Site architecture is the map from a visitor’s question to the next page that helps. Give each topic hub a practical job, give each detail page a distinct purpose, link where the reader’s question changes and hunt for pages that nothing links to. Google says it will likely understand your pages as they are, so restructure to help people, not to chase a theory.

Structure is for the visitor first

Google’s starter guide says that organizing a site logically can help search engines and users understand how pages relate. It also says something less often quoted: do not drop everything to reorganize, because search engines will likely understand your pages as they are today, regardless of how the site is organized.[1] That is a good brake on over-engineering. A tidy structure is worth building because visitors move through it, not because it is a ranking switch.

The common pattern is a hub with detail pages around it. The hub answers the starting question: what are the options and how do I choose? Each detail page does one job in depth. Links connect them in the places where a reader would naturally want the next answer. Pages with no links to them, called orphans, are the failure the pattern prevents. Google finds pages on your site through the links on your pages, and recommends that every page you care about is linked from at least one other page.[2]

Hub pageSports massage Explainer Comparison Prices Booking Injury guide Case note Orphan pageNothing links to it
A hub with six detail pages and one orphan. Solid lines are hub links, and dashed lines are contextual links inside the content. The orphan can be crawled only if something else, such as a sitemap, points to it, and no reader can reach it by browsing.

Give the hub a practical job

A hub that is only a list of titles adds nothing the navigation does not already provide. A useful hub does the choosing work. It explains the situation the reader is in, summarizes each option in a sentence or two with a descriptive link to the detail page, and helps someone decide where to go next. Ask the hub test: could a first-time visitor pick the right next page in thirty seconds?

Hub outline

Opening
Who this is for and the decision it helps with, in two sentences.
Options
One short block per detail page: what it covers, who it suits, a descriptive link.
Decision aid
A small table or a few “if you… then read…” lines.
Next step
The one action the topic leads to, such as booking or requesting a quote.
Ownership
The single hub URL for this topic, so no other page tries to be the overview.

The risk is duplication. If the hub repeats each detail page at length, the two compete for the same queries and a reader gets the same text twice. Keep the hub as the overview and the choice, and leave the depth to the detail pages. The search intent chapter gives the test for whether two pages have different jobs.

Make each detail page distinct, with an address that says so

Each detail page needs one job that no other page on the site has. Name it plainly in the URL. Google recommends descriptive URLs with words that are useful to users, such as /pets/cats.html, and not strings of random identifiers, because parts of the URL can appear in results as breadcrumbs.[1] It also suggests grouping topically similar pages in directories, noting that on sites with more than a few thousand URLs this can help Google learn how often different sections change.[1]

URL habits that help and hurt
DoAvoid
/services/sports-massage/: readable words, one topic per address./p?id=2387: nothing to tell a person or a crawler what is there.
/guides/hamstring-strain/: the folder says what kind of page it is.Folders that mirror the org chart, such as /team-b/q3/.
Standard parameters, such as ?sort=price&page=2, used only where needed.Fragments that change content, such as /#/potatoes. Google generally does not support them; use the History API.[3]

Once a URL is public, treat it as stable. A renamed folder creates dozens of redirects and a window in which links and signals are split. Choose carefully the first time, and use the redirect guidance in the crawl and index chapter when you must change.

Place links where the question changes

Four kinds of internal link do different work. Mixing them up produces the two classic problems: pages nobody can reach, and pages with a hundred links.

Internal link types and their jobs
TypeJobWatch for
NavigationHelp anyone find the main sections from any page.A menu so large that nothing stands out.
ContextualAnswer the next question where it arises, inside the content.Generic anchors, or links stacked without context.[2]
RelatedOffer two or three logical next reads at the end of a page.An automatic list that ignores what the reader just learned.
BreadcrumbsShow where the page sits in the hierarchy and give a way up.Trails that do not match the real structure.

Breadcrumbs also have a structured-data form. Google describes a breadcrumb trail as indicating a page’s position in the site hierarchy, and it can show the trail in results.[4] The starter guide adds that Google learns breadcrumbs from the words in the URL automatically, so markup is a refinement and not a requirement.[1] Write contextual anchors with the guidance from the titles and links chapter. Google’s advice is to use descriptive text, and its list of practices for AI features includes making content findable through internal links.[5]

Find orphaned and overloaded pages

Auditing structure is a comparison exercise. Gather three lists: the URLs in your sitemap, the URLs a crawler reaches by following links from the home page and the URLs that received organic visits in the last three months. The differences are the findings. Most crawlers report inbound internal links for each page, which is the figure you need.

What each pattern in the lists usually means
PatternInterpretationAction
In the sitemap, no inbound internal linksAn orphan. It may be valuable, or a leftover.If useful, link to it from the pages where its topic comes up. If obsolete, redirect or remove it.
Two URLs earning impressions for the same queryOverlapping pages. In Search Console, open the query and check the pages tab.[6]Merge them with a permanent redirect, or give each a clearly different job.[7]
A hub with dozens of links and no guidanceAn index, not a decision aid.Cut to the choices that matter and add the “if you… then…” guidance.
A page that only appears at the end of a long pathImportant but buried.Link it from the hub and from the pages that lead to the decision.

Record each decision in the same sheet: URL, job, inbound internal links, organic visits, overlapping page and action. Do the work in small batches and re-crawl afterwards to confirm that the orphan count fell. Keep the exercise going as the library grows, because every new page is a new chance to create an orphan.

Apply it to your website

  • Each topic has one hub URL and each detail page has one distinct job.
  • The hub helps a visitor choose, with a short summary and a descriptive link for every option.
  • URLs are readable, stable and grouped by topic, with no content-changing fragments.
  • Every page you care about is linked from at least one other page, in context and with descriptive anchor text.
  • Breadcrumbs, where used, match the real hierarchy.
  • Overlapping pages are merged with permanent redirects or given clearly different jobs.
  • You compare the sitemap, a crawl and analytics regularly to find orphans and buried pages.

Sources and further reading

Sources were consulted in October 2026 and the numbered notes above point here. Platform documentation changes often, so check the linked page before you act on a platform-specific detail. Worked scenarios are illustrative unless identified otherwise; methods and worksheets are Vantle’s editorial guidance.

  1. Google Search Central: SEO Starter Guide, “Organize your site”Logical organization, descriptive URLs, directories for large sites, and the advice not to restructure needlessly.
  2. Google Search Central: Link best practicesHow Google finds pages through links, anchor text, and linking every page you care about.
  3. Google Search Central: URL structure best practicesCrawlable URL requirements, including the limits on fragments and the use of the History API.
  4. Google Search Central: Breadcrumb structured dataHow a breadcrumb trail expresses a page’s position in the site hierarchy.
  5. Google Search Central: AI features and your websiteBest practices include making content findable through internal links.
  6. Search Console Help: Performance report (Search results)Query and page dimensions for spotting competing URLs.
  7. Google Search Central: Site moves with URL changesPermanent redirects from old URLs to their replacements.