Every attribute combination in a catalogue is a potential page, and almost none of them should be one. Deciding which qualify is the single decision that most affects whether a store has a clean, rankable structure or thousands of thin near-duplicates.
The test, in order
1. Is the term searched independently?
Not whether people could search it, but whether they do. “Waterproof hiking boots” is a real query. “Size 9 blue waterproof hiking boots” is a filter someone applies, not something they type.
2. Do the results show category pages?
Search it. If the first page is category and listing pages, a category page can compete. If it is all product pages or guides, a category page is the wrong shape for that query.
3. Do you have enough products?
A category with three products is thin by construction and offers a poor experience. Roughly a screen’s worth is a reasonable floor, though the honest number depends on the category.
4. Can you write something specific about it?
If the only distinguishing text you can produce is the attribute name repeated, there is nothing to rank with. A category needs a reason to exist beyond filtering.
5. Is it stable?
A category whose products rotate out entirely each season becomes empty, which produces a soft 404. Attributes tied to stock rather than to product type are usually filters.
A term passing all five justifies a page. Failing any one of them usually means it should stay a filter.
Category, subcategory or filter
- Category: a primary product type with clear standalone demand.
- Subcategory: a genuine segment within it that people search separately. See category vs subcategory.
- Indexable filter: rare, and only where an attribute has real demand and you will treat the page as a proper category with its own content.
- Filter, not indexed: everything else, which is the large majority.
What happens when you get it wrong
Too many pages
Near-identical categories competing with each other and with their parent. Crawl capacity consumed by pages nobody searches. This is the more common error and the harder to unwind, since consolidating dozens of pages afterwards is slow work.
Too few pages
Real demand going unmet because the segment only exists as a filter. Less common, and easier to correct — you create the page.
Given the asymmetry, the safer default is fewer pages. Adding one later is straightforward; removing fifty is not.
Checking what you already have
- List categories with fewer than a handful of products.
- Find categories whose only content is the product grid.
- Check Search Console for categories with impressions but no clicks, and for categories with neither.
- Look for pairs of categories receiving impressions for the same query.
Frequently asked questions
How many products before a category is justified?
Demand matters more than count, but a category with only two or three products is thin regardless. If demand is real and inventory is thin, the honest answer may be to stock more rather than to publish the page.
Can I make filtered URLs indexable instead?
For a small number with genuine demand, yes, provided you treat them as real categories with their own content. Indexing filters generally is the fastest route to index bloat.
What about brand pages?
Brand plus product type is usually searched and often justifies a page. Brand alone competes with the brand’s own site and is rarely winnable for a retailer.
Should I create categories for seasonal terms?
Only if they persist year to year, and keep them live out of season rather than unpublishing. Creating and deleting seasonal categories destroys whatever they accumulate.
Consulting CTA
If your catalogue has grown more categories than demand justifies, book an SEO consultation for a structural review.