What is site architecture?

A navigation diagram describes relationships and routes through pages. It is more than a list of folder names. Google’s ecommerce guidance explains that links between pages help it understand their relationships and relative importance. That guidance does not establish one ideal layout for every website.

Flat and deep describe a tradeoff

A flatter arrangement offers shorter paths but can overload a menu or listing. A deeper arrangement groups choices into narrower sections but requires more steps. Select the organization that makes the actual tasks understandable.

Navigation design choices, not ranking predictions
ArrangementPotential benefitWhat to inspect
FlatterFewer intervening pages for a taskA crowded menu or an unhelpful mixed listing.
DeeperMore specific categories and choicesImportant pages hidden behind unnecessary steps.

Instead of enforcing a universal click limit, try the journey to a real priority page. Ask whether each intervening page helps the visitor choose the next destination.

Group navigation without isolating useful links

A hypothetical desk store might group products, assembly guides and care instructions into separate navigation sections. A product can still link to its relevant assembly guide. Keeping related tasks connected is more useful than forbidding every link between categories.

Some editors call tightly grouped sections “silos.” Treat that as an organization choice, not proof that isolating pages causes a ranking uplift. This page concerns page and link structure; it does not prescribe a topical-authority campaign.

Folder names do not establish link relationships

A page at /products/desks/frame/ may be one link from the homepage. Another at /frame/ may only be reachable through a long sequence of listings. The URL spelling cannot tell you which route exists.

Google’s guidance says it generally uses links rather than URL structure to infer site structure. Keep naming consistent, but inspect navigation and body links when diagnosing a missing route.

Separate discovery, crawling and indexing

A sitemap can help expose pages that lack normal navigation links; it does not add a path that visitors can follow. Google’s ecommerce guide notes this alternative discovery route. A discovered or crawled page is also not automatically indexed.

Google’s crawl-budget guide focuses on large or rapidly changing websites. It discusses crawl capacity, demand and unnecessary URL inventory. Do not assume that a small site needs a crawl-budget project because its navigation contains several levels.

Menus, category pages and relevant body links should provide usable routes to important pages. On an ecommerce site, relying exclusively on a search box leaves a gap: Google’s guidance says Googlebot generally does not submit searches to discover products.

For paginated listings, Google recommends sequential a-href links and distinct page URLs, with each page using its own canonical. It no longer uses rel="next" and rel="prev" to identify pagination. A load-more control alone should not be treated as proof that later page URLs are discoverable.

Review architecture against a real visitor task

  • Choose the pages that support the actual priority tasks.
  • Follow each path from the public starting page and record unnecessary detours.
  • Check labels, link destinations and the content of intermediate listings.
  • Inspect later pagination pages as well as the first screen of results.
  • Compare a link crawl with the known page inventory to investigate unreachable pages.

Test a proposed navigation change on those journeys before moving URLs. Fixing a missing link may solve the problem without a site-wide migration.

Avoid cosmetic rebuilds and artificial crawl barriers

A neatly drawn diagram does not prove the published menus implement it. A single enormous dropdown does not necessarily make choices easier. Blocking later listing pages solely to satisfy a depth target can remove useful discovery paths.

Google’s Starter Guide discourages restructuring an existing site merely for organization. Base a change on an observed navigation or discovery problem and record the routes it improves.

Questions readers ask

Must every important page be linked in the main menu?

No. The menu is one navigation surface. A useful category page or contextual link may provide a clearer path. Choose links according to the task rather than filling the main menu with every URL.

Does an XML sitemap fix an orphan page for visitors?

No. It can provide a discovery route for crawlers, but it does not create a visible navigation link. Investigate whether the page needs a relevant link from the public site.

Click Depth, Internal Link, XML Sitemap, URL Structure.

Primary references and scope

Documentation checked on October 13, 2026. Worked examples are hypothetical.