serkankaya.
Türkçe sürüm
Let’s work together ↗

Ecommerce Site Architecture: Category Tree and URL Structure

How to build ecommerce site architecture: plan the category tree, URL structure, internal links, breadcrumbs and filter pages step by step for your store.

Ecommerce site architecture is the structure that defines the hierarchy of an online store’s home, category, subcategory and product pages, and how those pages link to each other. The short answer: you build the category tree around the language customers search with, you make every product reachable through HTML links that start from the menu, and you keep URLs simple, stable and readable. Among filters, only the ones with search demand become separate pages.

I covered the overall framework of store SEO in my ecommerce SEO guide. This post focuses on the structure underneath that framework: the category tree, URL structure, internal links, breadcrumbs and filter pages. I started as a software developer before moving into SEO, so I look at this topic from both the development side and the search side.

Why is ecommerce site architecture important?

Ecommerce site architecture matters because Google largely infers the importance of a store’s pages from internal links. According to Google’s ecommerce documentation, the number of links needed to reach a page and the number of links pointing to it can be used to understand that page’s relative importance on the site.[1] In other words, a category two clicks from the menu and a product stuck on the sixth pagination page do not carry the same signal.

A corporate site has a few dozen pages, and its structure stays tidy almost by itself. A store, on the other hand, has thousands of products, hundreds of categories and many URLs generated by filters. Architecture decisions here affect templates, not individual pages. A linking mistake in the category template makes hundreds of pages harder to discover at once.

The second job of architecture is user experience. Visitors should find the product they want using words they recognize in the menu and category names. The structure that is right for search engines and the one that is right for customers are usually the same, because both come from the same search language.

Home page Category Subcategory Product
The basic ecommerce hierarchy: each level links to the one below with an HTML link, and products link back up through breadcrumbs.

How do you build a category tree?

A category tree is built by grouping products into main category, subcategory and product levels according to the groups people use when they search. The category tree is the skeleton of the architecture: the menu, URLs and internal links are all derived from it. That is why I start designing the tree from keyword data, not from the product list.

The first step is to map demand for the product types the store sells. Keyword research shows which groups are searched separately and which are searched under a broader heading. For example, “hair care products” is a general demand, while “sulfate-free shampoo” is a subgroup searched on its own. A group with separate, clear demand deserves its own category page.

The second step is deciding how many levels to use. For most stores I find two levels, main category and subcategory, to be enough. Very wide catalogs may need a third level. Rather than adding levels, building more clearly separated categories at the same level usually works better.

The third step is making sure each category matches a single search intent. Two categories that answer the same demand produce two pages competing with each other. I explained how I build the content of a category page separately in my category page SEO post.

Tree built on warehouse logic

  • Groups based on supplier or brand lists
  • Categories named with internal codes
  • Several categories answering the same demand
  • Many empty subcategories with no demand

Tree built on search language

  • Groups based on the product types customers search for
  • Categories named with search phrases
  • Each category matches a single search intent
  • A subcategory is opened only when separate demand exists
A category tree is built around how customers search for a product, not where the product sits.

How do you balance click depth and a flat structure?

Click depth shows how many links must be followed from the home page to reach a page, and it should stay low for revenue-driving pages. The common industry claim that “every product should be at most three clicks away” is not a rule set by Google. What Google recommends is linking from the menu to category pages, from there to subcategories, and finally to all product pages.[1]

A completely flat structure is not the answer either. Linking to hundreds of categories from the home page weakens the hierarchy signal and makes the menu unusable. A balanced structure keeps the hierarchy while pushing important pages forward with extra links. Google also suggests linking to a best-selling product from the home page or from content such as blog posts.[1]

What usually increases depth the most is pagination. When the products in a category are split into pages of 20, the products at the end of the list can end up very deep. In that case, opening a subcategory, increasing the number of products per page or adding related product links reduces the depth.

What should an ecommerce URL structure look like?

An ecommerce URL structure should consist of readable words, separate words with hyphens, contain no unnecessary parameters and stay stable even when the product changes. Google recommends using readable words instead of long ID numbers in URLs, preferring hyphens over underscores as word separators and trimming parameters that do not change the content.[2]

A common misconception needs correcting here. The folder structure in a URL is not the main way to tell Google about site hierarchy. Google’s ecommerce documentation states clearly that it generally does not look at URL structure to understand a site’s structure, and analyzes the links between pages instead.[1] So adding a deep category path to a URL does not improve rankings on its own.

This makes the product URL decision easier. A product can belong to more than one category. If the product URL includes the category path, the same product opens from several addresses and a duplicate URL problem appears. For products, I prefer a short, stable path that is independent of the category:

Category:      /hair-care/
Subcategory:   /hair-care/shampoo/
Product:       /product/sulfate-free-shampoo-250-ml/
Filter:        /hair-care/shampoo/?color=blue&sort=price

You also need a decision on Turkish characters. Google recommends percent encoding characters outside the ASCII range in links.[2] In practice, most stores in Turkey replace the Turkish letters “s with cedilla”, “soft g” and “dotless i” with plain “s”, “g” and “i”. Whichever path you choose, the rule must be the same across the whole site. Because Google treats URLs as case sensitive, writing all addresses in lowercase also prevents duplicate URLs.[2]

Internal links are what actually carry ecommerce architecture; Google largely infers a store’s structure from menu, category and in-text links. For Google to follow these links, they need to be provided with the <a href> tag. Google recommends not using JavaScript events attached to other HTML elements for navigation.[3]

I set up four kinds of internal links in a store. The first is menu links, which carry the main categories and subcategories. The second is the product list and subcategory links on category pages. The third is the related product and “same category” links on product pages. The fourth is in-text links from guides and blog content to categories.

