Why One Page Earns Links and Another Doesn't
A link gets added at a very specific moment. Someone is writing something — an article, an answer, a support doc, a lesson plan — and they hit a point they don’t want to explain themselves. So they hand it off: they name the thing, and they point at a page that already explains it. That handoff is the entire mechanism.
Everything else about earning links follows from that moment. If you want more links, you are not trying to persuade anybody of anything. You are trying to be the page somebody reaches for when the handoff happens.
The four conditions
For your page to catch a link, four things have to be true at the same time. Miss any one and the other three are wasted.
1. A need exists. Somebody has to be writing about your topic often enough that handoff moments actually occur. This is the condition people skip, and it is the one that kills most link-building projects. If nobody publishes about what you do — no trade writers, no bloggers, no forum answerers, no course authors, no internal documentation teams — then there is no moment for your page to be chosen at. You can publish the best page in the world about a topic nobody writes about and collect nothing. Before you build an asset, find the conversation it would be cited inside.
2. Your page is findable at the moment of need. The writer needs a source now, mid-draft, and they will spend about a minute finding one. In practice that means one of two things: your page ranks for the phrasing they’d search, or that particular person already knows your page exists. Nothing else reaches them. A brilliant page on page four of the results does not participate in the handoff.
3. It is safe to vouch for. A link is a small public endorsement. When you link out, you have staked a bit of your own credibility on the page being right. So writers screen fast, on cheap signals: is it accurate as far as they can tell, is it current, is there a plain author or organisation behind it, does it read like it was written to inform rather than to sell? A page that’s obviously a sales pitch with an explanation wrapped around it fails this screen even when the explanation is good. So does an undated page about something that changes.
4. It is linkable. This one is mechanical. The writer needs to point at the specific thing — so the specific thing needs to be a page. One page about one topic, at a stable address, with a heading that matches what they’re claiming. If your explanation of a concept is section four of a long guide, the writer’s options are to link the whole guide (vague, so they often don’t bother) or to link a fragment (fragile, and most people don’t). If the only obvious target on your site is your homepage, nobody links, because “see their homepage” isn’t a citation.
Most link problems are condition one or condition four
“We publish good stuff and nobody links to it” is almost never an effort problem. It is almost always a mismatch.
Either nobody is writing about the thing — condition one — in which case more effort on the same asset changes nothing, and the fix is to move to a topic that sits inside an active conversation. Or the material exists but isn’t pointable — condition four — in which case the fix is often just splitting: one narrow page per idea, each with a title that states the idea, instead of one enormous guide that contains all of them. That second fix is unglamorous and frequently doubles the number of citable surfaces on a site without a word of new writing.
The asymmetry you can’t get around
You cannot make anyone link to you. There is no lever. The decision belongs entirely to the other person, it happens in a moment you will never observe, and by the time you could ask, the choice is made and published.
What you can do is be the best available option at that moment. That’s the whole strategy, and it’s a genuinely different activity from persuasion. Persuasion asks “how do I convince this person?” Availability asks “when someone in this situation needs a source, what are their options, and can mine be the obvious one?” The second question has answers you can act on today.
It also explains why depth beats breadth. Being the obvious citation for one narrow thing means that every time that thing comes up, you are the default. Being a mediocre option for forty things means you’re the default for none of them, because there is always something more specific. The goal is not coverage. The goal is to be unimprovable on a small number of questions — the page a writer reaches for without pausing to compare.
About the evidence
Be sceptical of anyone, including us, who tells you exactly what earns links. Nobody publishes reliable data on citation behaviour: the decision happens privately, search engines say almost nothing about how they weigh links, and the industry’s “studies” mostly measure what already ranks rather than why anyone linked. What’s above is reasoning from how writing actually works — the mechanics of needing a source and reaching for one. It holds up, but treat it as a model, not a measurement. If a claim about links comes with a suspiciously precise number attached, assume it was invented.
The one thing worth stating flatly: a link either exists or it doesn’t, and the only durable way to get one is that a human decided your page was worth pointing at.
What to do next
Work the four conditions in order, because they’re sequenced.
- Start with condition one. Identify who publishes about your topic and what they need. That’s the subject of who actually links to things, and it should shape what you write before you write it.
- Then fix condition four on what you already have. Split the buried explanations out into their own pages, one idea each. Making pages easy to cite covers the details.
- Then build assets aimed at a real need: original data, small tools, diagrams, definitions, documentation, worked examples, templates. Each of those genres has its own post in this cluster.
- Then, and only then, tell people. The outreach side is covered in how to earn backlinks without buying them — but outreach on top of an unlinkable page is just email.