The data in your blog posts ages whether you refresh them or not.
We scanned 5,034 data claims across 961 SaaS blog posts from 46 domains and asked a simple question: how old is the data these posts are actually citing? One in five of the posts that make data claims are carrying numbers two or more years out of date. The share climbs the longer a post sits unmaintained, and the data that ages first is the borrowed statistic you cited from someone else.
Here is the part that surprised me: most of it is old, not wrong. The numbers have aged, not been debunked. But old is how decay starts, and a measurable slice has already finished the trip: 163 claims, about 3% of the total, are stale as presented, a years-old benchmark restated as if it were still current.
Old Data Accumulates the Longer a Post Sits
We sorted every dated claim by how old the cited data is today. A claim that points at a 2026 figure is fresh. A 2025 figure is aging. A figure from 2024 or earlier is two or more years out of date, the point where a reader should ask whether it still holds.
One in five of the posts that cite data carry at least one of those years-old figures. And the older the post, the more of its data has aged out:
In posts under a year old, 2.0% of cited stats are already two or more years out of date. By the time a post is two to three years old, it is 10.3%, about five times the rate. The share of cited data that has aged out climbs cleanly with the post's age. Older evergreen posts do not just have old data, they have more of it.
The Problem Switches On at the One-Year Mark
At the post level the curve has a sharper shape. Under 12 months, 10.0% of posts carry data two or more years old. Past the one-year mark it more than doubles to around 24%, and then it holds there: 23.2% at 12 to 18 months, 25.1% at 18 to 24 months, 23.9% at 24 to 36 months.
So the count of affected posts switches on at the one-year line and plateaus, while the amount of old data inside each affected post keeps climbing. A post crosses its first birthday and the odds it is carrying years-old data more than double. Nobody schedules that review, because nothing on the page changed. The data just got older while the post kept ranking.
Any post past 12 months deserves a look. The hidden cost of outdated charts compounds with every month past that line.
Borrowed "According To" Data Ages First
When we looked at which claims had aged out, the pattern was consistent: the borrowed statistic, the "according to a 2023 report" citation, is the one most likely to be carrying years-old data. A publisher's own measured numbers age the least, because a team that measured something once tends to remeasure it.
That tracks with how the claims got there. A first-party number is something you control and can refresh on your own schedule. A borrowed number is frozen at whatever year you pulled it from, and it keeps getting older while you move on. When you open an old post to update it, start with the borrowed citations. They are both the most likely to have aged and the hardest to notice, because the sentence around them still reads fine.
Old, Not Wrong: Why This Is a Maintenance Problem
Here is the honest framing, and it is the most important line in this study. When we checked each aged claim against the live source for evidence the number had actually changed, most had no confirmed contradiction. The data is old. In most cases it is not yet demonstrably wrong. But a distinct slice has crossed the line: 163 claims, about 3% of the total, are stale as presented, a benchmark from 2019 to 2023 restated in a 2026 post as if it still described today.
That is exactly why this is mostly a maintenance problem and not a fact-checking scandal. A statistic from 2023, cited in a post that still ranks in 2026, is a figure your readers are treating as current when the world has had three years to move. It has not been debunked. It has been left. The job is not to prove every old number false, it is to flag the years-old ones for a human to refresh before they drift from old into wrong.
This is also why a date-stamp audit misses it. "Last updated" measures whether someone touched the page. It says nothing about whether the 2023 figure in paragraph nine is still the number you would cite today.
The Sources Behind the Numbers Are Also Going Dark
While we had every claim and source in a database, we checked the citations themselves. The picture compounds the aging one. Across the borrowed claims, only 30% carry an external link in the claim's own paragraph; the other 70% name a source with no way to check it, or cite nothing. Of the verified links that do exist, about one in five (20%) is dead, gated, or broken. The full breakdown lives in the source-verification study.
Pair that with the aging curve and the failure mode becomes a system. A post cites a borrowed figure from 2023. The figure ages. The link behind it rots. The claim still reads fine in the body, and nothing in a normal refresh cycle visits either problem. The post keeps ranking, keeps converting, keeps pointing readers at a years-old number behind a link that may already 404.
What the Maintenance Actually Costs
The cost model assumes 0.5 hours per claim to find the data point, check it against a current source, and update it. Borrowed claims with no link take longer, because there is no source to check against and the auditor has to relocate the original.
The Variables That Move the Number
Blog size sets the floor. Average post age sets the multiplier. Audit frequency swings the widest, and it is the variable teams have never formalized.
The typical scenario: a 50-post SaaS blog with an average post age of 18 months, audited annually at $75 per hour. Roughly $2,000 per year in claim-maintenance debt. I have never seen a team put that number in a budget. They track whether a post was touched. The data inside it never gets counted.
The question that decides whether this model fits your team is what your update cycle actually reviews.
The cost model above assumes every claim is found during the audit. That assumption depends entirely on how deep the review goes. If the update cycle stops at structural edits and never reaches the data layer, the budget spent on refreshes produces zero claim corrections. The data inside a post ages on its own clock whether or not the intro gets refreshed, so an update that never reaches the data layer leaves the aging in place.
LiquiChart's content debt estimator runs the same cost model on your numbers.
If the result is worth a conversation, generate a shareable link and send it to whoever owns your content budget.
How We Measured It
We collected 961 posts from 46 SaaS domains and extracted every data claim in each: 5,034 claims total.
For each claim that references a specific year, we computed how old the cited data is today: a current-year figure is fresh, a one-year-old figure is aging, and a figure two or more years old is counted as aged. Claims anchored to a fixed event or a historical subject are exempt, because a stat about the 2020 election or a 2018 funding round is not data you are meant to update. This is the same deterministic freshness tier the Content Health Scanner writes for every claim, so you can reproduce a post's result by scanning it.
We deliberately separate two questions. "Is this data old?" is a calendar fact, and it is what this study reports. "Is this number now wrong?" is a stronger claim that requires positive evidence the figure changed, and we only flag that when the live source actually contradicts the cited value. Reporting the first without overclaiming the second is the whole discipline.
Source verification ran as a separate pass: HTTP requests against every unique external citation, classified by reachability, with publish and modified dates fetched from the reachable ones.
The Debt Is Already on the Books
Content debt stays off the quarterly plan. It stays off headcount. It shows up as ad hoc requests, late-night fixes before a sales call, and the realization that a chunk of the claims in your oldest posts are citing data from years you would not put in a slide deck today.
About three in four SaaS blog posts contain data claims. The average post carries several. Past 24 months, about one in ten of a post's cited stats has aged two or more years out of date, on top of the borrowed citations that were already old when they went in and the links behind them that are starting to rot.
Page-level refreshes leave it untouched, because nothing on the page changed. The data needs claim-level detection, and that requires living content infrastructure built around the unit that ages: the individual claim.
That is what living content changes. Maintenance targets the claim.
Pick your highest-traffic evergreen post and see how much of its data has aged.
Every quarter your content budget ignores claim-level maintenance, another cohort of posts crosses the one-year line and starts carrying numbers your readers will treat as current. The debt is already on the books. The only question is whether you price it before or after it prices you.