Documentation
Content Decay Map: spot pages losing traffic over time
A heatmap of your top pages by month or week that rates each one from severe decay to growing, so you know which articles to refresh first.
The Content Decay Map (titled Content Decay Heatmap in the app) lays your top pages out as a grid: one row per page, one column per month or week, each cell shaded by how that period compares with the page's own best. Slow declines that never show up as a sudden drop become obvious, and every row gets a status and a decay score so you can sort the refresh queue rather than eyeball it.
Load the heatmap
The map defaults to the last six months with no comparison; pick another range in the date picker if you want a longer view. Search Console keeps about 16 months; properties with stored history can go further. Then set:
- Page URL contains: narrow the map to a section, such as
/blog/. - Rows: the top 50, 100 or 200 pages by clicks in the window.
- Month or Week columns, and Path or Full URL labels.
- The shared filters (country, device, search type and others).
Press Load Heatmap Data. AutoPSEO fetches daily page rows (capped at 50,000; the status line tells you when the cap was hit) and builds the grid. A badge in the header shows where the rows came from: live Search Console, stored history or demo data. Exclude a page from the results and reload to remove noise such as a retired campaign page.
How each page is classified
Classification always uses clicks, converted to clicks per day in each period so that a partial first or last month does not look like a collapse.
- Peak is the page's best period before the current one.
- vs Peak compares the current period with that peak.
- Trend compares the current period with the average of the three periods before it.
- Decay score runs from 0 to 100: 70% weight on the drop from peak and 30% on a falling trend.
| Status | Rule |
|---|---|
| Severe | Current period at least 50% below the peak |
| Moderate | At least 25% below the peak |
| Growing | Within 10% of the peak and more than 10% above the recent average |
| Stable | Everything else |
| New | No traffic before the last two periods, so there is nothing to compare yet |
The cards count pages per status and show Clicks lost per period: for severe and moderate pages, the gap between their peak daily rate and today's, expressed per 30 days (or per 7 for weekly columns).
Read the grid
The metric toggle (Clicks, Impressions, Avg. Position, CTR) changes what the cells show and how they are shaded, not the status. Shading is relative to each row's best period, so a small page and a large page are comparable. The Threshold slider (4% by default) sets how big a change from the previous period must be before a ▲ or ▼ marker appears, and you can pick the shading colour. Partial periods are marked with an asterisk. Each row also shows the page's average position over the last seven days and an On-page shortcut that runs the On-Page SEO Check for that URL.
Decide what to refresh
A page that lost clicks while its position held usually lost them to a changed results page; one whose position slid has been overtaken. Compare the page with what now ranks above it, update facts and examples, cover the questions searchers ask now, and remove what is out of date. Google's guidance on helpful content warns against changing a page's date to make it look fresh when the content has not substantially changed, so make the update real. The playbook Fix content decay goes step by step.
Seasonal pages decay every year on schedule; check last year's shape with a longer range before rewriting them.
How AutoPSEO helps
For articles it manages, the content engine runs a decay check every week: an article whose clicks over the last 28 days fell at least 40% below the 28 days before (from at least 20 clicks) gets a refresh task, up to three per site a week, and articles younger than 45 days or changed recently are left to settle. See Keeping articles fresh. Your AI assistant can ask for decaying pages through the content_decay tool of the MCP server; it compares two periods directly instead of drawing the heatmap.
Related: Query Decay Map, Content Groups.