Content Decay: Find Articles Losing Traffic

Most blog dashboards flatter their owners. The total line holds steady because new posts keep arriving, while a quieter story plays out underneath: older articles slip a few places, lose a slice of their clicks, or keep appearing in results without anyone choosing them. Nobody files a ticket for a slow leak, so it goes unnoticed until the archive that once carried the blog has stopped carrying much at all.

This guide is about finding those articles with evidence rather than instinct. It covers how to build a watchlist from Search Console, how to cross-check it in GA4, how to separate genuine decline from seasonality and site-level problems, and how to turn each candidate into a recorded decision. It does not teach you how to rewrite an article from scratch, because that is a different craft. Any cut-off that matters has to come from your own data and your own capacity to act on it.

Where the Term Comes From and Where It Stops

"Content decay" is an industry label. Google does not document it as a concept, and its guidance talks instead about search traffic drops and about helpful, people-first content. That matters for how you use the word. If you treat it as an official diagnosis, every decline gets the same cure. If you treat it as a working label, you keep the freedom to ask what actually happened.

A definition you can test

Decay means a gradual decline in the organic clicks or impressions of a single article over several months, with no obvious external event that explains it. Three conditions keep the definition honest. The unit is one article, not the whole site. The period is months, not days. And the cause is not yet explained by an outage, a migration, a security warning or a manual action. If any of those apply, solve that first.

What a decline can look like without being decay

Google's guide to debugging search traffic drops groups causes into a handful of families: algorithmic updates, technical problems, security issues, manual actions, seasonality and shifting interest, and site moves or restructures. Each one can draw a curve that resembles slow decay on a single article. A holiday guide peaks and troughs on a calendar. A product that fades from the news shrinks the demand for the query itself. A migration reshuffles URLs while Google recrawls. Event-driven drops after a single ranking update are a separate diagnostic job, and we cover it in diagnosing traffic loss after a Google update. Here the target is the slow, article-level leak.

Why the distinction pays off

A wrong label wastes effort in both directions. Rewriting an article whose real problem is a stray noindex tag changes nothing, while ignoring an article that has quietly been overtaken by a better page lets the leak continue. The rest of this guide is organised so that cheap, system-level explanations are checked before anyone edits a word.

Set Up a Decline Watchlist

A watchlist is a standing table of articles whose visibility is moving the wrong way, with enough context to decide what to do. It starts with knowing what each tool actually measures.

What the Search Console numbers mean

The Search Console performance report (Opens in a new tab) defines clicks as the number of times a user clicked through to your site from Google Search, and impressions as how many times your site appeared in results. Click-through rate is clicks divided by impressions. Average position is the average position of the topmost result from your site, and in the table it is calculated for the URL or grouping on that row. Rows can be grouped by query, page, country, device, search appearance or date.

Two details from that documentation protect you from false alarms. First, totals in the table and the chart can differ because some dimensions are aggregated by property and others by page, so judge an article from the page-grouped table rather than from the headline chart. Second, the newest data can be preliminary and is drawn with a dotted line, so a comparison anchored on the last few days will mislead you. Exports also turn values shown as "~" or "-" into zeros, which means a zero in a spreadsheet may be missing data instead of a real count.

The fields worth keeping per article

The columns below are common industry practice and not something Google prescribes. They keep a decision explainable months later.

Choosing a cadence

A quarterly pass suits many blogs, and high-volume sites may review more often. That is practitioner habit rather than a documented rule. What matters is that the same table, the same filters and the same comparison windows are reused every cycle, so differences between cycles reflect the site and not the method. The query and page views are covered in more depth in our guide to the Search Console performance report.

Reading the Compare View Page by Page

Google's guide to debugging search traffic drops (Opens in a new tab) lays out a sequence that adapts neatly to article hunting. The steps below follow it, with the article-level adjustments added.

Start with the long view

The guide advises extending the date range to 16 months so you can see whether a drop repeats year after year. For a single article the question is simple: is the line falling in the same months every year, or has it been drifting in one direction without a calendar pattern? A repeating dip is the shape of seasonality. A one-way slope across more than a year is the shape that deserves a closer look.

