Playbook

Content decay: how to find dying posts and bring them back

Content decay is the slow traffic loss on pages that once ranked. Find it in Search Console, diagnose the cause, and refresh only what's worth saving.

Key takeaways
  • Split every page's click loss into impressions and CTR before you edit anything, because a CTR drop calls for a title and snippet fix, not a rewrite.
  • Sort decaying pages by absolute clicks lost, require the drop to hold for six to eight weeks, and start with pages ranking around positions 5 to 15.
  • Match each page's format to what ranks today, and merge or redirect cannibalizing URLs instead of refreshing both.

Most pages that rank well don’t stay there. They fade. A post earns a top position for a useful query, holds it for a year, then slides to the third page of results, and nobody notices until a quarterly review shows organic traffic down by a third. That slow slide is content decay, and it is the most common traffic loss on established sites. It is also one of the most fixable, because the data that reveals it already sits in Search Console.

The catch is that the fix is rarely the one people reach for first. A team sees a dropping page and starts rewriting it. Sometimes that works. Often the page was fine, and the query, the results around it, or a page on your own site had moved. Diagnose before you edit.

What content decay looks like in your data

Content decay is a sustained decline in organic clicks to a page that used to earn them, with no single event behind it. It’s a slope, not a cliff. A cliff usually points somewhere specific: a canonical tag pointing at the wrong URL, an accidental noindex, a manual action, or a core update that hit a whole section of the site. Decay is what remains once those are ruled out. A few weeks of fewer impressions, a click-through rate that drifts down, an average position that creeps from 3 to 6 to 9.

Two numbers matter most, and they have different causes. Clicks are impressions multiplied by click-through rate (CTR), so every drop comes from fewer impressions, a lower CTR, or both. Treat them separately.

Here’s a worked example with illustrative numbers. A guide gets 4,000 impressions a month at a 5% CTR, which is 200 clicks. Six months later it gets 3,200 impressions at a 3.1% CTR, which is about 99 clicks. The loss is about 101 clicks, and it splits two ways. The 800 missing impressions account for roughly 40 clicks at the old CTR. The lower CTR, applied to the 3,200 impressions still showing, accounts for roughly 61. Most of this loss is a click problem, not a ranking problem. The page may be exactly where it was and has simply stopped being the result people choose.

That changes the work. A ranking problem sends you toward the content and the competition. A CTR problem sends you toward the title, the snippet, and whatever now sits above you in the results.

One trap before you go further. Average position in Search Console is a weighted average across every query a page appears for. A page that ranks second for its main query and thirtieth for twenty long-tail variants can show an average of nine, and a page that hasn’t moved at all can look like it’s sliding because the mix of queries shifted underneath it. Read position at the query level before you conclude anything.

The refresh loop 01 Find Sort by lost clicks held for 6+ weeks 02 Diagnose Impressions vs CTR, then name the cause 03 Prioritize Value and how close it is to ranking 04 Refresh Smallest fix that matches the cause Then measure against pages you did not touch.
Four stages, in order. The diagnose step, outlined in the accent color, is the one most refresh programs skip.

Finding decaying pages in Search Console

Search Console shows what happened in search. Google Analytics shows what happened after the click, which matters for deciding what’s worth saving, but it can’t tell you why the clicks stopped. Search Console finds the decline; Analytics ranks it by value.

This is the sequence that works:

  1. Export a baseline before you need one. Search Console keeps performance data for 16 months, and you can’t backfill it. Export page-level and query-level data every month into storage you control.

Keep your own history: without a monthly export, you can't see a page's trend before the window closes, and you can't tell a slow slide from seasonality. A scheduled pull through the Search Console API is the simplest way to keep it running.

  1. Compare long windows, not days. Set the Performance report to the last three months and compare it with the same three months a year earlier. Daily and weekly swings are mostly noise, and seasonal topics dip every year. Year-over-year is the only comparison that cleanly separates a seasonal dip from real decay, and the 16-month retention window is what makes it possible.

  2. Sort by absolute clicks lost, not by percentage. A page that fell 80% from 12 clicks is a rounding error. A page that fell 30% from 900 clicks is the one to work on. Percentages look dramatic on small pages and hide the big losses on large ones.

  3. Require the decline to persist. Flag a page only when the drop holds for six to eight weeks. One bad week after a spike is noise, and a page that recovers on its own was never decaying.

  4. Open the Queries tab, filtered to the page. This tells you whether you lost one query or a whole cluster. Losing the single query that carried most of the traffic is a different problem from losing twenty small queries that add up to the same total. Search Console also hides some queries for privacy, so the query rows won’t sum to the page total. That gap is expected. Don’t chase it.

