First, confirm the decline is gradual and spread across many pages
Open the Performance report in Search Console and compare a long recent period with the same period a year earlier, filtered to the blog or resource section. Then switch to the pages view and sort by the change in clicks. Decay has a particular shape: many pages each losing a little, with no clear starting day. If instead the line falls off a cliff on a date, or a whole template disappears at once, the cause is a technical change, an indexing fault or a ranking update, and the rankings dropped diagnosis is the right page.
For the decaying pages, record three things from the same report: impressions, average position and click-through rate, each for the main query the page was written for. These three numbers separate the causes. Impressions falling while position holds means fewer people are searching. Position falling while impressions for the query hold means other pages are now preferred, either competitors' or your own. Clicks falling while both hold means the results page itself changed around the listing. The content marketing troubleshooting hub covers the other content symptoms if none of these fits.
Diagnosis table
Diagnosis
| Symptom | Check | Likely cause | Fix |
|---|---|---|---|
| Slow decline on articles that cite prices, rules, product features or years | When each page was last reviewed and whether its facts are still true | Out-of-date content, no review cycle | Refresh under the same address; set review dates |
| Position for a query wobbles; two of your URLs take turns in the results | Pages receiving impressions for the same query, per period | Newer posts competing with the original | Consolidate or differentiate the overlapping pages |
| Impressions fall while average position holds | Query impressions over a long period; demand estimates for the topic | Search demand shrinking | Re-weight effort toward topics with live demand |
| Position slides while impressions for the query hold | Who now ranks above you and what their pages offer | Competitors with fuller or fresher pages | Answer the question better, with current evidence |
| Older articles earn fewer clicks and are rarely crawled | Internal links pointing to each old article, then and now | Archive no longer linked from new content | Re-link old pages into their topic cluster |
Fix, verify and prevent each cause
Facts, examples and dates that went stale
Every article starts ageing on the day it is published. Prices move, rules change, products are renamed and screenshots stop matching the interface. Readers and search systems both notice: Google's guidance on helpful content asks whether content is accurate and trustworthy, and a page with a dated claim fails that test quietly.
Fix: re-read each decaying page against current sources, update facts, examples, links and screenshots, and record the review, keeping the same address, as the refresh step in the content governance guide describes; the guide to updating and correcting published content covers the updated date and the note that tell returning readers what changed.
Verify: over the following weeks the refreshed page's clicks for its main query stop falling and begin to recover, compared with untouched pages of similar age.
Prevent: give every page an owner and a next review date in the inventory, and review pages with time-sensitive claims on a shorter cycle.
Newer posts competing with the original
Teams that publish steadily often write a fresh post on a subject the archive already covers, and then a third. Each splits the signals the original had built, a pattern called Keyword CannibalisationTwo or more pages on one site competing for the same query and intent, splitting signals..
Fix: use the query-level evidence and the consolidate, differentiate or remove decision set out in the keyword cannibalisation guide, then point internal links for the query at the surviving page.
Verify: one URL per query receives most impressions in the next period, and its position steadies.
Prevent: before briefing a new piece, search the inventory for pages answering the same question and refresh the existing one instead of starting another.
Fewer people searching for the topic
Some topics fade: a tool falls out of use, a regulation is replaced, a trend passes. The page may still be the best answer and simply have fewer people asking.
Fix: confirm the fall in demand with the query sources the keyword research guide lists, then move refresh and new-content effort toward clusters with live demand; keep the page if it still serves customers, merge or retire it if it does not.
Verify: the traffic lost on fading topics is offset by gains on the clusters you re-weighted toward over the following quarter.
Prevent: revisit the keyword map each quarter so that the calendar follows current demand rather than last year's.
Competitors who answered the question better
A page that ranked well years ago may now face pages that are more complete, more current or more closely matched to what searchers want.
Fix: read the pages now ranking above yours, list what they cover that yours does not, and look at what the results page rewards, using the search and content checks in the competitor analysis guide; then rewrite for the current Search IntentThe goal behind a search query: to learn, to find a site, to compare or to act., adding the first-hand evidence and specifics a competitor cannot copy.
Verify: the page's average position for its main query improves in the following months while impressions for the query stay level.
Prevent: keep the monthly monitoring of rankings on the terms that matter, so a competitor's improvement is noticed while the gap is small.
Older articles dropped out of the link structure
New articles are linked from the homepage, the newsletter and each other; old ones slowly lose those links as listings move on and nobody adds contextual links back. An article with few internal linksA link from one page to another page on the same website. is harder for readers to find and is crawled less often.
Fix: restore the hub-and-supporting pattern described in the topic clusters guide, so that each older article is linked from its hub and from at least a couple of siblings, and add links from new pieces to the archive pages they build on.
Verify: a site crawl shows each re-linked article with more inlinks, and its impressions recover over the following weeks.
Prevent: make links to existing archive pages a required item in every brief and in the publishing check.
A prevention routine against content decay
Decay is the default state of an archive, so the defence has to be a standing routine rather than an occasional clean-up. Monthly, list the pages whose clicks fell over a sustained period, read their impressions, position and click-through for the main query, and assign each to one of the five causes above; the cause decides the fix. Feed those pages into the same backlog as new work, so that refreshes compete honestly with new pieces instead of waiting in a separate list nobody reads. Before any new brief is approved, search the inventory for an existing page on the question. Quarterly, check that every page has an owner and a review date, re-run the demand check for each pillar, re-read the pages that now outrank yours on the queries that matter most, and crawl the archive for articles with too few inlinks. Finally, keep a note of what was refreshed and when, so that the next review can see whether the refresh worked or whether the page should be merged or retired instead.
A travel insurer's guides slide for a year
A travel insurer's 120 destination guides earned about 54,000 search clicks a month at their peak and 39,000 a year later, with no single page losing more than a fifth of its traffic. Search Console showed three groups. For 31 guides, position had slipped while impressions held; they quoted entry rules and fees from two years earlier. For 14, a newer "checklist" post on the same country now took turns with the guide in the results. For 9, impressions had fallen at a steady position: the destinations had simply become less popular. The team refreshed the 31 stale guides, merged the 14 checklists into their guides with redirects, and moved new-writing effort from the 9 fading destinations to 6 growing ones. Six months later clicks were back to about 50,000 a month.
Before committing to a refresh programme, the content refresh test measures whether refreshed articles recover relative to decaying articles left untouched, which separates the fix from natural recovery.
Common mistakes
Changing dates without changing content. Writing a new post instead of refreshing the old one. Treating a dated drop as decay. Refreshing pages in alphabetical order rather than by lost traffic. Keeping pages alive for topics nobody searches for. Linking only new articles from new articles.
