How RankForge calculates the Structural Health Score

A structural audit produces dozens of metrics: orphan count, depth distribution, cluster cohesion, authority concentration, anchor quality, cannibalization. Handed all of them at once, most people can't tell whether their site is in good shape or bad, or whether last month's work helped. A single composite score solves two real problems: it gives a defensible at-a-glance verdict, and it gives a trend line you can track across crawls so 'did the restructure work?' has an answer.

The risk of any single score is that it hides its reasoning. RankForge's answer to that is a hard rule: the headline number is always the weighted composite of the dimension scores you can see on the cards below it. There are no hidden caps, secret penalties, or overrides that move the headline away from the visible parts. If a problem deserves to move the score, it does so through a visible dimension — never through an invisible thumb on the scale.

The composite is built from a set of independently scored dimensions, each capturing one structural concern. They map directly to the concepts in the rest of this Academy.

DIMENSION              WHAT IT JUDGES
─────────────────────  ──────────────────────────────────
Authority distribution  Is internal equity concentrated on
                        the right pages, or pooled/leaking?
Crawlability & depth    Can pages be reached, and how deep
                        do important ones sit?
Orphans & connectivity  How many pages get no internal links?
Topical clustering      Are related pages wired into pillars
                        & clusters, or isolated?
Anchor text quality     Do internal links describe their
                        targets, or waste the signal?
Cannibalization &       Do pages compete with each other or
  consolidation         consolidate cleanly?

Each of these has a deeper article: authority flow, crawl depth, orphan pages, topic clusters, anchor text, and cannibalization.

The dimensions are not weighted equally, because they don't have equal impact on whether a site's structure works. Weighting reflects leverage: how much a problem in that dimension actually suppresses rankings and how directly fixing it helps.

• chevron_rightAuthority distribution and connectivity carry more weight because they're upstream — if equity never reaches a page, nothing else about that page matters. A perfectly written, perfectly clustered page that's orphaned still can't rank.
• chevron_rightAnchor text quality carries real but smaller weight: it sharpens an already-existing link rather than determining whether authority arrives at all.
• chevron_rightDimensions only present on some sites are renormalized, not penalized. If a site is too small to have meaningful clusters, the score is computed over the dimensions that do apply — an absent dimension never silently drags the number down.

85–100   Strong. Authority reaches priority pages, few
         orphans, shallow important pages, real clusters.
         Work is fine-tuning.

70–84    Healthy with gaps. A handful of orphans or some
         depth/anchor issues. Targeted fixes, not a rebuild.

50–69    Structurally compromised. Authority pooling, many
         deep pages, weak/absent clusters. Real plateaus
         usually live here.

below 50 Architecture is working against the content.
         Orphans at scale, leaks, near-zero clustering.
         Restructure before producing more content.

Common reasons for a poor score cluster into a few patterns: authority trapped on the homepage and never distributed; a large tail of pages at depth 4+; orphans created by a migration or by faceted navigation; and content published as isolated articles that were never wired into clusters. Notably, none of these are content-quality problems — which is exactly why they're invisible to a conventional audit and why the score surfaces them.

• chevron_rightReconnect orphans with contextual links from relevant, already-linked pages — usually the fastest single improvement.
• chevron_rightFlatten the path to priority pages so they sit within ~3 clicks of a strong entry point, raising both crawl frequency and authority.
• chevron_rightWire isolated content into clusters: supporting pages link up to pillars, pillars link down to key supporting pages.
• chevron_rightConcentrate homepage and hub outbound links on genuinely high-priority destinations instead of spraying them across everything.
• chevron_rightFix the leaks — redirect chains, conflicting canonicals, and nofollow traps that bleed equity out of the graph.
• chevron_rightReplace generic anchors ('click here', 'read more') on important internal links with descriptive ones.

This is the most important section, because the score is easy to over-read.

It is not a ranking prediction

Rankings depend on content quality, search intent, backlinks, competition, and dozens of factors that have nothing to do with internal structure. The Structural Health Score measures one thing well: whether your architecture routes discovery and authority sensibly. A high score removes structure as the bottleneck; it does not promise position one, and a site can score 90 and still rank poorly because its content or backlinks are weak.

It is not a Google score

No search engine publishes a structural health figure. This is RankForge's model, built to be useful and honest, not a leaked Google metric. Treat it as a well-instrumented opinion about your architecture, not a verdict from the search engine.

It is bounded by what was crawled

The score reflects the pages actually crawled. If the crawl hit its page budget, or the site is largely client-rendered, or a mega-menu links everything to everything (which makes depth and authority metrics uninformative), RankForge caveats the affected dimensions rather than reporting confident noise. A score computed over a partial or degenerate crawl says so — honesty under degenerate input is part of the methodology, not an afterthought.

• chevron_rightIt judges structure, not prose. It can't tell you your content is thin or off-intent — only that the page is or isn't well-connected.
• chevron_rightIt sees internal links, not the live link graph Google has built over years; very new or very large sites are approximations.
• chevron_rightWeights are informed judgement, not ground truth. They're tuned against what correlates with healthy sites and will keep evolving.
• chevron_rightA single crawl is a snapshot. The signal compounds across crawls — the trend in your Crawl History is more meaningful than any one number.

Used as intended — a prioritization and tracking tool for the structural layer — the score earns its keep. Start with the pillar guide on structural SEO for the concepts, then run an audit to see your own dimensions broken out.