To track content decay you need the peak, and the peak is usually outside the window. Pages rarely die in a week. They leak clicks across two or three years, a few percent a month, which is slow enough that a page sitting at 60% of what it once earned reads as flat and healthy inside Search Console's 16 months.
That's the whole difficulty. Decay is defined by comparison to a better version of the page, and Search Console deletes the better version on a rolling schedule.
Get early access Free during early access. No card required.
What a shrinking window does to the diagnosis
A page that peaked 20 months ago has already lost its peak from your reporting. What's left is the tail of the decline, and a tail with no head looks like a plateau. You end up comparing a bad month to a slightly less bad month and concluding the page is stable.
The same limit hides which pages to prioritize. Ranking your pages by clicks tells you what is big today. Ranking them by the distance from their own historical maximum tells you what is bleeding, and the second list is the one worth acting on. You can only build it if the maximum is still stored somewhere.
How Month17 tracks it
Month17 archives every performance row daily and keeps it with no retention limit, so the comparison window grows instead of sliding. Its MCP server includes a content_decay tool that buckets each page's clicks into consecutive 30-day periods and returns the pages that declined in every one, counting a missing period as zero rather than skipping the page, so pages that decayed all the way to nothing still surface.
Because the tool takes the number of periods as an argument, the depth of the check is limited by your archive rather than by Google. Three periods is a quarterly health check. Twelve is a different question entirely, and it's one Search Console cannot answer at all.
What it won't tell you
A consistent decline is a pattern, not a diagnosis. Seasonal businesses decay every year on schedule and recover on schedule. Pages about retired products are supposed to fade. Somtimes a rewrite already happened and the decline you're looking at ended last month.
The tool finds candidates and the ranking work stays yours. What changes is that the candidates are real instead of whichver pages happened to look weak in the last 16 months. From there the page-update loop is the usual next step.
One caveat worth stating: an archive is not retroactive. It holds the window Google still has on the day you connect, and starts accumulating from there. The peaks already outside that window are gone for good, which is an argument for conecting sooner rather than for waiting until decay becomes a problem. The data loss calculator puts dates on it.