Skip to content
search-engine-optimizedby Netsolution

Internal linking

A link is only followed if it is an anchor element carrying an address: not a click handler, not a framework-specific attribute. The text of that link has to describe the page it points at. Internal linking decides what is reachable, and therefore what can exist in an index.

A far narrower definition than the one a developer has in mind, and it is the compulsory starting point of the subject.

A link is only crawlable if it is an HTML anchor element carrying an address attribute.

officialSource : Google Search Central, “Link best practices for Google”

Everything that looks like a link without being one therefore does not count: a clickable element calling a function, a button that changes the display, a navigation attribute belonging to a framework. To a visitor the difference is invisible — they click, the page changes. To a crawler, one is a road and the other a wall.

It is the most expensive defect on a site built like an application, and the hardest to spot, because nothing signals it: the site works perfectly, it is simply unreachable in part.

An engine discovers addresses by following the links it meets, and it does not download all the ones it discovers.

officialSource : Google Search Central, “In-depth guide to how Google Search works”

An old manual telephone switchboard: rows of labelled jacks up the full height of the panel, and three patch cords stretched upwards.
Every jack carries its label, and nothing connects itself: somebody decides to plug that cord, between this point and that one. Internal linking is built the same way, one link after another.Djembayz — CC BY-SA 3.0 · Wikimedia Commons, relevé le

Good link text is descriptive, reasonably concise, and relevant both to the page carrying it and to the page it points at.

officialSource : Google Search Central, “Link best practices for Google”

Three requirements, of which the third is the most interesting: relevant to the page carrying it as much as to the page it points at. A link is not merely an arrow, it is a sentence saying why you would leave here to go there. It is also why lists of links at the end of an article, with no context, are worth less than a link placed inside the reasoning that justifies it.

The documentation names the phrases to avoid:

The documentation advises against generic phrases such as “click here” or “read more”, and against piling keywords into link text.

officialSource : Google Search Central, “Link best practices for Google”

The test it offers is excellent and needs no tool: read the link text on its own, outside its sentence. If it does not say where you are going, rewrite it. “Read more” fails, “click here” fails, and a pile of keywords fails too — for the opposite reason, it says too much to be readable.

One useful special case:

When a link is carried by an image, that image's alternative text serves as the link text.

officialSource : Google Search Central, “Link best practices for Google”

A clickable thumbnail with no alternative text is therefore a link with no text: the same loss of information as “click here”, in a more discreet form.

Writing link text that is understandable outside its sentence

hypothesis

Effort
low
Time to effect
weeks
Risk
none
Durability
structural

A page no internal link points at depends entirely on the sitemap to be discovered, which makes it the most fragile page on a site.

hypothesis

That “hypothesis” status is owned up to: the documentation does not talk about “orphan pages” under that term. The reasoning rests on two documented facts. Discovery happens by following links, and the sitemap commits to neither crawling nor indexing. A page that depends only on the sitemap therefore has one access path, and that path promises nothing.

There is a second cost, this one editorial. A page with no inbound link is a page nobody on the site thought worth citing. That is often a sign it had no place, and finding it orphaned is an occasion to decide: link it properly, or merge it with the page that should have cited it.

This site applies the rule to itself and checks it at build time: a guide no other guide cites fails the build. That check was added after two orphaned pages slipped through — the check that counts is the one that reads the graph backwards, looking for the pages nobody points at.

Three simple rules are enough, and they have nothing to do with a quota.

The first: link when the reader needs it, at the point where they need it. A link placed at the moment the question arises is followed; the same link relegated to the bottom of the page is not.

The second: link what complements, not what resembles. Two very close pages linking to each other mainly signal that they should have been one page.

The third: guarantee a descending path. From the home page, every important page has to be reachable by a sequence of links a human would follow naturally. If that path does not exist, the problem is not the linking, it is the site plan.

What does not work, or no longer works

Place exactly three exact-match internal links per article, and avoid linking to your secondary pages so as not to dilute your link juice.

The numbered quota is published nowhere, exact-match anchors are the keyword piling the documentation advises against, and the fear of “diluting” leads to leaving pages with no inbound link — that is, to manufacturing the one linking problem that can actually be documented.

One last habit to revisit: treating internal linking as an end-of-drafting operation, done in passing. Links are the structure of the site, not its decoration. Deciding what a page cites means deciding what it assumes known and what it delegates — which is an editorial choice, and it is made while writing.

Sources

  1. Link best practices for Google — Google Search Centralaccessed on
  2. In-depth guide to how Google Search works — Google Search Centralaccessed on
  3. Learn about sitemaps — Google Search Centralaccessed on
  4. Optimize your crawl budget — Google Search Centralaccessed on

Block revalidated on

Frequently asked questions

Is a JavaScript link followed?
The documentation only guarantees the anchor element carrying an address. A click handler or a framework-specific attribute is not parsed reliably. Navigation that produces no real links makes part of the site invisible without anything signalling it.
How many internal links should a page have?
No number is published, and the useful question is different: is every important page reachable from another relevant page? A link count says nothing; a page with no inbound link, on the other hand, says something precise and bad.
Should the keyword be repeated in every link text?
No, the documentation explicitly advises against piling keywords into link text. The criterion it keeps is descriptive: the text has to stay understandable read on its own, and be relevant to the page you leave as well as the page you reach.
Does a sitemap replace internal linking?
No. It helps discovery, and the documentation states that it commits to neither crawling nor indexing. A page present only in the sitemap is connected to nothing, which makes it fragile and deprives it of the context its inbound links would give it.