Internal Linking: Mechanism Rather Than Architecture
One confusion needs to be cleared up immediately. A site’s architecture determines how pages are structured and their hierarchy. Internal linking, on the other hand, is the mechanism that circulates value between these pages once the structure is in place. The two subjects are complementary; they do not overlap.
This article deals exclusively with the mechanics: how many links, where they are placed, with what anchor texts, to what targets, and how to detect leaks. The question of structure and hierarchy is covered separately in our semantic silo method, which should ideally be read first, as effective linking only occurs within an already clean structure.
The reason this topic deserves separate treatment lies in its effectiveness. It is the least expensive and fastest SEO intervention to implement: it requires no new content, no external link acquisition, and no server intervention. It consists of making better use of what already exists.
Each page has a capital of link equity that it transmits to the pages it points to. A page emitting forty links transmits a portion to each of the forty destinations. The same page emitting six links concentrates that equity. This elementary arithmetic explains why a fifty-item menu prevents internal pages from ranking.
How Link Equity Truly Flows
The mechanism is simple to state, and its consequences are counter-intuitive.
Transmission via Link
A site receives link equity through its inbound links, mostly concentrated on its homepage. This equity is then redistributed step-by-step, dividing at each stage among the destinations. A page reached in four clicks from the homepage receives a very small fraction of what a page reached in one click receives.
This division explains a phenomenon observed in almost all audits: the homepage ranks well, service pages rank moderately, and the rest of the site is invisible. This is not a matter of content quality; it is a matter of distribution.
Links That Transmit Almost Nothing
Not all links are equal, and three categories transmit far less than one might imagine. Navigation links present on all pages are devalued because they provide no contextual information. Footer links are even more so. Finally, automatically generated related article blocks are largely ignored.
Conversely, a link placed within a paragraph that elaborates on the target page’s subject transmits strongly, because its context confirms the relevance of the relationship. This is why we systematically place links within the body of the text rather than in dedicated blocks.
Position on the Page Matters
A link located in the first third of the content carries more weight than an identical link placed at the end of the page. This internal hierarchy is intuitively understood: what a writer mentions early is what they deem important. It has an immediate practical consequence on how to organize paragraphs.
Anchor Texts: The Most Mismanaged Topic
The anchor text is the clickable text. It tells the search engine what the target page is about, and it’s one of the few signals you fully control on your own pages.
What a Good Anchor Text Contains
It describes the destination page and contains the vocabulary of its subject. It integrates into a sentence without distorting it. It is explicit enough for a reader to know where they are going without reading the context, which is also an accessibility criterion.
Empty anchor texts remain the most common mistake. “Click here,” “learn more,” “this link,” and “our dedicated page” transmit no information. They are all the more damaging because they occupy the place of a signal that cannot be compensated for elsewhere.
Variation, Without Over-Optimization
Repeating the exact same anchor text on thirty internal links pointing to the same page produces an artificial signal. Natural variation is preferable: exact anchor text sometimes, partial phrasing often, long expression from time to time.
This variation offers a second advantage. Each different phrasing broadens the lexical field associated with the target page, helping it rank for related queries that no single anchor text would have covered.
The Case of Links to the Homepage
The homepage already receives all navigation links. Adding contextual links from the body of articles is almost always wasteful: it doesn’t need this link equity, and it would be better used on an internal page. The only exception concerns content that genuinely presents the company as a whole.
Orphan Pages: The Invisible Leak
An orphan page receives no internal links. It appears in the sitemap, Google knows about it, but it receives no link equity and behaves as if it didn’t exist.
On a site with over a hundred pages, there are almost always some. They result from the deletion of parent pages during a redesign, content created for a one-off campaign, or articles published outside any site structure.
Detecting them requires cross-referencing two lists: that of known addresses, from the sitemap and Search Console, and that of addresses actually reached by an automated crawl starting from the homepage. The difference between the two constitutes the inventory of orphan pages.
These orphan pages also appear after a poorly prepared migration, a case detailed in our article on the warning signs of a redesign. The fix is quick, and the return is excellent, as it involves reactivating content already written and indexed. In the audits we conduct, this is regularly the action that produces the most visible gain within six weeks.
Dead-End Pages: The Other Leak
Symmetrical to the previous case, a dead-end page receives links but emits none. It accumulates link equity and does not redistribute it, making it useless to the rest of the site.
Thank you pages after forms, legal notices, and certain detailed product sheets often fall into this category. The remedy is to add two or three relevant contextual links, pointing to the pages you want to promote.
