We scanned 5,034 claims across 961 blog posts on 46 domains and checked how old the data behind each cited statistic had grown. In posts under a year old, 2.0% of cited statistics were out of date. In posts two to three years old, the rate was 10.3%, about five times higher.
Content decay is what happens to the data inside a published post as time passes: the statistics it cites age past the point their sources still support them, while the prose stays exactly as persuasive as the day it shipped. One in five of the posts that cite data were already carrying numbers two or more years old. Nothing on the page marks which ones. The decline you eventually watch in an analytics dashboard is a delayed reading of a process you could count today, one claim at a time.
What Content Decay Means
Search the term and the definition you'll find describes a symptom: a page's organic traffic and rankings sliding over months. That definition lives entirely inside your analytics tool and says nothing about the page itself.
Open the post and the trouble is harder to see. The benchmarks and cited figures inside it have been aging since the day you hit publish. The writing still holds up. The argument still lands. And the statistic in paragraph two now trails its source by three years, with nothing in your stack built to notice.
When Google's Lizzi Sassman and John Mueller were asked about content decay, they treated it as a term they didn't recognize and put gradual traffic loss down to other causes. For traffic, that reading is fair. A slow slide usually traces back to fading search interest, seasonality, or a competitor with a better page, and giving any of that a new name explains nothing. A sudden drop deserves its own diagnosis, and separating an algorithm update from decay is a read you can run from your own Search Console.
There's a second clock they left out. Cited data ages on a schedule set by whoever published the source, so a post can hold every reader and every ranking it ever earned while the figure it rests on falls years behind.
The unit that makes decay measurable is the claim: one verbatim statistic a post commits to. A post never crosses from fresh to stale in a single move. Individual numbers expire one at a time, and each expiry is countable the day it happens.
Content Decay vs Content Debt
The two terms travel together and mean different things. Content debt is a stock: the unsourced claims, orphaned figures, and unverifiable statements already sitting in your published library, measurable as a quantity on any given day.
Content decay is a rate. Leave the library alone for a year and more of its cited data will have drifted out of date, which means the debt grew without anyone editing a file. By the time a traffic drop announces the problem, the balance has been growing for months.
The Shape of Content Decay
Plot the drift against post age and it forms a curve. Across the 5,034 claims in our study, the share of a post's cited statistics that were two or more years old rose with every age band.
The curve climbs through every band we measured, and it starts climbing on publication day. Every post you've shipped is somewhere on it right now.
Counting affected posts instead of claims tells the same story. Under a year old, 10.0% of posts carry data two or more years old. In the 12 to 18 month band, 23.2% do. The bands after that hold at 25.1% and 23.9%. Once a post passes its first birthday, its odds of carrying aged data have more than doubled, and they stay there.
Which numbers expire first is predictable. Borrowed statistics, the "according to" figures lifted from someone else's research, age faster than anything a publisher measured itself, and borrowed is most of what blogs run on: 65.5% of the claims we scanned came from external sources, against 34.5% first-party. The fastest-decaying part of your content is the part you never controlled.
The causes barely involve the writing. A borrowed figure expires the moment its source publishes a newer one, whether or not you ever hear about it. A report moves, goes behind a paywall, or disappears, and the number it supported keeps standing on the page with nothing underneath. A benchmark shifts because the market it described shifted. A post can argue as well at three years old as it did at launch and still lose currency on all of these clocks at once, because none of them belong to the author.
Why Ranking Posts Hide the Most Decay
Refresh queues get built from traffic reports. A post that still ranks looks healthy, so it sinks to the bottom of the queue, and the posts that keep ranking longest tend to be the oldest and most cited on the site. Those are the posts that have had the most time to accumulate expired figures.
So your strongest page can sit at position three carrying more aged data than anything else you've published, and every system you watch will report it as fine.
The standard response is a refresh, and most refreshes touch the wrong layer. A new publish date and a rewritten introduction change how fresh the post looks without establishing that anyone rechecked a single figure. I've seen a refresh ship where the date moved and every number beneath it stayed put. Ahrefs' research on date-refresh tactics found that a date change with no meaningful content change can make the decline worse, with date-only refreshes moving ranking positions by as much as 95 places. That gap between the freshness signal and the state of the data is freshness theater, and it's why a freshly stamped post can keep decaying in the one dimension readers actually check.
How to Spot Content Decay Early
Very little of the aged data in our scan was provably wrong. A figure from a 2023 report may still be the best number available; it needs a recheck. About 3% of what we found had crossed the line into stale as presented, a years-old benchmark displayed as if current, and the rest had simply gone unexamined since publication. Unexamined never registers on a dashboard. It registers immediately in a count of claims.
The leading indicators sit in plain text: a cited statistic with no date attached, an "according to" figure whose source report is several editions old, a link that now resolves to a 404 or a paywall.
Teams discover aged data in very different ways, and the way yours does sets how much the drift ends up costing. How do you find out today?
Four of those five answers depend on something downstream breaking first: traffic, a reader's trust, or an audit calendar coming around. The fifth catches the drift before anything downstream can break.
Traffic, the calendar, an inbox: each of those is a proxy standing in for the figure itself, a downstream readout that only moves long after the number it depends on has gone old. That is how a post can sit at the top of its results page holding the most out-of-date data it has ever carried. The check is landing one layer above the claim, where the staleness actually lives, instead of on it.
Reading a post at the claim level is what the Content Health Scanner does: paste a URL and it extracts every statistical claim on the page and scores each one for age and staleness risk. A claim keeps the verbatim wording the page published and carries a state, current, stale, fixed, or expired, and when the scanner can't call one confidently it returns needs-review instead of guessing.
Without an account it runs one URL per day; Free allows three, Pro 10, and Visionary 50. A scan reads the page as it stands today, and when the moving part is the source itself, stale data detection on a monitored page watches that external URL and flags the claims citing it as soon as it changes.
The first URL I scanned was the post I was proudest of. Run yours, the evergreen one that still ranks, and read what comes back.
What Content Decay Costs
The cost shows up when you trace the citations. Across the 961 posts we scanned, 70% of borrowed statistics were unverifiable, with no external link at all. Of the links that did exist, about 20% were already dead, gated, or broken. Only 17.2% reached the original primary source. That's the citation layer your competitors publish on, and yours too, unless something is watching it.
Uncounted decay gets discovered by whoever next hits the page with a reason to check: a reader, a competitor, or Google. By then it has already cost a ranking or a citation. The durable fix is content that updates itself as the data behind it moves, which is what living content exists to do, and it starts with seeing the drift while no one outside your team has seen it.
Somewhere in your best post, a number expired while the rankings held. You'll find it first, or your reader will.