After this lesson you should be able to
- Explain what a hyperlink is at the level of the HTML that produces it
- Describe how a search engine discovers a page it has never seen
- Separate a link's role as a path from its role as a signal
- State accurately what passing link equity does and does not mean
A link is a citation
Strip away twenty-five years of industry language and a hyperlink is one thing: a citation. One document points at another and says, in effect, this other thing is relevant to what I am talking about. That is all the HTML does. A link is an anchor element with an href attribute holding a destination, and some text or an image inside it that a person clicks.
The academic parallel is not decoration, it is the actual intellectual origin. Citation analysis — counting which papers get cited, by whom, and how often — predates the web by decades. The insight that made web search work was that the web has the same structure. Pages cite each other. If you can read all the citations, you can say something about which documents the rest of the corpus treats as worth pointing at.
Hold on to the citation framing, because it survives every algorithm change. When you are trying to decide whether a link is worth pursuing, the useful question is not "what is the domain metric" but "would a reasonable editor cite this page here, for this reason, if search engines did not exist?" That question answers a surprising number of hard cases correctly.
How a search engine finds anything at all
A search engine has to solve a problem before it can rank anything: it has to know your page exists. There is no registry of all web pages. The primary mechanism for discovery is following links.
The process, simplified but not misleadingly so:
- Crawling. A program — Googlebot, Bingbot, and others — requests a page it knows about, reads the HTML, and extracts every
hrefit finds. - Queueing. Destinations it has not seen go into a list of things to fetch. Destinations it has seen get scheduled for a revisit, sooner or later depending on how important and how frequently changed the crawler judges them to be.
- Indexing. Content that passes the engine's quality and duplication filters gets stored in a form that can be retrieved for a query. Crawled is not the same as indexed. Plenty of pages are fetched and then never included.
- Ranking. At query time, the engine selects from what it has indexed and orders the results.
Two consequences follow immediately. First, a page with no internal links pointing at it and no external links pointing at it is, from the crawler's point of view, invisible unless you tell the engine about it directly through a sitemap or a submission. Second, the frequency with which your pages get recrawled is not uniform — it responds to signals, and links are among them. This is why a new site with genuine external links tends to get indexed faster than a new site without them, and why the first practical benefit of link building often shows up as pages appearing in the index rather than as rankings moving.
Internal links do this job too, and they are free. I have lost count of the audits where the fastest available win was not an external link at all — it was linking to an orphaned category page from somewhere in the site's main navigation.
A path and a signal at the same time
Every link does two distinct jobs, and confusing them is the source of a lot of bad decisions.
As a path, a link is something a crawler walks and a person clicks. Its value here is measured in traffic and discovery. A link on a busy page in front of the right audience sends real visitors, and those visitors do not care what any metric says about the domain.
As a signal, a link is something an engine counts and weighs. Its value here is measured in whatever contribution it makes to how your pages are assessed. That contribution is invisible, indirect, and impossible to isolate for a single link.
A link can be excellent as a path and worthless as a signal. A link from a major publication that carries a nofollow attribute — a marker telling engines not to treat it as an endorsement — may drive thousands of visits and considerable brand exposure while contributing nothing that the engine counts as a vote. A link can also be the reverse: technically followed, on a page nobody visits, contributing something to the graph and nothing to your business.
When somebody asks me whether a link is "good", my first question back is which job they want it to do. A campaign built to earn signal links and judged on referral traffic will look like a failure. A campaign built for reach and judged on domain metrics will look like a failure too. Both were measured against the wrong job.
What "passing equity" does and does not mean
Link equity — also called link juice, a term I would happily retire — is the industry's shorthand for the ranking value that flows along a link from the linking page to the linked page. The concept is real. The mental model most people have of it is wrong in three specific ways.
It is not a fixed quantity you receive
There is no bucket of authority on a page that gets divided and handed out in fixed amounts. What exists is a set of computations over a graph of documents, run repeatedly, in which a page's assessed importance depends on the assessed importance of the pages linking to it. It is recursive and iterative, not a transfer of a substance. Nobody outside the engine knows the current weighting, and it changes.
Not every link contributes
Links can be discounted or ignored for many reasons: an attribute on the link, a pattern the engine has learned to distrust, a page that is not indexed, a placement in a template rather than in editorial content, a source the engine has classified as unreliable. "I got a link" and "the engine counted a link" are different claims, and only the first one is something you can observe.
The relationship to page count is real but crude
The original formulation did divide a page's contribution among its outbound links, which is where the folk advice about avoiding pages with hundreds of outbound links comes from. The direction of that advice still holds — a link in a list of four hundred is worth less than a link in a paragraph of a focused article — but treating it as arithmetic you can compute is false precision. Modern systems weigh links by far more than a count.
What is safe to say: a link from a relevant, well-regarded, genuinely editorial page tends to help more than a link from an irrelevant, low-regard, or templated one. Everything beyond that is estimation.
Where beginners get this wrong
Four mistakes, in rough order of how often I see them.
- Treating all links as interchangeable units. A hundred links from one site is not a hundred times a link from one site. On the profile I use as a reference throughout this course, a single domain accounts for 219,159 links — 16.8% of the entire profile — because of one sitewide placement. Counting links there would tell you something wildly untrue about the site.
- Assuming a link works the moment it goes live. The engine has to crawl the linking page, index it, and process the link. That takes time, and sometimes it never happens because the linking page itself is never indexed.
- Ignoring the path job entirely. If nobody ever clicks a link, ask yourself honestly why an engine should treat it as a meaningful citation. That question is uncomfortable for a lot of link building, and it should be.
- Believing links are permanent. They are not, and the rate at which they disappear is one of the most under-appreciated numbers in this field. Lesson 1.5 puts a figure on it.
Questions
Does a link have to be clicked to count?
No. Search engines process links they discover during crawling regardless of whether any human clicks them. But clicks are evidence that a link is genuinely useful, and links nobody ever clicks tend to sit on pages that are themselves poorly regarded. Treat click-through as a proxy for quality even though it is not a direct ranking input you can point to.
Is an internal link a backlink?
Not in normal usage. A backlink is a link pointing to your site from a different site. Internal links — page to page within one domain — do real work for crawling, indexation and how importance is distributed inside your site, and they are entirely under your control. They are just not what anyone means by a backlink, and tools count them separately or not at all.
Does a link from a page that is not indexed do anything?
Very little, and possibly nothing. If the engine has not indexed the linking page, it may still have crawled it and seen the link, but a page excluded from the index is usually excluded for reasons that also make its outbound links uninteresting. Checking whether a linking page is indexed is one of the fastest quality filters available to you.
Do redirects pass link equity?
A permanent redirect passes most of what the original link contributed, and Google has stated for years that it does not lose PageRank through 301s. That does not make redirect chains harmless: each hop adds crawl cost and risk, and redirecting an irrelevant expired domain at your site is a manipulation pattern engines have handled for a long time. Use redirects to preserve real history, not to import authority.