Contextual Links: Why In-Text Beats Navigation

A contextual link is a link placed inside the body prose of a page, within a sentence, surrounded by relevant text, put there by an editorial decision. Navigation, footer, and sidebar links are boilerplate: they appear identically on every page regardless of content. The evidence and reasoning both point the same way: contextual links carry far more signal to search engines and far more value to readers, which is why an internal linking strategy measured in nav links is mostly measuring nothing.

What counts as contextual

The test is simple: would this link exist if this specific paragraph didn't? A link inside a sentence about crawl budget, pointing to a crawl budget guide, exists because of that sentence. A footer link to the crawl budget guide exists because someone edited a template once.

Contextual (in-text) links:

• Links within body paragraphs, wrapped around a phrase in a sentence.
• Links inside content-specific elements a writer placed deliberately: a comparison table cell, an image caption within the article.

Not contextual (boilerplate and near-boilerplate):

• Header navigation and mega-menus.
• Footer link lists.
• Sidebar widgets: including automated "related posts" and "recent posts" blocks. These are content-adjacent but algorithmically generated and template-positioned; they're better than nothing for crawl discovery, but they aren't an editorial statement.
• Breadcrumbs (useful for hierarchy and structured data, but boilerplate by nature).
• Tag clouds, category archives, "posts by this author."

The gray zone is the hand-curated "further reading" list at the end of an article: chosen per-page (editorial), but positioned outside the prose (structurally like a widget). Reasonable people file it either way; it's clearly better than an automated widget and clearly weaker than an in-sentence reference.

Why in-text links carry more signal

We should be precise about epistemic status here: Google doesn't publish its link-weighting formulas. But the reasoning below is grounded in things that are public: a granted Google patent, consistent public statements that anchor text and context matter, and mechanics anyone can verify. Together they make the conclusion about as safe as inferences get in SEO.

1. The surrounding text describes the relationship

A link is a labeled edge between two pages, and an in-text link gets three layers of labeling: the anchor text itself, the sentence around it, and the topical context of the whole source page. When your paragraph about espresso extraction links to a grind-size guide on the words "grind size," the link arrives packaged with meaning: this page, on this topic, considers that page relevant to this exact concept.

A footer link has none of that. Its "surrounding text" is other footer links; its anchor is usually a bare page title; and it appears on every page, including hundreds with no topical relationship to the target. There's simply less information in it, and search engines have long been better at extracting meaning from text than from bare graph edges.

2. Editorial choice is a scarce, costly signal

Boilerplate links are free: one template edit creates ten thousand of them. In-text links each cost a decision; someone had to be writing about the subject and choose to reference that page. Scarce, costly actions are informative; free, mass-produced ones aren't. This is the same logic that makes one earned editorial backlink worth more than a thousand directory listings, applied inside your own site.

3. The reasonable-surfer model

Google's "reasonable surfer" patent (Ranking documents based on user behavior and/or feature data, US 7,716,225, granted in 2010, originally filed in 2004) describes weighting links by the probability that a user would actually click them, based on features like position on the page, whether the link is in body text or a footer, font size, anchor text, and more. Under such a model, a prominent in-prose link in the main content earns more weight than a link buried in a footer among fifty others; because a reasonable visitor is far more likely to click the former.

The honest caveats: a patent describes something Google invented, not necessarily the exact system running today, and the details have surely evolved over two decades. But the patent's core idea (not all links on a page are equal, and placement plus context determine weight) is the concept to internalize, and everything Google has said publicly since is consistent with it. Boilerplate links being discounted is about as close to consensus as this field gets.

4. The mechanics of link equity

Whatever weighting applies, a page distributes its link equity across its outgoing links. A template that puts 80 boilerplate links on every page means every page's equity is heavily spent on the same sitewide destinations before its handful of in-text links get their share. That's a second, purely mechanical reason contextual links on lean templates punch above their weight; and a reason "add more footer links" makes structure worse, not better.

5. Readers behave the same way

None of this is only about algorithms. An in-text link answers a question the reader has at that moment, in the middle of the sentence that raised it. Footer links serve the rare visitor who scrolls to the bottom looking for a sitemap. If you had analytics on every internal link (worth setting up), you'd see the same asymmetry the reasonable-surfer model assumes: body links get clicked, boilerplate mostly doesn't.

The failure this explains: the "well-linked" orphan

Put the pieces together and a common audit blind spot becomes obvious. A crawler that counts all links reports that every page on your site has 60+ inbound links: nav, footer, category archives. Structurally fine, apparently. But filter to in-text links only and the picture inverts: a large share of pages, often the majority of a blog archive, have zero inbound contextual links. Nothing in any article's prose references them. They're orphans in the graph that matters.

Those pages are crawlable but uncorroborated; nothing on the site editorially vouches for them, no anchor text describes them, no reasonable surfer paths lead to them. When people say internal linking moved the needle for a site, resolving this (turning boilerplate-only pages into contextually referenced ones) is usually the needle that moved. The full playbook for doing this systematically is our complete guide to internal linking.

Practical ratio advice

How many contextual links, and what balance against boilerplate? Working rules that hold up:

• Density: roughly one in-text link per 200–300 words of prose. Enough that every substantial section connects somewhere; not so many that the page reads like a wiki. (LinkAgent's own balancing targets ~1 per 250 words with a cap of 8 per page: numbers chosen to stay comfortably inside what reads naturally.)
• Every page should have at least 2–3 inbound in-text links. Zero is an orphan; one is fragile. Cap the other direction too; funneling dozens of new inbound links at one target in a burst looks like what it is.
• Vary anchors across a target's inbound links. Ten inbound links are a profile; ten identical exact-match anchors are a pattern. Descriptive, varied phrasing (the craft covered in our guide to anchor text) is both safer and more informative.
• Keep templates lean. Every navigation item and footer column is spent from all pages simultaneously. A 40-link footer isn't a linking strategy; it's a tax on every page's equity.
• Prioritize adding contextual links to pages that only have boilerplate inbound links: the highest-leverage fix in most audits, because you're taking pages from zero editorial endorsement to some.

The workflow that makes the ratios sustainable: whenever prose mentions a concept another page covers, link it there, using words already in the sentence as the anchor; no retrofitted phrasing, no bolted-on "check out our guide" sentences. At archive scale, that "whenever prose mentions" step is exactly what's worth automating with an internal linking tool that finds the mention-without-link pairs across every page and proposes the anchor from the existing sentence, leaving you the approve/reject call.

The bottom line

Contextual links beat navigation links because they carry information navigation can't: surrounding text that explains the relationship, an editorial decision that vouches for the target, and a position real readers actually click; the intuition Google formalized in the reasonable-surfer patent. Treat boilerplate links as plumbing for reachability, treat in-text links as the actual signal, audit the two separately, and spend your effort where the signal is: getting every page you care about referenced, in prose, with anchors that describe it.