Use the comparison tab on purpose

In the date filter, the Compare tab offers the last three months against the previous period, or a year-over-year view. The solid line is the recent period and the dotted line is the comparison period, and the gap between them shows the size and timing of the loss. The guide also draws a line between two symptoms. If both impressions and clicks fell, check the common causes list. If only clicks fell, the cause may lie in your titles, your snippets, or competing rich results on the results page.

Sort the page table by click difference

With a comparison active, the guide suggests sorting the Pages table by Clicks Difference to bring the biggest losers to the top. To make that table useful for decay hunting, add a filter that keeps only your article URLs, for example by path. Product pages, category pages and landing pages follow different logic and would dilute the list.

Drop into queries for each candidate

Once an article stands out, repeat the comparison on its queries. Split what you find into the main query the article was written for and the long tail around it. A fall in the main query position suggests a competition or intent story. A loss of long-tail impressions suggests the article no longer answers the neighbouring questions people ask. These are hypotheses to test in the sections below, not conclusions.

Priorities without magic numbers

There is no universal percentage that makes a decline worth acting on. Rank candidates by what the article does for the business, then by the size of its loss, and write your own working threshold into the table so that the next review applies the same rule. Fictional example: two articles lose a similar amount of traffic. One of them feeds a signup page, and the other is a loosely related opinion post. The first goes to the top of the queue, and the second can wait a cycle.

GA4 as a Cross-Check

Search Console tells you what happened in the results page. GA4 tells you what happened after the visit started. Using both guards against acting on a measurement artefact.

The landing page report

The GA4 landing page report (Opens in a new tab) lists the first page of each session along with sessions, active users, new users, key events, average engagement time per session and total revenue. Its landing page dimension is the page path and query string associated with the first page view in a session. The documentation recommends adding session source and medium as a secondary dimension, which shows which sources send people to each landing page. The report is session-scoped, so one person landing on different pages in separate sessions counts once for each page.

How to read agreement and disagreement

This pairing routine is practitioner convention. Build the decline list in Search Console, then look at the same URLs in GA4 with a source and medium split. When both tools lean the same way, the decline is real. When only one shows it, look at measurement first: a tracking change, a consent flow change, a URL format mismatch or a filter error could explain the gap. A change in average engagement time per session is a hint about whether visitors still find what they came for, though it never proves content quality on its own.

What GA4 cannot tell you

Analytics sees visits, not the results page. It cannot show you impressions you never converted into clicks, the query a person typed, or the competitor that now appears above you. That is why GA4 confirms and Search Console explains, and why neither replaces looking at the live results.

Telling Decay From Everything Else

Before changing an article, work through explanations in order of cost. The order below mirrors the structure of Google's traffic drop guidance.

Rule out the system-level causes

Technical problems come first. The guide names server availability, robots.txt fetching and not-found errors as blockers of crawling, indexing or serving, and notes that a misplaced noindex tag can cause a slower decline that looks like decay. It points to the Page indexing and Crawl stats reports for these. Security problems such as malware or phishing can trigger warnings before users reach the site, and the Security Issues report is the place to look. A manual action for spam policy violations can lower rankings or remove pages. The guide also recommends the Data Anomalies page, because logging or processing problems can produce a drop that is not real. If one of these matches, you have found a cause that no rewrite will fix.

Check the timeline against ranking updates

The guide suggests comparing the date of a drop with the list of ranking updates. It also says that small position shifts, for example from 2 to 4, often need no action, and that radical changes are best avoided on pages that already perform well. Large drops across many terms call for a review of whether the site as a whole is helpful and people-first, and effects can take days to months to appear, so a few weeks of waiting are advised before re-analysing. If a single article slides slowly and the timing does not line up with any update, the decay reading becomes more plausible.

Intent drift

What people want from a query can change. A search that once rewarded a short definition may now surface comparisons, tools or step-by-step guides. That is a widely repeated observation in the industry and you can test it yourself: run the query, look at the first results, and note their page types and formats. If the results have moved toward a different format from yours, a refresh may need to rebuild the article, not just update its facts.

