A category page is the list of products an online shop groups under one label: walking boots, fridge freezers, sofa beds. It is the page that captures broad demand, from people who know what they want to buy but not yet which model.
On 26 September 2026 we checked seven commercial UK SERPs to see which pages Google actually puts on page one for those queries. Of 61 organic results, 41 were category pages and 1 was a product page.
Seven queries, 61 results: 67% of page one is a category page
We picked one query for each of the largest ecommerce sectors, phrased as a purchase and without a brand name. Then we classified by hand the page type of every organic URL on page one.
| Query | Organic results | Category pages | Product pages | Editorial content | Other |
|---|---|---|---|---|---|
| fridge freezer | 9 | 9 | 0 | 0 | 0 |
| sofa bed | 9 | 8 | 0 | 0 | 1 |
| walking boots | 9 | 8 | 0 | 1 | 0 |
| brake pads | 9 | 6 | 0 | 2 | 1 |
| wireless headphones | 9 | 5 | 0 | 3 | 1 |
| face sunscreen | 8 | 5 | 0 | 3 | 0 |
| extra virgin olive oil | 8 | 0 | 1 | 6 | 1 |
| Total | 61 | 41 (67%) | 1 (2%) | 15 (25%) | 4 (7%) |
The ratio between the two middle columns is the figure that matters. Forty-one category pages against one product page. On "fridge freezer" and "sofa bed" there are no product pages at all, and page one belongs to Currys, Argos, AO, John Lewis, Beko, Dunelm, IKEA, Habitat, Sofology and Furniture Village. On "wireless headphones" the lists from Argos, Very, PriceRunner and idealo lead, and Amazon ranks with an internal search page.
The exception teaches more than the rule. On "extra virgin olive oil" category pages drop to 0 out of 8: what ranks is two Reddit threads, Quora, three other forums, a Budgens recipe tag page and a single Amazon product page. It is a query that means both "what is it" and "buy some", and Google splits it. Before deciding that a category should rank for a keyword, look at the search intent exactly as the SERP shows it, not as we imagine it.
A second finding from the same check: six of the seven product SERPs showed an AI Overview, and only "brake pads" did not. The sources are telling, though. On "walking boots" the AI Overview cited five category pages (Scarpa, LOWA, Mountain Warehouse, Go Outdoors, Grisport); on "fridge freezer" it linked to the Currys and Argos category pages and cited those of Appliances Direct and Marks Electrical, next to Google Shopping product listings.
Why the category carries more weight than the product page
Search volume sits on the category name, not on the model code. "Sofa bed" is searched around 246,000 times a month in the UK (DataForSEO, Google UK, September 2026). "Tornsborg 2-seat sofa-bed dark grey" is not. Anyone doing keyword research for a catalogue sees it straight away: most searches concentrate on a minority of labels, and those labels correspond to listing pages.
The same mechanism works outside ecommerce. On a project we have run since 2020, Macropix, an LED screen manufacturer based in Milan, the heaviest keyword in the Italian market is "ledwall", 4,400 searches a month in Italy: a category word, not a model. The page that targets it is now first, "monitor pubblicitario" (advertising display) went from position 88 to position 2, and across the cluster the domain holds a share of voice of 25.55%, ahead of amazon.it at 18.08% (Semrush, July 2026). The full figures are in the guide to SEO for manufacturing and in the Macropix case study.
The market context shows what getting these pages wrong costs. According to the Office for National Statistics (Retail sales, Great Britain: August 2026, published 18 September 2026), 28.8% of all retail sales in Great Britain were made online in August 2026, up from 28.4% in July, and online sales values rose by 2.5% over the month. More than one pound in four goes through a website, and most shoppers reach it through a list of products.
On-page elements, in order of impact
This is the order we use when we rewrite a category tree, and it is not random: first what declares what the page is about, then what makes it understandable to the reader.
- Title tag: the category keyword at the start, the brand at the end. A title such as "Products" repeated across twenty categories is the quickest way to rank none of them. The rules on length and structure are in the guide to the meta title.
- H1: just one, the category name as the market searches for it. If the catalogue calls it "outdoor footwear" and people search for "walking boots", people win.
- URL: readable and stable. Google says explicitly that "Google generally doesn't look at the structure of URLs to work out the structure of a site": the URL serves people and internal consistency, while links communicate the architecture.
- Text: one part above the grid and one below, not a single block at the bottom.
- Breadcrumb: visible and marked up. It is the only element on this list that also has an effect in the SERP, and we cover it in the guide to breadcrumbs.
