The statistics page is the only link asset I can predict
We have published a lot of content that earned nothing. Comparison pages, opinion pieces, a few things I was personally proud of. They sat there. The one asset that reliably brought links was a page of statistics, and it did so without a single follow-up email from us after the first week.
I want to be specific about that, because "create linkable assets" is advice that survives on being unfalsifiable. Ours was a report of skills-based hiring statistics. Last time I checked it had earned somewhere north of a hundred and eighteen referring links, most of them from writers we never contacted. That is not a big number by agency standards. It is a large number for a company our size, and more to the point, it is the only number in our content programme that kept climbing after we stopped working on it.
Why writers link to numbers and not to arguments
Think about who is actually in a position to link to you. Mostly it is somebody writing a post on a deadline who needs one sentence of support for a claim they have already decided to make. They are not looking for your framework. They need a figure with a source attached so their own editor stops asking questions.
Your opinion piece cannot do that job, however good it is, because the writer already has an opinion. That is why they are writing. A number is the one thing they cannot generate themselves.
Aggregation gets you traffic, original data gets you the link
Most statistics pages are collections of other people's research. Those pages rank, and they pick up some links, but the links are fragile. Any writer who follows your citation to the original source will link to the original source instead, and the good ones do check.
At least some of your figures have to be yours. We ran a survey and published the methodology and the sample size next to the findings, which meant there was no upstream source for a writer to defer to. If you have customer data you can aggregate without breaking anything, that is the cheapest original research available to you. You already paid for the collection.
The details that decide whether it works
Give every statistic its own anchor. Writers cite individual figures, not pages, and if they can link to the exact line they will.
Put the year on the page and update the page rather than publishing a new one each January. You lose the compounding if you split the equity across three URLs, though I will admit the year-in-the-slug approach does help with a certain kind of query, so this one is a genuine trade.
Write the citation line for them. One sentence, formatted, ready to paste. It sounds trivial. It removes the last bit of friction between a writer and a link to you.
Do not gate it. I have watched teams put a form in front of their best link asset because somebody needed MQLs that quarter, and the links stopped that month.
What changed with the answer engines
Since large language models started answering hiring questions directly, we have been watching which of our pages get cited inside those answers rather than just ranked beneath them. The statistics report is disproportionately represented, and the reason is mechanical. Clean, attributable, individually parseable figures are the easiest thing in the world for a retrieval system to lift and credit.
So the same page now does two jobs. It earns the links from writers, and it earns mentions in the answers people read instead of clicking. I did not plan for the second job. It happened because the format that made the page easy for a human to cite also made it easy for a machine to cite, which in hindsight is not a coincidence at all.
What this costs
A survey with a real sample takes weeks and some money. The page itself took one writer a few days. Against that, we have not done cold link outreach for this asset since the launch window, and the outreach programme we ran before it consumed more hours per link than I care to write down publicly.
Go and check which page carries your profile
Pull your referring domains and sort by which page they point at. Most companies find that one page carries a wildly disproportionate share, and that the page is not the one they have been optimising. Then look at whether that page contains numbers, and whether any of the numbers are actually yours.
