Statistics Round-Ups Earn Links and Launder Errors

Statistics round-ups do earn links. A writer needs a figure to support a point, they need it now, and the page that lists dozens of figures on their topic is the fastest place to get one — so it gets cited. They are also the main mechanism by which unverifiable numbers become “common knowledge”, and most of them are built by copying each other. Both of those things are true at once, which is why this format deserves more suspicion than it gets.

Why they attract links

The appeal is pure convenience at the moment of need. Somebody is mid-paragraph, needs a number, and does not want to read a research report to find one. Your page is a shortcut, and shortcuts get bookmarked and shared.

That’s a real property and it does produce citations. But look at what it’s built on: you are a summary, not a source. Everything of value on the page was produced by somebody else. Which means the format has two structural problems and neither is fixable by writing better.

A careful writer skips you. Anyone with editorial standards traces the figure to whoever produced it and cites that instead. Your page was a search tool, not a reference. So the links you collect are disproportionately from people who didn’t check — which is not the audience you want vouching for you.

A careless writer cites you for a number you never verified. Now your domain is attached to that figure. If it turns out to be wrong, or to have described something quite different from what your one-line summary implied, that’s your page carrying the error.

The contrast with actually being a primary source is stark. A source’s citations get stronger over time as more people build on the work. A round-up’s citations are only as good as the weakest number on the page.

How laundering happens, step by step

The pattern is boringly consistent. Follow almost any widely repeated industry figure backwards and you find some version of this chain:

Round-up A states the figure with a confident one-line gloss and names round-up B as its source, or names nothing at all.
Round-up B — which was assembled the same way — attributes it to a vendor’s blog post.
The vendor’s blog post attributes it to a survey. Sometimes its own, sometimes a partner’s.
The survey either no longer exists, was never published in full, or did exist and measured something narrower than the sentence now travelling under its name.

At every hop, a qualifier falls off. The population shrinks or expands. “Among respondents who had already adopted the thing” becomes “of companies”. A figure about one vendor’s customer base becomes a figure about an industry. Nobody lies at any single step; the meaning just erodes.

Then repetition does the rest. Once a number appears in enough places, everyone assumes somebody must have checked, and asking becomes faintly rude. This is why you’ll see a figure attributed to a year that predates the technology it describes, still circulating, still being cited.

The honest version, if you build one

You can publish a round-up that isn’t part of that chain. It costs a great deal more than the usual version, which is exactly why most people don’t.

Name the original publisher and the year, in the visible text, for every entry. Not in a source list at the bottom — beside the figure, where anyone quoting it will carry the attribution along. Under this site’s no-external-links policy, sources are named in plain text rather than linked, and naming is what matters for honesty: a reader who has the publisher and the year can find and check the work themselves. A bare hyperlink with no name in the text is worse, because it decays into nothing.
If you couldn’t trace it to a primary source, it doesn’t go in. This is the rule that does all the work, and it will cut your list badly. Accept the shorter list. A page with a modest number of traceable figures is more useful, and much safer to be cited for, than a long one padded with hearsay.
State the population each figure describes. A survey of one vendor’s customers is not “the industry”. A poll of conference attendees is not “practitioners”. Say who was actually asked, and how they were reached, because that’s usually where a plausible number becomes a misleading one.
Record what you removed and why. A short note that certain widely repeated figures were excluded because they couldn’t be traced turns your page into something better than a list: a page that has done the checking. That’s a claim nobody else on the topic is making.
Re-verify on a schedule, and delete rather than let it rot. Figures go stale, publishers withdraw work, and a page that hasn’t been touched in years is quietly wrong. Removing a dead entry is a service; keeping it because the page looks thinner without it is how the laundering restarts. Cited pages need ongoing maintenance more than uncited ones, not less.

The version to refuse

Some round-ups shouldn’t be built at all:

Pages assembled from other round-ups. If your research consisted of opening the pages currently ranking and merging their lists, you have produced a copy of a copy and inherited every error in both.

“Updated for 2026” pages where only the title changed. Readers treat the year as a claim that the figures are current. If nothing was re-checked, the year is the falsest thing on the page.

Any figure you cannot state the origin of. If your honest answer to “where does this come from?” is “it’s everywhere”, leave it out.

It’s also worth naming the scale problem. Mass-produced, near-identical round-ups across many sites are the textbook scaled-content pattern that search engines discount — Google’s spam policies describe content generated primarily to manipulate rankings rather than help people in exactly those terms. A format whose main appeal is that it’s cheap to churn out is a format search engines have every incentive to stop rewarding.

What to do next

If you want durable citations on a numeric topic, own a number instead of collecting other people’s. Measure something you’re already in a position to measure, publish it, and let everybody else’s round-up cite you — that’s the position original data you already have puts you in.

And whichever you do, publish the method. For a round-up, the method is your tracing and exclusion rules. It’s the only thing that distinguishes your page from the ones it’s competing with.