  5. Remove the expected losses. Retired products, deliberate redirects, seasonal topics, and pages you’ve already decided to merge are not decay. Mark them and move on, so they stop taking up room in the queue.

Diagnosing why a page is losing clicks

Refresh programs tend to fail here. They assume the cause, then start rewriting. Most decay falls into one of four causes, and a fifth looks like decay but isn’t. Identify which one you’re dealing with before you touch the page.

Intent drift

Google decides what a query means partly by looking at what already ranks for it. When the top results change format, from long guides to tools, from articles to category pages, or from listicles to video, the query’s intent has moved. A well-written page in the old format gets pushed down, and adding paragraphs won’t bring it back.

The test takes a few minutes. Search the query in a private window and write down the format of each of the top ten results. If your page is a guide and seven of the ten are calculators or templates, you have your diagnosis. The fix is format, not prose. A page titled “What is a content audit” that explains the concept can’t compete with ten results that hand the reader an audit template. Add the template, the worked example, or the calculator, or accept that the page should be a different kind of page altogether.

SERP change: the click moved, the rank didn’t

Here impressions and position hold, but CTR falls. The usual causes are a feature that now sits above your listing (a featured snippet, a video carousel, an AI-generated answer), a competitor who added rich results you don’t have, or a title that no longer matches how people phrase the query.

Check the live results for the query and note what’s above you. If an answer box or AI overview now resolves the question outright, be honest about the ceiling. You may not win that click back. The realistic goal is to stay a cited source and the best next click for the follow-up question. That means a direct answer near the top, clear structure a snippet can lift, and a title that promises the specific payoff.

Staleness

Facts, versions, prices, screenshots, and links age on their own schedule. A guide that describes a tool’s old interface, cites a 2023 benchmark as current, or links to a page that now returns a 404 tells the reader the site isn’t maintained. It matters most when a competitor updated and you didn’t.

The tells are visible in the page itself: a year in the title that’s now two years back, screenshots of interfaces that no longer exist, references to features that were removed, and dead external links. Staleness usually produces a slow slide across many queries rather than a sharp drop on one, which is why it hides in monthly reports.

Cannibalization

Two of your URLs compete for the same query. Google picks one, then the other, and neither gets a stable position. Clicks split, and both pages look mediocre.

To check, open the Queries report, filter to a query where you’re losing ground, and see whether two URLs appear. Compare their impressions and positions. A healthy page ranks steadily. Cannibalized pages flip from week to week. The usual source is an older post that covers the same ground as a newer one, often written by someone else, with a slightly different angle.

Demand decline

This one looks like decay and isn’t. The page ranks exactly where it did, but impressions fell because fewer people search for the query. Seasonal topics, shrinking product categories, and technologies that have peaked all do this. If the query’s total volume dropped, no amount of editing will restore the traffic.

Check the query’s trend in Google Trends, and compare its impressions across your whole site. If the decline is market-wide, the right move is usually to leave the page alone or retire it.

Keep one table open while you work through a list. It maps what you see in the data to the cause to check first:

What you see in Search Console Most likely cause First check
Impressions and position fall together Competitors passed you, or the page went stale Compare today’s top ten with the results that ranked when you were strong
Position holds, CTR falls SERP change or a weak title and snippet Look at what now sits above you, and whether your title matches the query wording
Top results switched format Intent drift List the format of each top-ten result in a private window
Impressions fall, position holds Demand decline Check the query’s trend in Google Trends and across your site
Two URLs alternate on one query Cannibalization Compare both URLs’ impressions and positions on that query

These causes overlap. A page can be stale and cannibalized at the same time. Diagnose the one that explains the largest share of the lost clicks, fix it, and re-check the data before moving on to the next.

Prioritizing the refresh queue

Most sites have more decaying pages than they can refresh well. Triage is what keeps effort pointed at pages that will return something.

Score each candidate on three things:

  • Value. What is the page worth when it’s working? Use the organic landing-page report in GA4 to see whether that traffic converts into signups, purchases, or assisted conversions. A page losing 200 clicks to a glossary-style query may matter less than one losing 40 clicks to a query with buying intent.
  • Distance. How far has the page fallen from where it used to be? A page that slipped from position 4 to position 8 is close to recovery. A page that dropped from position 3 to position 25, behind a field of established sites, needs a much bigger push.
  • Cost. A title and snippet change takes an hour. An intent-drift rebuild takes days. Cheap fixes on valuable pages go first.

A workable order falls out of those three.