Category page text: where it goes and how much you need
No Google document gives a length for the text on a listing page. The 150-300 word range that circulates in US-written guides is industry practice, not a rule: it comes from guides such as those by Passionfruit SEO and Semrush, and should be treated as such.
The useful criterion is different: the text should answer the two or three questions a person asks before choosing within that category. For "fridge freezer" they are the 50/50 or 70/30 split, frost-free or not, and freestanding or integrated. These are the same points Google puts in the AI Overview and in the "People also ask" box above the results. Writing them above the grid, in four lines, works better than eight hundred words of filler under the footer, and the reverse of what you often hear also holds: if the category has nothing to say, leave it without text rather than add a generic paragraph. We have a dedicated guide on how to build that content, on creating SEO content.
There is a measured benchmark for the average quality of these pages. The Baymard Institute, which assesses more than 21,000 parameters across over 170 leading US and European ecommerce sites, rates 58% of product lists on desktop and 78% on mobile as "poor" or "mediocre" (benchmark updated September 2025). Baymard also finds that 61% of sites do not promote filters within the product list. The sample is US and European, and the category pages we open every week give us no reason to expect better numbers elsewhere.
Filters: which ones deserve an indexable URL
The question is not whether to index filters. It is which combinations have their own search demand, large enough to justify a page, and how to stop Google discovering all the others. These are two separate jobs.
We found proof that a filter can rank well inside the check itself: on "wireless headphones", in eighth organic position, there is a Very filter page, URL /browse/technology-headphones/headphone-features--wireless/headphone-style--over-ear-headphones. It is not a menu category, it is two attributes promoted to a page.
Google's official documentation on managing faceted navigation lists two concrete problems: overcrawling of filtered URLs, and slower discovery of new pages, because crawl resources are spent on useless combinations. The tools it recommends, in order of effectiveness: a robots.txt disallow on filtered navigation URLs; URL fragments after the hash, which Google does not crawl; the canonical, which reduces crawl volume over time but does not eliminate it; and rel="nofollow", less effective and only useful if applied to every single link pointing to that URL. Plus a rule almost nobody follows: when a filter combination returns no products, the page must return a 404 status code, not a redirect to a generic error page.
| Filter type | Example | Does it have its own search demand? | What to do with it |
|---|---|---|---|
| Attribute with measurable volume | 3 seater sofa bed | Yes, checkable with a tool (8,100 searches a month in the UK) | Indexable page, its own title and H1, link from the parent category |
| Recurring use-case attribute | wireless headphones for TV | Yes, seasonal or recurring (1,300 a month in the UK) | Indexable page, text that explains the use case |
| Aesthetic or size attribute | beige sofa bed 180 cm | Almost none | Navigation filter, URL not crawlable |
| Multiple combinations | colour + price + availability | None | robots.txt disallow |
| Sorting and views | ?sort=price, ?view=grid | None | Disallow, or a fragment after the hash |
| Combination with no results | size 14 in a range that stops at size 12 | Not relevant | HTTP 404, never a redirect |
Skipping this decision carries a double risk: duplicate content across dozens of near-identical variants, and crawl budget burnt on URLs that will never bring a visit.
The category passes on authority, it does not just receive it
Google describes the hierarchy it expects in one practical sentence: "add links from menus to category pages, from category pages to sub-category pages, and finally from sub-category pages to all products". And it warns what happens when that chain breaks: "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". In that case you need sitemaps or a Merchant Center feed.
The same documentation page, Help Google understand your ecommerce site structure, gives the criterion Google uses to weigh a page's relative importance: "The more links a page has to it within a site, the higher the relative importance of the page". In practice, a category linked only from the menu counts for less than one that also receives links from product pages, sibling categories and blog articles. This is the work described in the guide to internal links, and on large catalogues it matters more than half of the on-page work.
One last technical constraint from the same source: navigation links must be <a href> tags, not JavaScript events on DOM elements. A dropdown menu that changes page via script is a menu Googlebot does not follow.
Pagination: the first-page canonical mistake
Two official guidelines you rarely see written out in full. The first: the rel="next" and rel="prev" tags are no longer needed, "Google no longer uses these tags, although these links may still be used by other search engines". The second, from Google's page on pagination and incremental page loading: "Don't use the first page of a paginated sequence as the canonical page". Pointing the canonical of pages 2, 3 and 4 to page 1 is the default setting of many plugins, and it removes from the index the products that only appear on the later pages.
