Internal Link Equity: How PageRank Flows Through Your Site

A retailer asked why their highest-margin product category — the one they wanted to rank — sat on page two while a thin “shipping policy” page outranked it for related terms. Nothing was wrong with the category page’s content. The problem was structural: the shipping policy was linked from the footer of all 40,000 pages, and the money category was linked from exactly three. The site had spent years pouring its internal authority into a utility page nobody searches for, and starving the page it actually wanted to win.

This is the recurring failure of internal linking: the link graph you built distributes authority according to its own topology, not according to your business priorities. PageRank is indifferent to which pages you care about. It flows along the edges that exist, concentrating wherever many links converge and draining wherever few do — and on most sites the convergence points are navigation chrome and the drains are the commercially valuable pages buried three clicks deep. Anchor text and external profiles get the attention, but the internal graph is the lever you fully control and most rarely model.

PageRank as Flow Through a Graph

Strip away the search-engine mystique and PageRank is a flow model over a directed graph. Each page is a node; each internal link is a directed edge. Authority circulates through the edges, and at equilibrium each page holds an amount proportional to how much flows in.

The mechanic that matters operationally: a page divides its outgoing authority among its outbound links. The original formulation splits it roughly equally — a page with N outbound links passes about 1/N of its transmissible equity through each — modulated by a damping factor (canonically 0.85) that models the chance a browsing user simply stops. Two consequences fall directly out of this.

First, links dilute. Every additional outbound link on a page reduces what each of the other links carries. A page with 10 thoughtful outbound links passes a tenth of its equity through each; bloat it to 200 links — mega-menus, footers, “related” widgets, social icons, tag clouds — and each link now carries a fortieth as much. The page hasn’t gained authority to give; it has sliced the same authority thinner. This is why a focused page can be a more powerful linker than a link-saturated one.

First-glance counterintuitive, second consequence: equity concentrates at convergence points. A page linked from every other page — the classic sitewide footer or main-nav link — receives a share from all of them, accumulating far more inbound equity than any single contextual link provides. That’s enormous leverage, and the question is only whether it’s aimed at the right targets. Aimed at your hub categories, sitewide links are the strongest tool you have. Aimed at the privacy policy, the careers page, and the login form — the default for most footers — they pour the site’s concentrated authority down a drain.

Modeling the Graph: From Crawl to Flow

You can’t redirect a flow you can’t see, and intuition about a 40,000-page site’s link graph is worthless. The graph has to be built from a crawl and the flow computed over it.

The construction is mechanical. Crawl every internal URL, record every internal link as a directed edge from source to target, and you have the adjacency structure of the site. Then run the iterative PageRank computation over it — seed every node with equal authority, repeatedly redistribute each node’s authority across its outbound edges with the damping factor applied, iterate until the values stabilize. The output is a relative internal-authority score per URL: not Google’s actual PageRank (which folds in external links and signals you can’t see), but a faithful model of how your internal structure alone distributes whatever authority enters the site.

That relative model is enough to find the pathologies, and they jump out once computed. The pages hoarding internal equity are usually the ones you’d never optimize: footer-linked utilities, the homepage (always the top sink), high-volume but low-value tag and filter pages. The pages starved of it are frequently the commercial targets — deep product categories, conversion landing pages, the cornerstone content you built links to externally but never routed internally. The gap between a page’s modeled internal authority and its business value is the actionable signal: a high-value page with low internal authority is an equity-starvation finding, and it’s almost always fixable with a handful of well-placed links.

The same crawl exposes the structural defects that leak equity entirely.

The Four Structural Leaks

Orphan pages. A URL with zero inbound internal links receives no internal equity — it sits at the floor the damping factor allows. It may be indexed if the sitemap surfaces it, but it competes from the bottom with no internal support. Orphans accumulate quietly: a product page whose category link was dropped in a redesign, a campaign page published outside the nav, content migrated without its inbound links. The crawl finds them as the set of indexable URLs (often cross-referenced against the sitemap) with no internal referrers. The fix is contextual inbound links from relevant, well-linked pages — not a footer dump, which dilutes, but in-content links from topically related pages that already hold authority.

