Website structure: an SEO guide

What is a website’s structure?

A website’s structure is the organization that arranges its pages in levels, like the branches of a tree. The home page is the root; then come the main sections, then the subsections, then the detailed content.

It is drawn as a diagram before building or redesigning a site. Once live, it takes shape in four ways: the menu, the breadcrumb trail, the URLs and the internal links. These four elements should tell the same story. When they diverge, visitors get lost, and so do search engines.

Why does site structure matter for SEO?

Google says so in its guide to ecommerce website navigation structure, updated on December 10, 2025. It analyzes the relationship between pages from the links that connect them. The more internal links a page receives, the more important it appears compared with the others. The same guide recommends a logical hierarchy. The menu leads to categories, categories to subcategories, then to detail pages.

Two technical constraints follow. First, Google only reliably follows links coded as <a href> (Google, crawlable links). A menu built with clickable JavaScript elements can therefore remain invisible. Second, Googlebot types nothing into the internal search engine. A page reachable only through a search does not exist for it.

The structure also shapes how authority is distributed. A page linked from the home page and the menu receives more than a page buried at the fifth level. The mechanism is detailed in the guide to PageRank.

Flat or deep site structure: which should you choose?

A flat structure shows many sections from the first level; a deep structure shows few, but multiplies the levels. The Nielsen Norman Group, a reference in web usability, sums up the trade-off in a 2013 article on flat and deep hierarchies. A flat structure works when categories are distinct and recognizable. A deep one becomes necessary when there are too many to fit on a single level.

Is the 3-click rule reliable?

The “3-click rule” comes up in almost every guide. No published study supports it, the same institute points out in a 2019 article devoted to the 3-click rule. A test by Joshua Porter, in 2003, found no link between the number of clicks and abandonment or satisfaction. What matters is that each click clearly brings the user closer to their goal.

Benchmark	Status	Recommended use
Strategic pages 3 clicks or fewer from the home page	Agency convention, not documented by Google	Alert threshold in a crawl, not an absolute goal
Beyond 5 clicks, a page is almost invisible	Practitioners’ observation	Priority for moving up through internal linking
Main menu of 5 to 10 items	Agency convention	Avoid dilution and overload
No orphan strategic page	Documented common sense: without a link, no discovery through crawling	Systematic check

Depth is measured in clicks from the home page, with a crawler, not in the number of directories in the URL. A short URL can hide a page six clicks away, and vice versa.

What is an SEO silo?

An SEO silo groups pages on the same topic in one branch of the site structure. Internal links stay concentrated within that branch. The idea: strengthen the perceived topical consistency and direct authority toward the silo’s parent page.

A strict silo almost forbids links between branches. That is its limit: a reader moving from technical SEO to content needs a bridge. The semantic cocoon, a method formalized in France by Laurent Bourrelly, refines the idea with links organized according to meaning and the reader’s journey. The guide to the semantic cocoon compares silo, cocoon and cluster.

How do you build a website’s structure?

1. Take stock of content and demand: existing pages, offers, customer questions, real queries. Assign each search intent to a single page (search intent).
2. Group into topic areas: a topic area can start from the product or from the customer’s need. Each group gets a parent page, reachable from the menu.
3. Limit the levels: three levels are enough for most sites. A subtopic can start as a section of a page before becoming a page of its own.
4. Set the URLs: short, readable, in words of the audience’s language, separated by hyphens (Google, URL structure). Decouple the URL from the title so that a reorganization breaks nothing.
5. Set up navigation: menu, HTML breadcrumb trail marked up with BreadcrumbList, links to sibling pages.
6. Add contextual links: links in the body of the text, which connect the branches. The method is detailed in the guide to internal linking.
7. Test before going live: crawl the staging site, measure depth, look for orphan pages.

Example: the Elev8 Lab site structure

This site applies the method. Its structure has 6 topic areas in the menu, 14 pillar pages and 3 levels at most: topic area, pillar, sub-pillar. Addresses stay short and stable. On the French version of the site: pillars at the root (/seo/), sub-pillars below them (/seo/technique/), articles under their pillar (/seo/maillage-interne/). Moving an article from one sub-pillar to another therefore does not change its address.

Site structure, URL, menu, breadcrumbs: what does what?

Element	Role for the visitor	Role for Google	Common mistake
Site structure	Understand the offer at a glance	Infer the hierarchy of pages	Mirroring the org chart rather than needs
Menu	Reach the main sections	Links present on every page	Mega menu that links everything, everywhere
Breadcrumb trail	Know where you are, go up one level	Links to parent pages, can be displayed in results	Missing or generated only in JavaScript
URL	Read the address, share it	Stable identifier of the page	Changing URLs with every redesign
XML sitemap	None	List of URLs to crawl	Including error or non-canonical pages

The XML sitemap helps discovery, but it does not replace navigation. The W3C accessibility rules point in the same direction on the visitor side. The WCAG 2.2 Multiple Ways criterion requires more than one way to find a page: menu, search, site map. The guide to the XML sitemap explains what it should contain.

How do you find the structure of an existing site?

• Crawl the site: a tool like Screaming Frog rebuilds the levels, the depth of each page and the orphan pages. It offers a tree view.
• Read the XML sitemap: often at /sitemap.xml or declared in robots.txt. It gives the list of URLs, not their hierarchy.
• Follow the menu and breadcrumbs: they show the intended structure, not always the real one.
• Cross-reference with Search Console: pages that get impressions without being linked from the site reveal orphans.

To draw a target site structure, a spreadsheet is enough. Diagram tools such as GlooMaps, FlowMapp or Figma help share it with teams. They bring nothing to SEO in themselves.

Which site structure mistakes should you avoid?

• strategic pages reachable only through internal search or a filter;
• empty or single-page categories, which add a level without adding anything;
• two sections that target the same intent and compete with each other;
• faceted filters that create thousands of crawlable URLs with no search demand;
• a redesign that changes every URL without a redirect plan: see the guide to SEO migration.

A site structure is not set in stone. Review it when a section grows, when a new offer appears, or when data shows that important pages remain too deep. It is part of the technical pillar of SEO, just like crawling and indexing.