Unsourced Claims: Why Citations Matter and How to Fix It

Somewhere on your page you wrote a sentence that sounds authoritative. Maybe "73 percent of buyers abandon slow sites." The check caught it because there is no citation attached, and a reader has no way to confirm whether that figure is real or whether someone made it up on a slide deck years ago and it has been passed around as fact ever since. That last part is the genuinely annoying bit, because a lot of these zombie statistics have no traceable origin at all.

What this check flags

The check looks for factual and statistical assertions that carry no citation, link, or named source. Percentages, dollar figures, rankings, "studies show," "experts agree," and comparative claims ("the fastest," "the most popular") all trip it. It is not flagging your writing for having a point of view. It is flagging claims that a careful reader, or an AI engine, would want to verify and cannot.

Why unsourced claims hurt

Three reasons, and they stack. First, trust. Google's Search Quality Rater Guidelines state plainly that trust is the most important member of the E-E-A-T family, because an untrustworthy page has low E-E-A-T no matter how expert or authoritative it otherwise looks. Unsupported claims read as untrustworthy. Second, accuracy. When you cite a primary source you give yourself a paper trail, and you give the reader a way to confirm you are right. Third, and this is newer, AI search engines reward sourcing directly. Generative engines behave like risk minimizers: they prefer to cite content that makes specific, verifiable claims backed by credible sources, because that lowers their odds of repeating something false. Industry research on generative engine optimization found that adding well-sourced statistics measurably improved a page's visibility in AI responses. Perplexity built its whole product around cited answers, where every factual claim carries a source. If your claim has no source, it is a weaker candidate for that citation.

Opinion versus factual claims

This trips people up, so let us draw the line cleanly. An opinion is yours to state. "We think image-heavy pages convert better for ecommerce" is a judgment, and you are allowed to hold it. A factual or statistical claim presents something as objectively true about the world: "image-heavy pages convert 27 percent better." That second sentence is making a measurable claim, and a measurable claim needs a source. The test is simple. Could a reasonable person check this against reality? If yes, it needs a citation. If it is clearly your interpretation, frame it as such and move on.

How to cite well

Link the primary source, not a blog that links to a blog that links to the study. If a university, a government agency, or a peer-reviewed paper produced the number, point straight at them. Name the study or organization in the sentence itself: "according to the 2024 Baymard Institute checkout study" beats a bare hyperlink, because the named authority does work even before the reader clicks. Add a date so the claim ages honestly. And keep the number faithful to the source. Rounding 68 to "roughly 70 percent" is fine. Turning "in one survey" into "most studies confirm" is not.

The special danger of fabricated stats

Inventing a statistic is the worst version of this problem, and it is worth being blunt about it. A made-up number that looks precise is more dangerous than a vague claim, because the precision invites trust you have not earned. If someone fact-checks it and finds nothing, your whole page loses credibility, not just that line. AI engines can cross-reference, readers can search, and the original-source trail either exists or it does not. This site does not invent statistics, full stop. If a number cannot be sourced to something real, it does not get published here, and it should not get published on your page either.

How to fix it

Go claim by claim. For each flagged statement, do one of three things. Source it: find the primary origin, link it, name it, date it. Soften it: if it is genuinely your view, rewrite it as an opinion so it stops posing as a verified fact. Or cut it: if you cannot find a source and it is not really your opinion, the number was probably never solid, so remove it. Do not paper over the gap with "studies show" and no study. That phrasing is exactly what the check is trained to catch.

Common mistakes

Citing a secondary source that itself never names the origin. Linking a homepage instead of the specific page with the data. Quoting a statistic with no date. Writing "experts agree" as a substitute for an actual reference. And the big one, reusing a stat you have seen everywhere without ever checking whether it traces back to anything. Popular does not mean sourced.

FAQ