Content decay: spotting the posts quietly losing traffic

Three months ago a client of ours lost 38% of organic traffic on a post that had ranked first for two years. Nobody touched the page. Google didnt penalize anything. What happened was content decay: that silent decline that hits good content while you are busy shipping new stuff. It isnt an algorithm update, it isnt a penalty, it is entropy. And almost every blog older than 18 months currently has between 20% and 40% of its catalog stuck in that state, generating half the traffic it could. The catch is that nobody audits until total traffic drops and the CFO asks why.

The first thing I do is open Search Console and export 16 months of impressions and clicks per URL, broken down by month. In Looker Studio I compare the current quarter against the same quarter the previous year. Any URL with a click drop above 25% goes onto the suspect list. Then I cross-reference with average position: if position dropped more than 3 spots, it is a ranking issue. If position held but CTR collapsed, it is usually a SERP feature stealing the click. That split changes the whole rewrite strategy. For the second case, it is worth revisiting Title tags that convert: 7 patterns tested on real SERPs and Does meta description still matter? What CTR data shows before you touch the body.

After isolating the decaying URLs, I sort them into four buckets. First: decay from factual staleness, classic for posts with a year in the title or tools that changed. Second: decay from migrated intent, when the SERP shifted from informational to commercial or the other way around. Third: competitive decay, when somebody published something objectively better. Fourth: technical decay, usually a botched migration, a wrong canonical or a crawl issue. Each bucket needs a different treatment. Lumping everything into a "general update" is like prescribing antibiotics for a virus. For bucket three, Rewrite or rebuild: making the call with SERP data helps you decide between surgery and demolition.

For the technical diagnosis I run Screaming Frog on the URL, check headers, compare rendered HTML against served HTML and look at Googlebot logs from the last 60 days. Absurd things come up here: a post with an accidental noindex since a migration, a canonical pointing at the homepage, JavaScript hiding half the content on mobile. Once Googlebot stops visiting the page, decay accelerates. Worth combining this with Log file analysis: what Googlebot is actually doing and Canonical tags: common mistakes bleeding your organic traffic. In 4 out of 10 audits I run, the issue is technical, not editorial. Rewriting without checking this first is a waste of writer hours.

When the issue really is editorial, I use a simple prioritization matrix: clicks lost in the last 90 days multiplied by the average per-session value of the URL (pulled from GA4). That gives me an impact ranking in dollars. I take the top 20 and open one doc per URL with three sections: what the current SERP expects (top 10 analysis), what the post delivers today and the gap. Here I also cross-reference Search intent: 4 types and how to map them on the SERP because half the decays I see are informational posts trying to rank for queries that turned transactional. Rewriting the opening paragraph wont fix that, you need a fresh angle.

When executing the refresh we keep the URL, only update the published date if the rewrite exceeds 40% of the body, and republish with a visible changelog at the footer to signal freshness to readers and to Google. We update title, meta, first paragraph and add new blocks that answer current PAAs. Internally we redirect links from related posts to reinforce internal authority, something I broke down in Smart interlinking: the internal authority map. We monitor for 6 weeks with a Search Console drop alert. On average, 70% of decaying posts recover at least 60% of lost traffic within 90 days, at a cost of 4 to 8 hours per post.

The practical takeaway: schedule a decay audit every quarter, locked into the calendar, before planning any new posts. Allocate 30% of editorial capacity to refresh, not 5%. Old content ranking well is worth more than new content fighting for position. Start tomorrow by exporting 16 months of GSC data, filtering for drops above 25%, and attacking the top three of the impact matrix. You will recover more traffic in two weeks than in three months of publishing new stuff.