Internal authority flow

Every internal link is a pipe that carries a share of the linking page's authority to its target. A page divides the authority it can pass across all its outbound links and sends a portion down each. The target receives that portion, adds it to its own, and passes the total on through its links. Run that across the whole site and authority settles into a distribution — some pages rich, some poor — determined entirely by the shape of your link graph.

This is internal PageRank in plain terms: importance propagating recursively along links. It's the same mechanism described mathematically in how RankForge calculates authority flow — this page is the conceptual version you can reason about while restructuring a site.

You control your entire internal graph, and it's the cheapest ranking lever you own — no outreach, no new content required, just links you can add today. Yet most sites distribute authority by accident: the navigation and a sprawling footer decide where equity goes, money pages receive only boilerplate links, and authority earned by one strong page leaks straight back to a feed instead of feeding a target.

[ Home 100 ]
   |        \
  50         50        <- split evenly across 2 outbound links
   v          v
[ Hub 50 ]  [ Page 50 ]
  | | |
 ~16 each              <- a hub linking to 3 children passes
  v v v                   ~1/3 of its (already reduced) equity
[ leaf ][ leaf ][ leaf ]   to each, and so on, decaying per hop

Those two rules — split and decay — are why crawl depth and authority are two views of one problem, and why a focused page that links to four things is a better authority conduit than a mega-menu linking to four hundred.

• chevron_rightOrphans — pages with no inbound internal links receive only a tiny baseline and nothing else. The flow never reaches them.
• chevron_rightMega-menu dilution — a page linking to hundreds of destinations passes a negligible share to each. Spreading authority everywhere is the same as sending it nowhere.
• chevron_rightDead ends and nofollow traps — pages that don't pass authority onward (or whose only inbound links are nofollowed) collect equity and never circulate it.
• chevron_rightRedirect chains and broken canonicals — every extra hop bleeds equity, and split duplicate URLs shatter a page's authority across address variants.
• chevron_rightDeep burial — valuable pages many clicks from any strong source get only the heavily-decayed remainder, regardless of how good they are.

Not all interventions move the numbers equally. Roughly in order of effect per hour spent:

1. Adding contextual links from the highest-authority pages to starved ones. The largest effect available, because you are moving from where the most is held.
2. Trimming bloated global elements. Removing links from a sixty-link footer increases what every remaining footer link carries, across the whole site at once.
3. Shortening paths to important pages. Each hop removed avoids a round of splitting and decay.
4. Fixing dead ends. Giving a section that receives and never passes a few relevant outbound links puts trapped authority back into circulation.
5. Adding more pages. Usually neutral to slightly negative in the short term — the same total is divided further until the new pages earn links of their own.

A per-page authority score invites two mistakes, and both are common enough to be worth naming before you look at one.

Mistake one: treating it as a ranking prediction

Internal authority says which pages your own site supports best. It has nothing to say about relevance, competition, or external links, so a page can hold the highest internal authority on the site and rank nowhere because the query is dominated by stronger domains. The score answers what your structure is doing, not what Google will do.

Mistake two: comparing across sites or crawls

The numbers come from a distribution that sums across whatever was crawled, so they are only meaningful within one site and one crawl. A score of 40 on your site and 40 on a client's site mean nothing to each other. Even on the same site, a crawl that reached more pages spreads the same total across more of them.

What to do instead: read the ordering and the gaps. Which pages sit at the top, which of your commercially important pages sit near the bottom, and how large the distance is between them. That comparison is stable and actionable, and it is where every useful finding comes from.

Authority flow is computed rather than observed, and knowing roughly how keeps you from over-reading the output.

1. Build the graph. Every page is a node; every internal link is a directed edge. What counts as an edge matters — client-rendered links and nofollowed links may or may not be included, and that choice changes the result.
2. Give every page a starting share, so a page with no inbound links is not at zero.
3. Distribute iteratively. Each page passes a share of what it holds along its outbound edges, split between them, with a portion lost at each pass.
4. Repeat until the numbers stop moving. The distribution converges, and the converged values are the readout.
5. Weight by link type. A raw model treats every edge equally; a useful one weights body links above navigation and footer links, because search engines do.

Step 4 is why structure behaves as a system rather than a set of independent pages: because the calculation runs to convergence, one new link changes every page's share slightly. It is also why the same link is worth adding on one site and pointless on another — the effect depends on everything else in the graph.

Modelling flow across a whole site produces a lot of numbers. Almost all the actionable findings fall into three shapes.

Pooling

Authority concentrates in a small set of pages — typically the homepage and top-level navigation items — and circulates between them. It reaches the pages that convert only through long, diluted paths. This is the most common finding on almost any site, and the fix is usually a handful of contextual links from those strong pages rather than a restructure.

Leaking

Authority reaches a page and stops. Dead-end pages, sections that link only inward, and pages with no outbound internal links accumulate what they receive without circulating it. One is harmless; a whole documentation section or a set of terminal landing pages is a meaningful amount parked where it does nothing.

Starving

Commercially important pages holding almost none of the site's authority, usually because they are linked only from navigation. Product, service and pricing pages are the usual victims — everyone assumes the nav is sufficient because the page is easy to reach.

  pooling      strong pages linking mostly to each other\n               [home] <-> [nav] <-> [nav]     money pages: starved\n\n  leaking      authority arrives and stops\n               [strong] --> [docs page] --> (nothing)\n\n  starving     reachable but unsupported\n               [nav] --> [pricing]   (no contextual links anywhere)\n

Starving is the one worth checking first, because it is both the most commercially expensive and the cheapest to fix — a few contextual links from relevant content usually moves it.

• chevron_rightAdd contextual, in-body links from authoritative, topically relevant pages to the targets you want to rank — these are the high-value pipes.
• chevron_rightConcentrate homepage and hub outbound links on a small set of genuine priorities instead of linking to everything.
• chevron_rightKeep important pages shallow — within about three clicks of a strong entry point — so they receive less-decayed authority and get crawled more often.
• chevron_rightFix the leaks: collapse redirect chains, resolve conflicting canonicals, and stop nofollowing internal links you actually want crawled and fed.
• chevron_rightPrune low-value links on important pages. Because equity splits, removing a pointless link makes every remaining one stronger — see anchor text for making the ones you keep count.