years ago, and every one of them is decaying on a clock you’re probably not watching. This post is the audit we run against our own archive: how to find the pages that are quietly losing ground, and what to do about each one.

Read how SEO actually works in 2026 for the foundation this sits on, and Generative Engine Optimization for why freshness now matters to citations, not just rankings.

What content decay actually is

Content decay is the gradual loss of rankings, traffic, or AI citations on a page that hasn’t itself changed — because the world around it did. A competitor published something more current. A statistic you cited went stale. Search intent for the query shifted. Or an AI engine started preferring a more recently updated source for the same claim.

It’s easy to miss because nothing “breaks.” The page still loads, the copy still reads fine, and unless you’re watching the trendline, a slow bleed from position 4 to position 11 over four months looks identical to normal traffic noise.

The majority of a blog’s traffic comes from posts published more than six months ago. If you’re only investing in new content, you’re maintaining a shrinking share of what’s actually earning you visits.

The three signals that flag a decaying page

Run these three checks and you’ll find almost every decaying page on your site:

Impressions holding, clicks falling. In Search Console, a page still gets served for its queries but the click-through rate is sliding. This usually means a competitor out-freshened you, or an AI Overview or featured snippet is now answering the query before anyone reaches your result. Check the query list under that page — if a “People also ask” or AI Overview snippet now sits above you for the exact phrase, that’s your answer.

Both impressions and clicks falling on a query that hasn’t changed. This is a ranking drop, not a SERP-feature problem. Pull the position history for that page’s top 5 queries. A steady decline over 60-90 days, with no algorithm update in that window, usually means the content itself has gone stale relative to what’s now ranking above it.

AI engines stopped citing a page that used to get cited. Re-run your flagship queries through ChatGPT, Perplexity, and Gemini every quarter (see AI visibility tracking tools if you want this automated). A page that dropped out of citations without any Google ranking change usually means the content aged out on the freshness signal specifically — the underlying claim is still correct, but a more recently dated source is now the one being lifted.

The quarterly pull

Once a quarter, pull three reports and cross-reference them — this takes about an hour for a site our size:

1. Search Console: Pages, sorted by clicks, last 12 months vs. previous 12 months. Flag anything down more than 20% year over year that isn’t explained by seasonality.
2. Search Console: Queries for each flagged page. Is the query volume itself declining (intent shifted, not worth chasing), or is the query stable and you’ve just lost position (worth fixing)?
3. GA4: Engagement rate and average engagement time on the flagged pages. A page losing search traffic AND losing engagement on the visits it still gets is a weaker fix candidate than one where the content is still doing its job for the people who land on it.

Anything that clears all three checks — traffic down, query volume stable, engagement still healthy — goes on the refresh list. Everything else goes into the decision tree below.

Refresh, consolidate, or retire

Not every decaying page deserves a rewrite. Sort each flagged page into one of three buckets:

Refresh when the topic is still live, the query volume is stable or growing, and the page’s core structure is sound but the specifics (numbers, examples, screenshots, tool recommendations) have aged. This is most of your list. Update the actual claims, don’t just touch the metadata — more on that below.

Consolidate when you have two or three thin, overlapping posts competing with each other for the same query instead of one strong one. Merge the best material into the strongest post, 301-redirect the others, and let internal links that pointed at the retired posts point at the survivor instead. Cannibalization is a common cause of “decay” that looks like a ranking problem but is actually a duplication problem.

Retire when the topic itself is dead — a deprecated tool, a discontinued platform feature, an event that already happened. Don’t leave it live to decay quietly; either redirect it to the closest living page or, if nothing fits, let it 410 and drop it from the sitemap. A dead page still costs you crawl budget and can drag down a topic cluster’s average quality signal.

What an actual refresh should touch

The single most common refresh mistake: bumping the modified date without changing anything the reader or the AI engine would notice. It doesn’t work, and it’s easy for both systems to tell the difference between a real edit and a timestamp-only touch-up.

A refresh that actually moves the needle changes:

• The numbers. Any statistic older than 12-18 months gets re-sourced or removed. Stale numbers are the single fastest way to lose an AI citation to a competitor’s more recent post.
• The examples and screenshots. If a tool’s UI changed or a recommendation is no longer current, update it — this is often the fastest way to lose reader trust even when the underlying advice is still right.
• The structure, if it predates your current formatting standard. Older posts written before you adopted question-shaped H2s or FAQ schema are worth restructuring, not just re-dated — see how to write a blog post that ranks AND gets cited for the current bar.
• Internal links. Add links to anything you’ve published since that’s now relevant, and check that every outbound internal link still resolves — a broken internal link on a page you’re trying to revive undercuts the whole exercise.

Only after the substance changes do you update dateModified in schema and the visible “updated” date. The date is a record of the edit, not a substitute for one.

What we run for clients

We run this audit quarterly against our own archive — over 50 posts at this point — and it’s usually the highest-ROI hour on the content calendar, ahead of anything net-new. For clients, we fold it into the ongoing content engagement: a quarterly decay pull, a refresh/consolidate/retire pass, and a re-check of AI citation status on anything we touch.

If your site has content that used to perform and doesn’t anymore, tell us what you’re working on.

FAQ

How often should I audit content for decay? Quarterly is the right cadence for most sites — frequent enough to catch a ranking slide before it compounds, infrequent enough to stay sustainable alongside net-new publishing. Most teams that skip audits entirely only notice decay once traffic has already dropped substantially, which makes recovery slower.

Does updating the “last modified” date actually help rankings or AI citations? Only if the date reflects a real edit. Search engines and AI engines both weight freshness as a signal for how current and trustworthy a claim is, but a modified date with no corresponding content change is easy to detect as stale and, over time, erodes trust in your dates generally. Change the substance first — statistics, examples, structure — then update the date.

What’s the difference between content decay and keyword cannibalization? Decay is a single page losing ground to the outside world — a competitor, a stale fact, a shifted intent. Cannibalization is two or more of your own pages competing for the same query and splitting your ranking signal between them. They often look identical in a traffic report, which is why the quarterly pull checks query-level detail, not just page-level totals, before deciding whether to refresh or consolidate.

Is it better to refresh an old post or write a new one on the same topic? Refresh, in almost every case. A page with existing backlinks, indexation history, and some accumulated authority will out-rank a brand-new page on the same topic faster than that new page can catch up — and splitting the topic across two competing URLs invites the cannibalization problem above. Write new content for genuinely new topics; refresh for topics you’ve already covered.

Related reading on this site:

• How SEO actually works in 2026 — the internal-linking and technical foundation this audit protects
• GA4 for GEO: measure the AI-search traffic GA4 hides — the measurement layer for catching decay in AI channels specifically
• AI visibility tracking tools — automating the quarterly citation re-check
• How to write a blog post that ranks AND gets cited — the structural bar a refresh should bring an old post up to