Dead-end pages. The inverse: a page with inbound links but no useful outbound internal links — a leaf that receives equity and circulates none onward. Authority flows in and stops. Beyond the equity it fails to pass along, a dead end is a poor user and crawler experience: nowhere to go next. Comprehensive pages with no “related,” “next,” or contextual links are the usual culprits. Adding relevant outbound links keeps the flow moving and the crawler progressing.

Deep pages. Crawl depth — the minimum number of clicks from the homepage — correlates with authority because equity attenuates with each hop through the damping factor. A page eight clicks deep has had authority decay across eight redistributions before any reaches it. Important pages buried deep are structurally disadvantaged regardless of content quality. The remedy is flattening: link valuable deep pages from shallower, higher-authority pages — a hub page, a relevant category, the homepage when warranted — pulling them up the depth gradient.

Equity sinks. Pages that receive heavy internal linking but neither convert nor need to rank: the aforementioned footer utilities, but also faceted-navigation parameter explosions and tag archives that multiply into thousands of near-empty URLs, each soaking up a share of every link pointed at the facet. These don’t just fail to use their equity; they dilute every page they sit beside in the navigation. The fix is selective: nofollow is largely ignored for internal sculpting in the modern algorithm, so the real tools are removing the links, consolidating the targets, or restructuring the navigation so concentrated sitewide equity points at hubs instead of dead inventory.

Source-to-Target Recommendations: The Actionable Output

A list of starved pages isn’t a fix; it’s a problem statement. The useful output of internal-link analysis is a set of specific source → target recommendations: for each high-value, low-authority target, which existing pages should link to it. That turns the finding into an edit.

The selection logic is two-part. The source must have authority to give — there’s no point linking a starved page from another starved page; you route equity from pages the model already ranks highly. And the source must be topically relevant to the target, because a contextually relevant in-content link both passes more useful signal and serves the user, whereas a forced link from an unrelated page is noise. So the recommendation engine ranks candidate sources by their modeled internal authority and their topical proximity to the target, and surfaces the best handful: “link your high-authority buying-guide article to this starved category page, with relevant anchor text, from within the body content.” That’s a concrete, defensible edit — not “improve your internal linking.”

Anchor relevance matters here, and it’s worth separating from external anchor-text strategy. The concern with external backlinks is the distribution of anchors across a profile — over-optimization, exact-match ratios, the footprint of a manipulated link-building campaign. Internal anchor relevance is a different, simpler thing: the anchor on an internal link is a topical signal you fully control, describing the target to both the crawler and the user, and it should accurately and naturally describe the destination. “Premium hiking boots” pointing at the hiking-boots category is correct internal practice; the external-profile worries about exact-match saturation don’t apply to your own contextual internal links, which are expected to be descriptive. Get the source and target right, write an honest descriptive anchor, and the link does its job.

The Practical Sequence

Crawl the site and build the directed link graph — every internal link as an edge. Compute relative internal PageRank over it to get a modeled authority score per URL. Then overlay business value: rank your pages by importance (revenue, conversion, strategic content) and compare against modeled authority. The high-value, low-authority set is your starvation list, ordered by the size of the gap.

Run the structural checks alongside: orphans (zero inbound), dead ends (zero useful outbound), excessive depth on valuable pages, and equity sinks hoarding footer-concentrated authority. For each starved target, generate source-to-target recommendations — high-authority, topically relevant pages that should link to it — and make the edits with descriptive anchors. Re-crawl and recompute to confirm the modeled authority moved toward the pages you intended, then watch the field rankings over the following weeks, since internal-structure changes take crawl-and-reprocess cycles to register.

VisibilityIQ models the internal link graph from the crawl, computes relative internal authority per URL, and reconciles it against page value to flag equity-starved high-value pages, orphans, dead ends, excessive crawl depth, and equity sinks — then produces specific source-to-target link recommendations ranked by source authority and topical relevance, so each finding ships as a concrete edit rather than a graph the user has to interpret. The analysis treats internal linking as the controllable distribution layer it is, distinct from the external anchor profile.