Site Architecture
How a site’s pages are grouped and linked: sections, URLs, navigation, breadcrumbs and internal links.
Reviewed by Alexander Yarovenko · Updated: 2026-10-04
Site architecture is the way a site’s pages are grouped and connected: sections, URL paths, navigation, breadcrumbs and internal links. Google says a logical organization can help search engines and users understand how pages relate. The documentation sets no ideal click depth and no required shape, so architecture is judged by whether people and crawlers can reach and understand each important page.
What it is and where the boundary lies
Architecture is a design of relationships between pages, expressed in four visible layers: the URL path, the menu, the breadcrumb trail and the links inside content. It is not a single setting. Google also cautions against reorganizing in a hurry: search engines will likely understand your pages as they are now, so a rebuild must have a clear reason, and URL changes need redirects.
How it differs from similar terms
| Concept | Scope | Typical question |
|---|---|---|
| Site architecture | the whole structure | how are sections and pages organized |
| Internal linking | links between pages | does every important page get links |
| URL structure | the addresses | are paths readable and consistent |
| Click depth | distance from the home page | how many clicks to reach a page |
| Breadcrumbs | a navigation trail | where is this page in the hierarchy |
Why it matters
Crawlers discover pages by following links, and visitors move through menus and trails. Google notes that for sites with more than a few thousand URLs, grouping similar pages in directories can help it learn how often each directory changes. For small sites the effect is mainly for readers: a clear structure shows what you cover and where to go next.
What Google documents and what it does not
- Documented: descriptive URLs with readable words, a simple URL structure, crawlable links (
<a href>), directories that group topically similar pages on large sites, breadcrumbs derived from URL words and optional markup. - Not documented: a “three-click rule”, a perfect depth, a mandatory silo or a ratio of categories to pages. Those are industry conventions; test them against your own data.
How to apply it
- List the page types: home, section hubs, detail pages, supporting articles, utility pages.
- Draw the hierarchy on paper before touching URLs: each page gets one parent topic.
- Make URLs readable and stable, in the audience’s language, without long IDs; keep parameters in the common
key=value&key=valueform. - Link hubs to children and children back to hubs with real
<a href>links, so the structure is visible without scripts. - Add breadcrumbs that mirror the hierarchy, and markup if you want to describe them explicitly.
- Keep every indexable page reachable by at least one link from another page (no orphan pages).
- Control endless variants such as faceted navigation, then check the result in a crawl.
Practical example
A crawl of a mid-sized shop from the home page gives this table:
| Finding | Likely cause | Check |
|---|---|---|
| products in the sitemap but never found by the crawl | no internal links, listing built by script | search the rendered HTML for <a href> |
| the same product under three paths | products attached to several categories | choose one canonical path and link to it |
| thousands of filter combinations | unrestricted parameters | review parameter handling and internal links to them |
Each finding has a single owner and a single fix, which is what turns an audit into a change list.
Common mistakes
- Links built only with scripts: Google reliably parses
<a>elements with anhref; other formats may not be extracted. - Fragments as URLs: content switched with
#fragments is generally not supported; use real URLs or the History API. - Reorganizing for its own sake: without a reason, a rebuild risks lost links and redirects to maintain.
- Treating click depth as a ranking rule: the documentation gives no number.
- Several paths to one page: choose one and use it everywhere.
- A menu that differs from the URL hierarchy: readers and breadcrumbs get conflicting signals.
How to validate the result
Crawl the site from the home page and compare the discovered URLs with the sitemap and with Search Console’s indexing reports: missing pages usually point to missing links. Inspect a few key URLs in Search Console to see the rendered HTML. After a restructure, watch redirects, indexed page counts and the Performance report across a full reporting period.
More questions
Should the structure be flat?
Google gives no depth target. Keep important pages reachable and the hierarchy logical; the number of clicks follows from the content.
Do I need breadcrumbs?
They help readers see where they are. Google learns breadcrumbs from URL words and also supports markup.
Does URL structure matter for a small site?
Readable URLs help people; Google says search engines will likely understand the pages anyway. Do not change working URLs without need.
Next practical step
See how to plan the structure in the lesson Site architecture, then crawl your site and mark every page that has no incoming link.
Related concepts
- Internal linking — links that build the structure.
- Content hub — a hub page with related pages.
- Hub and spoke model — a common content layout.
- Crawl depth — distance from the starting page.
- URL parameters — a source of endless variants.