A particular case deserves attention: pages that receive many external links. An article that has been widely cited concentrates significant link equity, and the absence of internal outbound links prevents this value from benefiting the rest of the site. Identifying these pages and carefully linking them is one of the most profitable interventions in a linking project.
How Many Links Per Page
No official threshold exists, and the circulating figures are approximations. The useful reasoning is twofold.
From a readability perspective, a reader should be able to follow the text without being interrupted every two words. In a three-thousand-word content piece, about ten contextual links go unnoticed; thirty transform the text into a directory and degrade the experience.
From a distribution perspective, each additional link divides the transmitted link equity. On a pillar page that needs to boost five child pages, adding fifteen links to secondary content reduces what reaches the five priority destinations.
The practical rule is to prioritize rather than count. On each page, identify the two or three destinations you genuinely want to promote, place them early and in a strong context, and treat the rest as navigation convenience.
Building Internal Links While Writing
Internal links are not added as an afterthought during a review pass. They are written with the text, and this difference in method changes the quality of the result.
Specifically, each new page starts with a list of three to five destinations to cover: its parent page, two sibling pages, a service page. The writer then looks for where each naturally integrates. If no suitable place presents itself without forcing it, it’s a sign that the link shouldn’t be there.
This discipline avoids two opposing flaws. Complete oversight, on sites where no one is responsible for internal linking. And mechanical insertion, where links are added at the end of writing to meet a quota, which produces artificial phrasing that readers sense.
Upward Linking: The Forgotten Topic
We instinctively think of downward linking, from the parent page to its children. Upward linking, from children to the parent, is at least as important and much more neglected.
Each child page must point to its pillar page, with an anchor text that contains the pillar page’s vocabulary. These upward links build the pillar page’s position for the most competitive generic query, one that the child page alone could not target.
Lateral linking between sibling pages completes the system. Two to four links to immediate siblings are sufficient, placed where the subject calls for precision. Beyond that, you disperse the equity.
These three directions—upward, downward, and lateral—form the complete flow. A site that practices only one leaves a significant portion of its potential untapped.
Detecting Cannibalization Through Internal Linking
Two pages targeting the same query weaken each other, and internal linking either exacerbates or reveals the problem depending on how it’s done.
The symptom is visible in Search Console: for a given query, two different addresses appear alternately from one week to the next, each plateauing on the second page. Google cannot choose, and neither page progresses.
Internal linking then acts as an arbiter. By systematically directing internal links with the relevant anchor text to only one of the two pages, you clearly indicate which one should carry the query. This is a gentle, reversible correction, often sufficient when both pages have a reason to exist.
When they don’t have distinct reasons to exist, merging is preferable: a single, more comprehensive content piece, the old address redirected, and internal links updated. This operation typically gains several positions in a few weeks.
Tools for Mapping Your Internal Links
Manual mapping is impossible beyond fifty pages. Three families of tools are sufficient.
A complete site crawler explores all pages by following links, exactly like a bot, and produces a list of reached addresses, their depth, and the number of inbound links for each. This is the central tool for any internal linking audit.
Search Console then provides the internal links report, which indicates how many links each page receives according to Google itself. This source is valuable because it reflects what the search engine has actually retained, which sometimes differs from what the crawler observes.
An audience measurement tool completes the picture by indicating which pages actually receive traffic. Cross-referencing the three reveals the most useful inconsistencies: a page that receives many links but no traffic, or vice versa.
The Five-Step Audit Method
Here is the sequence we apply to an existing site, usually over two to three days of work.
- Fully crawl the site and record for each page its depth, and its number of inbound and outbound links.
- Cross-reference with the sitemap to identify orphan pages, then with traffic data to pinpoint underutilized pages.
- Identify poorly served strategic pages: those that should rank but receive fewer links than secondary pages.
- Analyze anchor texts to detect empty anchor texts, excessive repetitions, and inconsistencies between the anchor text and the target content.
- Write the correction plan page by page: which links to add, which to remove, which anchor texts to rephrase.
Executing the plan is then quick, as it involves editing existing content. The first effects are measurable between the fourth and tenth week, allowing time for the site to be recrawled.
Internal Linking from Editorial Content to Commercial Pages
This is the junction point between SEO and conversion, and it is handled with a timidity that proves costly. The logic is the same as that applied to website creation: a site is meant to generate inquiries, not just visits. A blog that attracts traffic without ever linking back to service pages generates notoriety but not clients.