Overtaken by a better page

A competitor may have published something deeper or more current. The useful response is not imitation but finding the gap your own readers would value. Google's helpful content self-assessment asks whether content offers original information, reporting, research or analysis, and whether it goes beyond the obvious. A rewrite that restates the same material in new words rarely recovers a position.

Competing with yourself

Over the years a blog often ends up with several articles aimed at one query. The older one then loses ground as the newer one gains, and the combined traffic barely moves. Check which of your own URLs appear for the primary query. The method is covered in keyword cannibalization: how to detect it, and the usual outcome is a merge.

Is the market itself shrinking?

The traffic drop guidance recommends checking Google Trends when queries lose clicks and impressions, to see whether the decline is specific to your site or part of a wider trend. If demand is falling everywhere, improving the article will not bring it back. Sometimes the right decision is not to invest, or to fold the topic into a broader piece.

The signals inside the article

Only after those checks is it worth reading the article critically. The questions in Google's helpful content guidance (Opens in a new tab) give a useful frame: does the content provide original information or analysis, does it contain easily verified factual errors, and will a reader come away feeling they learned enough to reach their goal? Practical signs of an ageing article, as industry observation, include outdated product names and prices, dead outbound links, screenshots of an old interface, an introduction written for an earlier audience and a headline built on last era's phrasing. None of these proves a cause alone, but a short list combined with the right metric pattern strengthens the case.

Dates are a record, not a lever

The same helpful content guidance warns against changing the date of pages to make them seem fresh when the content has not substantially changed, and against adding or removing large amounts of content merely to look fresh. Google's publication dates documentation (Opens in a new tab) asks for a prominent, labelled visible date, such as published or last updated, and for the visible date and structured data values (datePublished and dateModified) to match. It also says the date must describe the page's publication or update, not the events the page covers. So an old date is not a defect in itself. A new date on an unchanged article is simply a false record.

Four Outcomes for a Declining Article

Every candidate needs a decision, and there are four productive ones plus a fifth that is easy to forget: leave it alone and watch.

Refresh

A refresh keeps the URL and rebuilds the content with current information, real examples and a reviewed structure. The steps, the checklist and the republishing routine are in our guide to updating old blog posts. In your decision record, only note what changed and why, so the later review has something to compare against.

Merge and redirect

When several articles target the same query, fold the useful parts into the strongest one and send the others to it. Google's redirects documentation (Opens in a new tab) lists consolidating duplicate URLs, merging sites, moving sites and handling removed pages as typical uses. A permanent redirect, using a 301 or 308 status, signals that the target should be treated as the canonical address and suits a change you do not plan to reverse. A temporary redirect keeps the source page in results. The documentation also reserves JavaScript redirects for cases where server-side options are not possible. The destination should answer the same need as the old page; sending readers to an unrelated page is poor experience, and that point is sound practice rather than a quotation from the documentation.

Remove

Deleting a page that has no remaining value is legitimate, but first look at its internal links, its inbound backlinks, and its place in menus and sitemaps. Google's documentation on HTTP status codes (Opens in a new tab) says 404 and 410 are handled like other 4xx errors, that newly encountered 404 pages are not processed, and that URLs already indexed are removed from the index. It also notes that a page returning a 2xx code that looks like an error is reported in Search Console as a soft 404. After removal, confirm that the server returns a genuine error code rather than a friendly page with a 200 status. If a sensible successor exists, redirect instead of deleting.

Leave and monitor

Not every decline calls for action. Google's guidance notes that small position changes often need none and warns against radical edits to pages that perform well. Write "leave" in the record along with the date you will look again, so that it is a deliberate deferral and not an oversight.

Fictional example: a four-article watchlist

Fictional example: a watchlist contains four articles. One is a seasonal guide whose long-view line dips every year in the same months, so the decision is to leave it. Two articles compete for the same primary query and the newer one is ahead, so the decision is to merge and permanently redirect the older address. The fourth describes a retired tool, and a current article already covers the topic, so the decision is to redirect. No article needed a complete rewrite. The scenario carries no figures and only illustrates how often the right move is putting existing content in the right place.

