Ecommerce SEO carries all the usual fundamentals and then a technical surface no other site type has. The failures below are specific to stores, and they matter more than on-page work because each one wastes crawl budget at a scale a brochure site never approaches.
Ordered by how much damage each does. Fix the top of this list before optimising anything.
1. Faceted navigation
The single largest source of wasted crawl budget on most stores.
Filters combine. Colour, size, brand, price band and sort order multiply into thousands of URLs, nearly all near-duplicates of each other, all consuming crawl capacity that should be reaching products.
- Establish how many URLs facets can generate. Multiply the filter options. The number is usually alarming
- Block or
noindexfilter combinations by default. Indexing them is the exception, not the rule - Disallow sort and view parameters —
?sort=,?view=,?page_size=. These never deserve indexing - Identify the handful of facets with real search demand — "black running shoes" may genuinely be searched. Promote those to real, statically-generated pages with unique content, not to indexed filter URLs
- Ensure filter links do not create infinite crawl paths where each combination links onward to further combinations
- Compare indexed page count to catalogue size. A large multiple is the finding
2. Product variants
- Decide the variant model deliberately — one canonical page covering variants, or separate pages where demand genuinely differs
- Canonicalise variants to the parent where they are near-identical
- Give a variant its own page only if it has its own demand and content that differs meaningfully. "Same product, different colour" is not different content
- Check variant URLs are not generating parameter duplicates alongside the canonical page
Separate indexable pages for every size and colour split relevance across pages that should have been one, and multiply the crawl surface for no gain.
3. Out-of-stock and discontinued products
A decision, not a default — and most stores never make it explicitly.
- Temporarily out of stock: keep the URL, keep the page, say so on it. The ranking and links it earned are preserved for when it returns
- Permanently discontinued with a successor: 301 to the successor
- Permanently discontinued with no successor: 301 to the closest category, or 410 if genuinely nothing is equivalent
- Never delete silently. A 404 discards the equity and sends a searcher nowhere
- Check whether the platform auto-removes URLs when stock hits zero. Several do, and it is rarely what you want
Because facets and variants generate their damage automatically once the catalogue grows, the architecture decisions here belong at build time — which is part of
stores built with the crawl surface designed rather than inherited.4. Category pages
Usually where the higher-volume commercial terms rank, and usually the most under-invested pages on a store.
- Treat categories as landing pages, not lists. They need their own content
- Unique description on every category, above or below the grid, that helps someone choose
- No duplicate category content across near-identical categories
- Pagination handled correctly — self-referencing canonicals on page 2 onward, never canonicalised to page 1
- Internal links from categories to their best products, not only to the first twelve alphabetically
- No empty categories indexed
5. Product pages
- Unique descriptions. Manufacturer copy appears on every retailer selling the item, which is the definition of duplicate content
-
Productstructured data with price and availability, kept accurate - Review markup only where reviews genuinely exist on the page
- Images with descriptive alt text and stable, indexable URLs
- Breadcrumbs matching URL depth, with markup
- The information a buyer needs to decide — sizing, materials, delivery, returns. Missing information is a conversion problem before it is an SEO one
6. Technical foundations
The usual checks, with ecommerce weighting.
- Crawl budget spent on products, not parameters. Verify in Search Console
- XML sitemap contains canonical product and category URLs only — no facets, no parameters, no out-of-stock removals
- Site speed on product and category templates, which is where traffic lands
- Mobile checkout genuinely usable, tested on a real device
- HTTPS everywhere, no mixed content
- No session identifiers in URLs
- Internal search result pages not indexed
Priority, honestly
If the catalogue is large, items 1 to 3 outrank everything else on this list. A store where facets generate fifty thousand crawlable URLs across a two-thousand product catalogue has a crawl problem that no amount of product-page optimisation will overcome — the crawler is spending its budget on pages that cannot rank.
The general technical audit, ordered by dependency, is in the SEO audit checklist. Run that first; this list is what it does not cover.
And once the crawl surface is right, the pages that actually convert are the ones worth the remaining effort — product page SEO covers where search intent and conversion meet.
Working through this list?
Most teams get through the audit and stall on the remediation. That is the part we usually take on.