The rule we apply is one to two commercial links per editorial content piece, placed where the reader understands the problem and is looking for a solution. In practice, this is after the section that explains the method, never in the introduction. Placed too early, the link interrupts the demonstration and discredits the content.
The choice of target page matters as much as placement. An article on a technical problem should link to the service that solves it, not to the homepage or the contact form. The closer the match between the content’s subject and the destination page, the more the link converts and the more signal it transmits.
This mechanism also works in the other direction, and that is even rarer. A service page benefits from pointing to two or three content pieces that delve deeper into technical aspects of the subject. This demonstrates expertise, retains visitors in the research phase, and pushes link equity to content that needs it.
Managing Internal Linking Over Time
Internal linking degrades without maintenance, just as surely as a database fills up. Three mechanisms are responsible.
Deleting pages leaves behind all the links that pointed to them. If the page is redirected, the link works but passes through an intermediary, which slightly dilutes the transmission. If it is not, the link is broken, and the visitor lands on an error page. The rule is to update inbound links at the same time as deletion, not just to set up a redirect.
The accumulation of content gradually unbalances the distribution. Each new article adds outbound links from the pages that mention it, thereby diluting existing links. On a site that publishes weekly, the distribution structure after two years bears no resemblance to the initial one.
Finally, changes in site structure invalidate internal linking assumptions. A page that becomes a child of another must have its upward links rephrased, and sibling pages updated. Without this revision, the internal linking describes a structure that no longer exists.
A quarterly review is sufficient to contain these three drifts. It involves re-running a full crawl, comparing it with the previous report, and correcting discrepancies. Half a day per quarter on a medium-sized site, integrated into what a serious maintenance contract covers.
The Most Common Internal Linking Errors
Six mistakes recur on almost all sites we audit, and none require special skills to correct.
Linking only from the menu, thinking navigation is sufficient. Using empty anchor texts on half the links. Placing all links at the bottom of the page, in a block no one reads. Massively pointing to the homepage, which doesn’t need it. Forgetting upward linking from children to the parent. And letting pages receive link equity without ever redistributing it.
A seventh point deserves to be added because it is more subtle: linking without hierarchy. Treating all destinations as equivalent amounts to prioritizing nothing, and the structure then transmits no signal about what matters. Internal linking is a decision-making tool as much as a circulation tool, and refusing to decide negates its purpose.
Internal Linking on an Online Store
E-commerce poses specific problems that editorial linking does not encounter.
The first is volume. With eight hundred product sheets, internal linking cannot be done manually. It is designed at the template level: complementary products, products from the same category, return to the parent category. The quality of the result then depends on the quality of product categorization.
The second is depth. A product sheet accessible in five clicks from the homepage will receive almost nothing. Intermediate categories must be limited, and important products promoted from higher-level pages.
The third is volatility. A sold-out product that disappears leaves behind all the links that pointed to it. A clear policy is necessary: keep the page, indicating unavailability, or redirect to the category. This is a point we systematically address in e-commerce SEO, and it differentiates platform choice, a topic covered in our WordPress vs. Shopify comparison.
What Internal Linking Does Not Replace
A clarification to avoid excessive expectations. Internal linking redistributes existing link equity; it does not create it.
On a domain with no external authority, optimizing internal flow amounts to better distributing a very small sum. The gain exists, but it remains proportional to the available capital. This is why internal linking work is conducted in parallel with authority building, never in its place.
Nor does it compensate for a lack of content. A poorly written page that does not meet the intended search intent will not rank just because it receives links. Internal linking serves pages that deserve to rank; it does not promote others.
Finally, it does not replace any technical prerequisites. A site whose deep pages are not crawled due to a crawling budget issue will not benefit from improved internal linking until server response time is addressed, a topic developed in our article on Core Web Vitals.
Key Takeaways
Internal linking is the mechanism for distributing link equity, distinct from the architecture that determines the site structure. It is the SEO intervention with the best effort-to-result ratio, because it requires neither new content nor external acquisition.
Three checks are enough to know where you stand. Are your strategic pages reachable in three clicks? Are there pages that receive no internal links? Do your anchor texts describe the target page, or do they say “learn more”?
The rest comes down to writing discipline: links within the body of the text rather than in automatic blocks, placed early, with varied and descriptive anchor texts, and an assumed hierarchy between the two or three priority destinations and the rest. The projects we follow in our achievements apply this rule from the design stage.