  1. Close, valuable, cheap. Pages sitting roughly between positions 5 and 15 on queries with real impressions, where the cause is CTR or staleness. Do these first. The gap is small and the fix is quick.
  2. Close, valuable, expensive. Intent drift or cannibalization on pages in the same range. Schedule these as real projects with an owner and a date.
  3. Far and competitive. Pages stuck at position 20 or lower where the top results are much stronger. Competing head-on rarely pays. Ask whether the page should exist as a standalone, or whether its useful parts belong in a stronger hub page.
  4. Demand-driven. Leave alone, consolidate, or retire.

Don’t rank the queue by total traffic alone. A large page that has lost its buyers is a worse use of a week than a smaller page with a clear, fixable gap.

What to actually change on the page

Start with one principle: change the smallest thing that addresses the diagnosed cause. A full rewrite feels productive, but it throws away whatever was already working, and afterward you can’t tell which change did what. A page that ranked for years on a set of queries carries much of its value in its existing structure, its URL, and its inbound links. Keep those unless you have a specific reason to change them.

Then match the change to the cause.

  • For CTR loss, rewrite the title tag so it uses the query’s own wording and promises a specific payoff. Rewrite the meta description to say what the reader will get. Move a direct, one-paragraph answer to the top of the body, so there’s something a snippet can lift. Change one thing at a time, and wait before changing the next, so you know what worked.
  • For staleness, update every dated claim, replace screenshots of old interfaces, fix or remove dead links, and delete references to features that no longer exist. Add the new information that matters, such as a current tool, a newer benchmark, or a changed rule, and attribute it to its source. This is usually a targeted edit that takes an afternoon, not a rewrite.
  • For intent drift, change the format to match what now ranks. Add what the top results provide: a template, a calculator, a worked example, a side-by-side comparison, or a checklist. Compare your H2s with the three strongest competitors and add what’s missing only where you can add something you can verify. If the dominant format is a different kind of page entirely, build that page, then differentiate or redirect the old one.
  • For cannibalization, pick the stronger URL. Fold the weaker page’s unique material into it, redirect the weaker URL with a 301 to the survivor, and point internal links at the survivor. If the two pages really do serve different intents, retarget the weaker one to a different query and make that difference explicit in its title and first paragraph.

Don't fake freshness: changing the visible date without changing the substance gives readers the same answer under a new label. It also leaves you unable to tell later whether the refresh did anything. If the facts didn't change, leave the date alone.

Don’t overlook internal links, the cheapest signal you control. Link to the refreshed page from newer, related articles, using anchors that name the topic. If you’re weighing a refresh against a new post on the same subject, how often to publish to rank covers the cadence side of that trade-off. My position is that a refresh of a page that already ranks usually beats a new page that competes with it. Once the page is in better shape, its new material can also feed other formats. Turning one article into ten assets shows how to do that without publishing duplicates.

Measuring a refresh without fooling yourself

Log every change before you publish it: the URL, the date, the cause you diagnosed, and exactly what you changed. Without the log, each later decision becomes a guess, and you’ll end up crediting random swings to edits you happened to make that month.

Then give the change time. Google has to recrawl the page and reassess where it belongs, and a rebuild takes longer to register than a title change. Judge a refresh on six to twelve weeks of data, not the first two. Early movement in either direction is usually noise.

Compare like with like. Use the same query set, the same period length, and a year-over-year comparison for seasonal topics. Watch the query level as well as the page level. A page’s total can rise while the queries you targeted fall, if it picked up new long-tail traffic that happens to offset the loss.

Keep a control group. Note a handful of similar decaying pages you didn’t touch during the same period. If they recovered at the same rate, your refresh may not be what moved the number. A core update or a seasonal swing could be doing the work, and you want to know that before you repeat the same edits on the next twenty pages.

When to stop refreshing

Not every page affected by content decay is worth saving, and sometimes the honest answer is to let it go. Retire, merge, or redirect when the topic’s demand has gone, when a real rebuild still leaves the page behind a stronger competitor for a query that matters, when it duplicates a better page you already have, or when its traffic never converted anywhere. Retiring is a legitimate outcome. Redirect the old URL to the closest relevant page, remove the internal links that point to it, and log the decision the same way you log refreshes.

The process itself is unglamorous. Export, diagnose, prioritize, change the smallest thing that addresses the cause, and measure against a control. Most content decay programs fail at the second step, because diagnosis feels slower than editing. Skip it and you’ll rewrite pages that never had a problem while missing the ones that did.

The discipline here is research before writing, a human voice, and judging each page on its own merits. That’s the same discipline AutoSEO Stack builds in from the start. It researches what buyers search for, writes genuinely useful articles, and publishes them on a schedule, with the quality bar set per page. If you want to see that running on your own site, start with the features.