Google notes that if category pages do not link directly to every product in a category, Googlebot may not find all products by crawling alone. Googlebot also does not type queries into the site search box.[1] That is why products reachable only through site search carry an orphan page risk from an architecture point of view. If linking to every product is not possible, Google recommends using a sitemap or a Merchant Center feed.[1]

A breadcrumb is the visible form of the architecture on the page. A trail such as “Home › Hair care › Shampoo” on a product page shows users where they are and links to parent categories. BreadcrumbList structured data passes this trail to Google in a machine-readable form; Google uses this markup to categorize the page’s information in search results.[5]

How are filter pages included in the architecture?

Short answer

How are filter pages included in the site architecture?

Filter pages are included in the architecture by selecting only the combinations with real search demand. These combinations are treated like a subcategory, with a clean URL and their own title and description, and are linked from the category page. Crawling of color, price and sort combinations that see no demand is limited.

  1. Check search demand for filter combinations with keyword data.
  2. Give the combinations with demand a clean URL, title and description.
  3. Link these pages from the category page with normal HTML links.
  4. Limit crawling of filter and sort URLs that have no demand.
  5. Add only the filter pages you want indexed to the sitemap.

Faceted navigation is the component that breaks ecommerce architecture most easily. Every combination of color, size, brand, price range and sort options can generate a new URL. Google notes that crawling of filter URLs that do not need to appear in search results can be blocked with robots.txt, or that filter information can be carried in the URL fragment.[4]

The key here is not to think of filters as “all open” or “all closed”. A combination searched on its own, such as “women’s running shoes”, deserves a subcategory spot in the architecture. A combination such as “blue, size 42, sorted by price ascending” is only a tool that narrows the list at that moment, and it does not need to be crawled.

How do you change the architecture of an existing store?

The architecture of an existing store is changed gradually, with a 301 redirect map that pairs old and new URLs. Rebuilding the category tree or the URL structure is effectively a site migration. Every old address left without a redirect risks losing the links and rankings it has built up.

In the projects I have worked on from the software side, I always start this job with a spreadsheet. It lists each old URL, its new destination, its status code and the organic traffic it received recently. Addresses that bring traffic are mapped first. Old categories with no direct equivalent are redirected to the closest new category; there is no bulk redirect to the home page.

The work does not end when the change goes live. Menu and in-text links should point directly to the new addresses, not be left to redirect chains. The sitemap should contain only new URLs, and canonical tags should be updated to the new addresses too. In the weeks after a migration, I closely monitor the page indexing report and the 404 list in Search Console. I laid out the audit order in detail in my ecommerce SEO checklist post.

Ecommerce site architecture check0/8

Conclusion: architecture is built before the first product is uploaded

Ecommerce site architecture consists of a category tree built around customers’ search language, the internal links that carry that tree, and simple, stable URLs. Since Google reads structure from links, the most important job is making every product reachable through HTML links that start from the menu, in few steps and at a single address. Among filters, only those with demand become part of the architecture.

As a next step, I recommend crawling your store with a site crawler and comparing the click depth report with the non-indexed pages in Search Console. You can see how I handle architecture together with ads and tracking on my performance marketing page, my technical SEO approach on my SEO services page, and my background on my about page.

Key terms

Site architecture
Site architecture is the structure that defines the hierarchy of pages on a website and how they connect through menus, categories, breadcrumbs and in-text links. It affects how search engines discover pages and understand their relative importance.
Category tree
Category tree is the hierarchical classification that groups an online store's products into main category, subcategory and product levels. It is also called a taxonomy. The menu, URL and internal link structure are usually derived from this tree.
Click depth
Click depth is the measure of how many links must be followed from the home page to reach a page; it is also called crawl depth. Site crawlers report this value for every URL, which makes it easier to find important pages that sit too deep.
Orphan page
Orphan page is a page that is live on the site but receives no link from any other internal page. Google can only find such a page through the sitemap or external links, and may treat it as less important.

Frequently asked questions

Why is ecommerce site architecture important?

Site architecture determines how Google finds products and which pages it considers important. Google explains that it can use the number of links needed to reach a page and the number of internal links pointing to it as signals of relative importance. The same structure also decides how many steps a visitor needs to reach a product.

Should product URLs include the category path?

A category path in the product URL is not required. Google mostly infers site structure from the links between pages, not from URLs. For products listed in more than one category, a stable product URL that is independent of the category reduces duplicate URL problems.

Is the 3-click rule a real Google rule?

No, the 3-click rule is not a rule set by Google; it is a common industry target. What Google says is that important pages should be reachable through menu and category links, and that important products should get more internal links. In large catalogs the goal is to keep revenue-driving pages shallow rather than to hit a fixed number.

Should filter pages be part of the site architecture?

Only filter combinations with real search demand should be part of the architecture. These pages are treated like a subcategory, with their own clean URL, title and description. Crawling of color, price and sort combinations that see no demand is limited with robots.txt or a similar method.

How do you change the architecture of an existing store?

First, list all current URLs and the pages that bring traffic. Then prepare a 301 redirect map that shows the new destination for every old URL. After the change goes live, internal links, the sitemap and canonical tags are updated to the new addresses and Search Console is monitored regularly.

Sources

  1. Google Search Central: Help Google understand your ecommerce website structure
  2. Google Search Central: URL structure best practices for Google Search
  3. Google Search Central: Make your links crawlable
  4. Google Search Central: Managing crawling of faceted navigation URLs
  5. Google Search Central: Breadcrumb structured data