A website gets evaluated as one system, not as isolated URLs.
Website hierarchy analysis checks a proposed or existing site at whole-site level — service pages, parent categories, location content, internal linking and coverage gaps together — because a URL decision made in isolation misses the relationships that justify or undermine it.
Ten things checked at whole-site level
- Service pages — do they map to confirmed, validated services?
- Parent categories — does every category contain genuinely distinct children?
- Child services — do they earn separation from their parent, or duplicate it?
- Supporting pages — do they serve a page that needs them, or float unattached?
- Location content — does each page reflect a confirmed, distinct service area?
- Page relationships — which pages should link, and why?
- Internal linking requirements — where does a natural link exist and go unbuilt?
- Missing coverage — which confirmed services have no adequate page?
- Potential overlap — which pages compete for the same intent?
- Unresolved structural issues — what doesn't have enough evidence to decide yet?
Why 382 URLs aren't 382 separate decisions
A site with 382 proposed URLs shares one evidence system, not 382 independent ones. A decision about a service page's naming affects every location page beneath it; a merge decision on one pair of pages changes the internal-linking requirements for a dozen others. Hierarchy analysis holds the whole set in view at once, which is what makes a coherent 382-to-1 outcome possible instead of 382 locally reasonable choices that contradict each other site-wide.
A typical layer structure
Every URL has a job
A page earns its place by entity, function, intent and relationship to the rest of the site — not because a keyword exists that a template can fill. A page with no defined job is the clearest early signal of an unjustified URL, before any traffic or ranking data exists to confirm it.
Reducing cannibalisation without guessing
URL similarity alone isn't enough. Keyword similarity alone isn't enough. Being located beside another page in the hierarchy isn't enough. Any one of the three raises a question worth investigating; none of them, alone, answers it — which is why a MERGE or REMOVE decision requires independent evidence beyond structural resemblance.
When the evidence for a merge falls short, the outcome is REVIEW, not a default merge.
Feeding relationship analysis and page intent
Once the hierarchy is scored, relationship analysis decides which pages should link and page intent analysis assigns each surviving page a single primary purpose — both steps depend on the whole-site view established here.
See how linking decisions get made. See how intent gets assigned.