Technical work on a shop is not the same job as on a brochure site, because a catalog generates pages by itself. Filters multiply into millions of combinations, every color and size can become its own address, and products go out of stock and come back. Those three mechanisms account for most of the technical damage in ecommerce, and all three are decided by configuration rather than by content.
Faceted navigation multiplies into a crawl problem
Filters for size, color, price, brand and rating combine into an effectively unlimited number of addresses, most of them near-duplicates with nothing to offer a searcher. Left open, crawlers spend their budget on combinations while real product pages go unvisited. The fix is deciding which facets may be indexed because people search for them, and blocking the rest through configuration rather than hoping.
Variants need one page unless people search for each
A shirt in nine colors and six sizes can be fifty-four addresses or one page with options. One page concentrates the signals and is usually right. The exception is where people genuinely search for the variant, a specific colourway, a particular capacity, in which case that variant earns its own page. Decide per product family rather than by a platform default nobody chose.
Decide what happens when a product goes out of stock
Deleting the page throws away everything it earned; leaving it unchanged frustrates buyers; redirecting everything to a category loses the specificity. The workable answer depends on whether it is coming back: keep and say so if it is, redirect to the closest genuine replacement if it is not, and never leave a page claiming availability that does not exist. Structured data must match reality.
Pagination and infinite scroll have to leave crawlable paths
Category pages that load more products only when a person scrolls give a crawler the first page and nothing else, which can leave most of a catalog undiscovered. Whatever the interface does for people, there must be real linked addresses for every page of results, and a sitemap covering every product regardless.
Speed is a revenue number on a catalog
Product and category pages carry many images and many third-party scripts, and most shop traffic arrives on phones. Image weight is nearly always the largest cause, followed by tag-manager scripts added for trials nobody removed. Auditing what actually loads, and removing what nobody can name a purpose for, is usually the fastest improvement available on any shop.
Questions people ask about ecommerce technical SEO
What is the biggest technical problem on an ecommerce site?
Faceted navigation. Filter combinations multiply into effectively unlimited near-duplicate addresses, and crawlers spend their budget there instead of on real product pages.
Should each colour and size have its own page?
Usually not. One page with options concentrates the signals. The exception is where people genuinely search for a specific variant, which is a per-product-family decision rather than a platform default.
What should happen to out-of-stock products?
Keep the page and say so if it is returning; redirect to the closest genuine replacement if it is not. Never leave structured data claiming availability that does not exist.
Does infinite scroll hurt?
It can leave most of a catalog undiscovered, because a crawler sees only the first set of products. Whatever the interface does for people, there must be real linked pages and a complete sitemap.