Question

E-commerce Site Architecture: Building for Scale

4 min read

Short answer

Store architecture that scales rests on links more than on URLs. Google's ecommerce structure guidance says "Google generally doesn't look at the structure of URLs to work out the structure of a site. Instead, it analyzes the linkages between pages". So the work is: a category hierarchy that links from menus to categories, to subcategories, to every product; one stable, clean URL per page; filter and sort URLs kept out of crawling; and a plan for how all of that holds when the catalog grows tenfold.

Sources

  1. 1developers.google.com
  2. 2developers.google.com

Ask next

What Google Uses to Understand a Store

From the same guidance:

  • Links show importance. Google can use “the number of links it needs to follow to reach a page and the number of links to a page to infer the relative importance of a page”.
  • Every product needs a link path. “If category pages don’t include direct links to all products in a category, Googlebot might not find all of your products by crawling alone”, and Googlebot “generally doesn’t try to submit searches into a search box”.
  • Links must be real links: use <a href> tags, not JavaScript events on other elements.
  • Promote what matters: link best sellers from the home page and other content to signal their importance.

Where full linking isn’t possible, a sitemap or Merchant Center feed helps Google find the rest.

Category Hierarchy

  • Mirror how shoppers choose: department, type, then the attribute that splits large groups (for example, men’s, shoes, running).
  • Link down every level: menu to categories, category to subcategories, subcategory to all its products, with pagination where lists are long.
  • Keep important pages close to the home page in clicks, through the menu, category links and featured links.
  • Split categories when they become hard to browse, not at a fixed product count, and merge categories too small to help.
  • Use merchandising collections (“Summer Essentials”) alongside the attribute hierarchy, as linked pages with their own purpose rather than near-copies of a category.

Category page content is covered in category page optimization.

URL Structure

Google’s ecommerce URL guidance is about avoiding crawl and duplication problems:

  • One URL per piece of content: minimize alternative URLs that return the same page; “Google cannot determine” that /product/black-t-shirt and /product?sku=1234 are the same “by looking at the URL alone.”
  • Consistent case if your server treats upper and lower case alike.
  • No fragments for different content: “Google does not use fragment identifiers in indexing.”
  • Descriptive words in paths: “The words in URLs may help Google better understand the page.”
  • Standard parameters: ?key=value, no repeated parameters.
  • A unique URL for each paginated page.

Google adds that on an ecommerce platform “you can most likely skip this section”, since platforms handle much of it. Hierarchical paths (/mens/shoes/oxford/) and flat ones (/products/oxford/) both work; choose one and keep it stable, because changing URL patterns later means redirects for every page.

Filters, Sorting and Variants

  • Filters and sort orders generate URLs fast. For those that don’t need to rank, Google’s first suggestion is robots.txt disallow; a noindex doesn’t save crawling because the page still has to be fetched. See faceted navigation SEO.
  • Variants (color, size) either share one product URL with options on the page, or get their own URLs where shoppers search for them; in the second case, keep each variant indexable with its own canonical and group them with variant markup, rather than combining noindex with a canonical.
  • Tracking parameters shouldn’t appear in internal links.
  • Main menu: top-level categories and the pages that matter most, as crawlable HTML links.
  • Mega menus: fine if the links are in the HTML; check the rendered page.
  • Footer: policies, help, key categories; not a list of everything.
  • Breadcrumbs: on category and product pages, with BreadcrumbList markup; Google shows breadcrumbs in desktop results. See BreadcrumbList schema.
  • Related products: crawlable links chosen for shoppers.

Internal linking principles are in internal linking strategy.

Planning for Growth

Document the rules before the catalog outgrows them:

Decision Write down
URL patterns Category, product, filter, pagination formats
Hierarchy Levels, naming, when to split or merge
Filters Which are blocked, which are landing pages
Variants One URL or one per variant, and the markup
Stock changes What happens to out-of-stock and discontinued pages

Then test the rules against expected growth: more categories, more filters, new markets. Discontinued product handling is in out-of-stock and discontinued products, and international structure in international SEO fundamentals.

Monitoring a Growing Catalog

  • Crawl stats in Search Console: the share of requests going to filter, sort or parameter URLs.
  • Page indexing report: products not indexed, duplicates, unexpected exclusions.
  • Crawls of the site to find orphan products, deep pages and redirect chains; see crawl analysis tools.
  • Server logs for what Googlebot actually requests; see log file analysis.

Platform-specific constraints are covered in e-commerce technical SEO.

Two decisions are costly to change later, the URL pattern and the variant model, because changing either means redirects across the catalog; settle both before the catalog grows.

Link profile