usdigitalmarketingindex.com SEO
How should ecommerce parent and child categories be arranged for search visibility?
Updated November 15, 2026
Arrange categories so that each level answers a different breadth of search, and so that every category and product can be reached by following links. A parent category covers a product type as a whole. A child category covers a subtype that shoppers look for by name. Attributes that merely narrow a list, such as size or color, are filters, and should not each become a category. Then connect the levels with ordinary links from menus to parents, parents to children, and children to products, which is the path Google describes for finding the pages of a store.
Parent/category role definitions
Give each level a job.
| Level | Covers | Serves a search like | Links to |
|---|---|---|---|
| Parent category | A whole product type | running shoes | Its child categories, and featured products |
| Child category | A named subtype with its own range | trail running shoes | Its products, and back to the parent |
| Product page | One item | a specific model name | Its category, related items |
A parent should help a shopper choose a direction: show the subtypes, explain how they differ, and link to each. A child should present its range. When a parent is only a longer list of the same products as its children, the hierarchy is adding a click and nothing else.
Facet versus child category decision
The recurring question is whether a way of narrowing a list deserves its own category page or should remain a filter. Decide with evidence.
| Condition | Evidence to check | Outcome |
|---|---|---|
| Shoppers search for the subtype by name | Keyword data and your site search log show the phrase | Child category |
| The subtype has a meaningful range of products | Inventory count that will stay above a handful | Child category |
| The subtype needs its own explanation or buying advice | You can write content that differs from the parent | Child category |
| The attribute only narrows a list | Size, color, price band, in stock | Filter |
| The combination is one of many permutations | Brand plus color plus size | Filter |
| The subtype is searched by name and is one product deep | One or two items | Link to the products from the parent; no category |
Keep the two mechanisms distinct. A child category is a page you want found in search. A filter is a convenience for someone already on the page. Google’s guidance on pagination and incremental loading addresses filtered and sorted variations of a list separately, suggesting they be kept out of the index or blocked from crawling (Google’s guidance on pagination and incremental page loading). A hierarchy that turns every filter into an indexable URL multiplies near-identical lists.
Navigation and crawlability
Google’s guidance for ecommerce site structure is explicit about the link path: “add links from menus to category pages, from category pages to sub-category pages, and finally from sub-category pages to all product pages” (Google’s guidance on ecommerce site structure). Three consequences follow.
- Products reachable only through the site’s search box may not be found. The same page states that “Googlebot generally doesn’t try to submit searches into a search box as part of crawling a site”.
- Every product should be linked from at least one category. The guidance calls it “strongly recommended to link to all products that you wish indexed”, and names a sitemap or a Merchant Center feed as the fallback where that is not possible.
- Links should be real links. The guidance says to use anchor elements with an href and not JavaScript events on other elements for navigation.
Depth matters as well. The same page says Google can use the number of links it must follow to reach a page, and the number of links pointing to it, to infer the page’s relative importance. By that reasoning, a category buried five levels down, with few links pointing to it, can be read as a minor page.
Category hierarchy examples
A sample tree for a fictional outdoor retailer, with the kind of search each page is meant to answer:
/footwear/ "outdoor footwear"
├── /footwear/running-shoes/ "running shoes"
│ ├── /footwear/running-shoes/trail/ "trail running shoes"
│ └── /footwear/running-shoes/road/ "road running shoes"
├── /footwear/hiking-boots/ "hiking boots"
│ ├── /footwear/hiking-boots/waterproof/ "waterproof hiking boots"
│ └── /footwear/hiking-boots/lightweight/ "lightweight hiking boots"
└── /footwear/sandals/ "hiking sandals"
Filters on every list (not categories): size, width, color, price, brand
Each page in the tree targets one query family, and no two pages target the same one. Waterproof appears as a child category under hiking boots because the fictional shop treats it as a named subtype with its own range. Under running shoes, where few waterproof models exist, the same attribute would stay a filter. The decision is per branch, driven by the conditions in the matrix.
Two failure patterns are worth checking for in an existing store.
- Duplicate branches: the same child reachable as /boots/waterproof/ and /waterproof/boots/ with the same products. Choose one path and link only to it.
- Orphaned children: categories that exist as URLs and appear in no menu and on no parent page.
How long category lists should be split across pages, and how sort orders create duplicate URLs, are implementation topics with their own guides: search-friendly pagination for a category with thousands of products and how sorting parameters produce duplicate category URLs.
Build in this order: define the query family for each level, apply the facet test to every proposed child, wire the links from menu to product, then check that each page has one path.
Add your comment to this answer