Ecommerce SEO becomes more difficult as the catalog grows.

A store with a small catalog can often be understood by both users and crawlers almost accidentally. A larger catalog with colours, sizes, multiple categories, filters and pagination can generate many URLs and several versions of effectively the same information.

At that point, SEO stops being a copy problem and becomes an architecture problem. Google’s ecommerce documentation covers site structure, crawlable links, product data, URL design and structured data; its guidance on faceted navigation addresses the crawl space filters can create.

At a glance

  • Build the catalog around search intent: category for broad terms, product for specific ones.
  • Decide per variant whether it deserves its own indexable URL.
  • Treat filters as UX, not as SEO landing pages.
  • Make the important taxonomy crawlable; keep parameters and empty states controlled.

Build the catalog around search intent

A useful hierarchy follows intent:

HOME → CATEGORY → SUBCATEGORY → PRODUCT

Categories and products answer different searches. A product page answers “brown leather belt model X.” A category answers “leather belts.” Broad transactional searches usually fit category pages better than individual products, which is why the taxonomy should reflect how people search, not only how the warehouse is organised.

Google’s ecommerce guidance recommends making important pages reachable through crawlable links. Products and categories that matter should not be discoverable exclusively through search, JavaScript events or filters.

A healthy structure links the home page to priority categories, categories to products and subcategories, product pages back to their category and to related products, and editorial content to relevant categories and products. A sitemap helps discovery; it does not replace internal architecture.

Product variants need a deliberate URL strategy

Variants are one of the easiest ways to create duplication. A product in three colours and four lengths is twelve purchasable variants — but that does not automatically mean twelve indexable pages.

Variant situation Recommended treatment
No independent search demand One canonical product URL
Meaningful independent demand Dedicated URL per variant
Mixed Canonical to the parent, index selected variants

Google recommends consistent canonical handling when variants use separate URLs, and provides ProductGroup structured data for describing product variants.

Faceted navigation: excellent UX, dangerous crawl space

Filters are useful and also one of the most common technical SEO problems in large stores. Every additional facet multiplies the number of possible URLs. Colour, size, material and sort combinations can create an effectively infinite URL space, wasting crawl resources and slowing discovery of valuable pages — something Google specifically warns about.

The answer is not to index every filter. It is to ask whether a filtered combination represents meaningful search demand:

URL pattern Index?
/leather-belts/ Yes — curated category
/belts/?material=leather Sometimes — if it maps to a real term
/belts/?material=leather&size=93&sort=price-desc No — parameter state

Separate navigation from SEO landing pages

Treat UX facets and SEO categories as two systems. Facets are dynamic and not intended for indexing. SEO categories are curated, stable and indexable. This gives buyers flexible navigation without turning every filter state into a search result. The same principle governs on-site search and filtering.

Empty categories should not become indexable dead ends

If a filter combination cannot return products, avoid leaving a crawlable, indexable dead end. Google’s faceted-navigation guidance recommends returning HTTP 404 for URLs that have no results; for temporarily empty but valid categories, define a separate policy based on whether the page still offers useful content.

Product structured data

Product markup helps Google understand identity, price, availability, offers, variants, ratings where legitimate, and shipping. For purchasable products, Google distinguishes merchant listings from editorial product snippets. Structured data does not guarantee a rich result, but it improves understanding and eligibility for richer commerce appearances.

For products with variants, Google supports ProductGroup containing individual Product entries, with properties describing what varies — colour, size, material or pattern. This is preferable to making Google infer variant relationships from page similarity alone.

Merchant Center and structured data solve different problems

Structured data lives on the page; Merchant Center provides a product feed. Google’s ecommerce documentation recommends using both where appropriate, and feeds give more control over rapidly changing data such as stock and pricing in large catalogs.

The governing principle is consistency. The price on the page, in the structured data and in the feed should agree. Conflicts create ambiguity.

Product information quality affects SEO and UX

Thin product pages are not improved by repeating a keyword. A good product page answers real purchase questions: material, dimensions, compatibility, use, care, delivery, availability, variants and specifications.

Useful product data also feeds internal search, filters, structured data, Merchant Center, comparisons and recommendations. Product data architecture is simultaneously an operational, UX and SEO concern.

Image SEO

Ecommerce imagery is content. Use descriptive alt text, appropriate dimensions, modern formats, stable aspect ratios, meaningful surrounding text and correct lazy or eager loading. The primary product image should not be lazily loaded if it is the LCP candidate above the fold; gallery images usually can be. The performance side of this is covered in Core Web Vitals for ecommerce.

Pagination, infinite scroll and sitemaps

Interfaces can load products progressively, but crawlers still need discoverable URLs. Do not assume that because a user can scroll indefinitely, Google will discover every product the same way.

A sitemap should contain canonical, indexable URLs only — not sort URLs, filters, tracking parameters, duplicate variants, internal search results or empty categories. For larger stores, splitting sitemaps by content type simplifies monitoring.

What to monitor in Search Console

Review indexed URLs, discovery of new products, whether categories rank for category intent, which canonical Google chose, product markup errors and Core Web Vitals field issues by template.

Ecommerce SEO problems are frequently template-level problems. Fixing one template can improve hundreds of URLs.

The core principle

Do not let every possible catalog state become an SEO page. Make the important commercial taxonomy explicit and crawlable, and keep filters, sort states, parameters and empty combinations controlled. Good ecommerce SEO is not creating the maximum number of pages; it is making the valuable pages exceptionally clear.

Continue reading

This is the structure work behind Web & Growth and Commerce at FRN.Q.