The other two rules on the same page: each page in the sequence has a unique URL with a parameter such as ?page=2, and page numbers must not go after the hash. For "load more" and infinite scroll you still need a URL reachable through a link, otherwise the sitemap is the only way left to discover them. Paginated URLs do get indexed and can rank: in our check, Argos ranks first among the organic results for "wireless headphones" with page 3 of its category.
Structured data: less than you have been told
Two wrong tips circulate about category pages: mark up the grid with Product, and add ItemList to get a carousel. The documentation says otherwise.
Google's page on structured data for ecommerce lists six types: BreadcrumbList for the hierarchy, Organization for company details, LocalBusiness for physical shops, Product and ProductGroup for products and variants, Review for reviews, VideoObject for videos. ItemList is not among the types recommended for listing pages. And the carousel rich result built with ItemList is supported for four content categories: lists of courses, movies, recipes and restaurants. Not products.
So on a category page the markup you need is BreadcrumbList, while Product belongs on product pages. The rest is work that produces no visible result in the SERP, and the guide to schema markup goes into detail on the required fields.
The mistakes we find most often in audits
- Categories built around warehouse logic instead of how the market searches: "miscellaneous accessories", "professional range", "new in".
- A category and a subcategory targeting the same search demand and taking positions from each other. The fix is merging, not adding text.
- The canonical of every filtered page pointed at the parent category, including the filters that had their own demand and could have ranked.
- The same descriptive text on twenty categories with only the name changed.
- Categories with no products in stock left online and indexed.
- No links between sibling categories, when that is the connection that supports both navigation and the internal distribution of authority.
How to tell whether a category is performing
The Performance report in Google Search Console, filtered for a single category URL, shows the queries that page appears for. The pattern to act on first is stable impressions with falling clicks: the page is seen and not clicked, and nine times out of ten the problem is the title, or the fact that the answer is given above the results. The second useful pattern is queries in positions 8 to 20: these are the categories one step away from page one, and where the work pays off soonest.
For a category's keyword cluster it is also worth tracking share of voice, meaning how much of the total visibility available on those queries the domain takes. It is the metric that tells you whether you are growing or the market around you is. The full set of metrics for a catalogue is in the guide to ecommerce SEO.
If you would rather someone else did this work, our SEO services always start from the category tree, before the product pages, for the reason shown in the first table of this article.
Frequently asked questions
No Google document gives a minimum or maximum length. The 150-300 word range found in US-written guides is industry practice, not a rule. The useful criterion is to cover the two or three questions a person asks before choosing within that category, with part of it above the product grid. If the category has nothing specific to say, it is better to leave it without text than to fill it with a generic paragraph repeated across twenty pages.
Only those with their own measurable search demand, for example 3 seater sofa bed or wireless headphones for TV. Those become pages with a title, an H1 and a link from the parent category. All other combinations, colours, prices, sorting and multiple intersections, stay as navigation filters and should be kept out of crawling with a robots.txt disallow, as Google's documentation on faceted navigation recommends. When a combination returns no products, the page must return a 404 status code, not a redirect.
BreadcrumbList. Google's official page on structured data for ecommerce recommends BreadcrumbList, Organization, LocalBusiness, Product and ProductGroup, Review and VideoObject, and does not include ItemList among the types for listing pages. The ItemList-based carousel rich result is only supported for lists of courses, movies, recipes and restaurants, so on a product grid it produces no visible result in the SERP. Product markup belongs on product pages, not on the category.
It depends on how many labels have their own search demand. A subcategory makes sense when it captures a query the parent category does not cover and when it has enough products not to look empty. If two levels target the same query they will take positions from each other: in that case merge them and keep a single page.
Yes. Text is not what gets a category page ranked: what holds it up is the combination of title, H1, a consistent URL, the number of products available and the internal links it receives. Text helps cover the secondary questions and gives AI-generated answers something to cite, but it is not the condition for ranking.
Yes, they are the natural landing page for campaigns on broad product queries, because they offer choice rather than a single model. Landing page experience is one of the components of Quality Score, so a tidy, fast category with visible filters also lowers the cost per click on the paid side.
One thing the industry rarely says. In our 26 September check, six of the seven product queries showed an AI Overview, but on "walking boots" and "fridge freezer" the sources it pointed to were category pages and Google Shopping listings, not blog posts. On a query like fridge freezer Google does not stop at a summary: it sends people to a page where they can compare and buy, because that is what they are trying to do. As long as that holds, the category page is the page on a site that AI answers are most likely to send people to rather than replace. It is worth putting it back at the centre of the work, before the product pages and before the blog.