Checkpoints at 30, 60 and 90 Days

Once you act, an experiment starts. Google notes that effects can appear over days to months, and advises waiting a few weeks before analysing again. A 30, 60 and 90 day rhythm is a practitioner convention that fits that advice and is not a schedule Google publishes. Stretch it if your site is crawled slowly.

Day 30: confirm the plumbing

Verify that the page is indexed, that the canonical address is the one you intended, and that every redirect resolves correctly. In Search Console, compare the period since the change to the period before it. Do not expect a traffic gain yet. The goal at this stage is to confirm that the technical side was implemented as planned.

Day 60: read the direction

Look at the position of the primary query, the impressions trend and the click difference. If impressions rise but clicks do not, the title and description may need their own attention. If position has not moved at all, revisit the original hypothesis.

Day 90: record the result

Write the outcome into the record in plain terms: recovered, partly recovered, unchanged or worse. A worse result is information too, because it tells you which hypothesis to drop or whether to reverse the change. If the article is still sliding, ask whether the topic itself has lost value.

Make it a loop

Reuse the watchlist at each cycle with the same columns, so that the accumulated decisions show which kinds of articles age fastest and why. That history becomes a planning input as much as a maintenance log.

Mistakes That Waste a Review Cycle

• Calling every drop decay: content edits made before technical, security, migration and update causes are excluded can be effort in the wrong place.
• Skipping the long view: without the 16-month range, a yearly dip is mistaken for a problem.
• Reading one metric alone: clicks and impressions moving together or apart narrows the cause.
• Changing dates without changing content: Google lists this among practices to avoid.
• Missing self-competition: the real culprit may be a newer article on your own site.
• Redirecting to an unrelated page: a destination that does not meet the reader's need damages the experience.
• Overhauling pages that perform well: the traffic drop guidance advises against radical changes in that case.
• Keeping no record: three months later nobody remembers what was changed or why, and the result teaches nothing.

Frequently Asked Questions

Is content decay an official Google term?

No. It is an industry term and does not appear in Google's documentation as a concept. Google organises traffic drops by cause: technical issues, security problems, manual actions, algorithmic changes, seasonality and site moves. Use the phrase as a working label for slow, article-level decline.

How do I find articles losing traffic in Search Console?

Extend the date range to check seasonality, use the Compare tab to set the last three months against the previous period or the same period last year, then sort the Pages table by Clicks Difference. For each candidate, look at its queries and see whether clicks and impressions fell together.

What does GA4 add to decay detection?

The landing page report shows sessions and average engagement time per session for each article, and adding session source and medium as a secondary dimension shows where visits come from. It helps you check that a Search Console decline also appears on your own site.

Will changing an old article's date help it recover?

No. Google says that changing dates to look fresh when the content has not substantially changed does not help. The date should record a real change, and the visible date should match the structured data.

Should I delete a declining article or redirect it?

If a successor page serves the same purpose, a permanent redirect is the usual choice. If there is no demand and no successor, returning a 404 or 410 removes it. Check internal links and backlinks first.

When should I judge the result of a change?

Google says effects can take days to months to appear. Day 30, 60 and 90 checks are a practitioner rhythm. The first confirms the technical setup, and meaningful interpretation needs a few weeks of patience.

Conclusion

Content decay is a useful name for something specific: one article slowly losing visibility with no event to blame. Finding it takes a method. Extend the date range to rule out seasonality, compare periods in Search Console and sort by click difference, confirm the picture in GA4, then clear technical, security, update, intent, competition and self-overlap explanations in that order. Record a decision for each candidate, whether to refresh, merge, redirect, remove or leave alone, and check back at 30, 60 and 90 days. No step guarantees a particular ranking outcome.

If your decision for some articles is a rewrite and you want writing support, you can review the scope and terms of the SEO content writing service on the Akua Adforce product page.