Tell category pages and product pages apart first
Every check below needs the type of each page. A category page lists products. A product page sells one product, with its variants. Words in the address, structured data and the layout a page shares with others all help tell them apart.
Check the split before you act on it. Category pages often carry Product markup for their product cards, so markup alone calls them products.
PerfBee gives each page one of eight types after a crawl, with fixed rules and no AI. When a group of pages is wrong, correct it once in Content Intelligence, and the choice holds after every crawl.
Four checks for category pages
- Category pages that list no products. An empty category gives a visitor nothing to pick, and Google may treat an empty list as a soft 404. If it is empty for good, let its address answer 404 or redirect it to the nearest category with products. If it is empty for now, keep it out of the sitemap and the menus until it fills again.
- Category pages that Google is told not to index. A noindex, or a canonical that names another page, leaves the category out of the results or gives its credit to the other page. Remove the noindex unless you closed the category on purpose, and point the canonical at the page's own address.
- Category pages many clicks from the start page. Every product sits at least as deep as its category. Keep categories within 3 clicks of the start page, through the main menu, the start page or the parent category. Google sets no click limit: 3 is a working rule for a category.
- Category pages missing from the sitemap. A category that may be indexed but is not in the sitemap is found only through links. Add it, or look for your shop's sitemap setting for categories.
Check that a category links every product
A product that no category or brand page links to may still be in the sitemap. A visitor who browses never reaches it, though, and Google finds it only by weaker routes.
Put each product in a category, and check that the category lists it on its first page or on a page that the first page links to. When a script builds the list, put plain links to the products in the HTML, so that Google and AI crawlers can read them.
Keep your own site search out of Google
Every search a visitor types can become a page in Google's index. These pages are thin, look alike and use up crawl budget, the visits Google spends on your site. Google advises against letting them be indexed.
Add a noindex to the search result pages, or disallow their address in robots.txt. Do not do both, because Google cannot read the noindex of a page that robots.txt blocks.
Give each page of a list its own canonical
The second and later pages of a category often point their canonical at the first page. Google reads a canonical as a hint to drop the page, so products linked only from later pages may never be reached through the list.
Set the canonical of each page of the list to its own address. Keep the links to the next and previous pages as plain links in the HTML.
Run the seven checks in Fix Planner
The seven checks run after a crawl and show up as fixes in Fix Planner. Choose a page type to see only its fixes.
The noindex check needs a detailed crawl, the depth check a complete crawl, and the sitemap check a crawl that read the sitemap.
Each check that finds pages becomes one fix, with the pages and how to fix them. Page types and the seven checks cost no credits. The Fix Planner help article has every step.