E-commerce SEO

Where e-commerce SEO starts – the order that saves budget

I start with what Google sees and indexes, then map keywords to pages, and only then add content and links. In a store the reverse order is the most expensive one: category copy and links won't help while Google wastes its time on thousands of filter URLs or indexes three versions of the same product page.

In practice I run stages 1 and 2 in parallel during the first weeks and start stage 3 with the few categories that have the highest revenue potential. That way the first changes reach production quickly instead of after months of analysis.

Keyword research for online stores – mapping queries to categories, filters and products

Every group of keywords with the same intent gets exactly one target page. Broad product queries go to a category, queries with an attribute (brand, material, use) to a subcategory or a selected filter, model names to a product page, and questions or comparisons to a guide.

Where the keywords come from

• Search Console: queries the store already gets impressions for but ranks poorly on. This is the fastest source of growth.
• On-site search: search events in GA4 show the words customers use for products, often different from category names.
• Keyword tools (Ahrefs, Semrush, Senuto for Poland): volumes, difficulty and the terms competitors rank for.
• The results page itself: if Google shows store categories for a query, I build a category; if it shows guides, I write a guide instead of forcing the query onto a product list.

Finally I check for cannibalisation: whether Search Console shows several store URLs alternating for one query. The usual pairs are a category and a filter, or a category and a blog post on the same topic. I resolve it by changing content, internal links or merging pages, not with a canonical alone.

Category pages – the anatomy of a page that ranks and sells

The category page is the most important SEO page in a store, because it answers the broadest, most profitable queries. It has to give Google a clear topic and copy, and give shoppers fast access to products. So the copy must not push the product grid below the fold, but it can't be missing either.

What a category page needs

• An H1 with the exact category name, matching the main keyword, and a title tag that starts with it. H2s and H3s in the copy add context: types, uses, how to choose.
• A short intro above the products (2–3 sentences) and a longer part below the grid: how to choose, how the product types differ, common questions. I write it for the buyer, not to hit a character count.
• Links to subcategories and to selected filters that have their own search demand, plus guides on the topic.
• Plain a href links to products. Google strongly recommends that category pages link to every product you want indexed and warns against navigation built on JavaScript events instead of links.
• Breadcrumbs visible on the page and described with BreadcrumbList data, so the hierarchy is clear to people and crawlers.
• A fast first screen: product images in a modern format with explicit dimensions, and no large banners delaying LCP.

For a beauty and fashion store I rebuilt headings on category and product pages on the same principle: the H1 is the precise entity name, and H2s and H3s are layers of context that support long-tail queries. The project is described in the Beauty E-Commerce SEO/GEO case study.

Product pages – what I check and fix

A product page wins with information competitors don't have: concrete specs, answers to buyers' questions, original photos and reviews. A manufacturer description that appears on dozens of stores gives Google no reason to show your page in particular.

Images matter twice in a store, because Google Images and Lens are a separate traffic source. Google's documentation notes that alt text is the most important image metadata, that CSS background images aren't indexed, and that the text near an image provides context. I cover this in more depth in the article on image SEO and visual search.

Nobody can hand-write thousands of descriptions for a large catalogue, so I combine SEO with AI product descriptions: the model works from product attributes in the database, and quality rules keep the output from turning into template text. I always refine the pattern by hand on the best-selling products first.

Filters, variants and duplicates – how a store creates its own competition

Most duplicates in a store don't come from copied content but from the way the platform generates URLs. The same product list is available with a sort parameter, a filter, a tracking parameter and a page number, and the same product under several category paths.

Google sets out these rules in its documentation on ecommerce URL structure: every variant that should be found needs its own URL, fragment identifiers after # aren't used for indexing, and the number of alternative URLs returning the same content should be kept to a minimum. For filters, Google's advice is that URLs you don't need in search results shouldn't be crawled at all.

I cover this in full, with server log analysis and robots.txt rules, in crawl budget optimization. In e-commerce SEO I always address it, but to the extent the store really needs: a small catalogue often needs only canonical and link fixes, a large one also needs changes to how filters are generated.

Pagination and out-of-stock products – two decisions stores often get wrong

Each page of a product list should have its own URL and its own canonical, and an unavailable product shouldn't automatically disappear with a 404. Both mistakes either cut Google off from products or drop pages that have links and traffic.

Pagination

• Google treats paginated URLs as separate pages and advises against using the first page as the canonical for the whole series.
• Google no longer uses rel=next and rel=prev, although other search engines may still read them.
• With “load more” buttons and infinite scroll I make sure the subsequent pages are also reachable through plain links, because Google mainly follows URLs in href attributes.

Out-of-stock and discontinued products

Google's John Mueller said in 2022 (as reported by Search Engine Journal) that Google may treat an out-of-stock product page much like a soft 404 and drop it from search results. So I don't leave empty pages that just say “out of stock”. The decision depends on whether the product will return and whether it has SEO value:

Product structured data and Google Merchant Center – two sources that must agree

Google recommends using both structured data on product pages and a Merchant Center feed, because crawling isn't guaranteed to find every product. Merchant Center participation is required for listings in the Shopping tab, while structured data powers product rich results in regular search.

Google launched free product listings in the Shopping tab in April 2020, first in the US, with plans to expand worldwide by the end of that year. According to Merchant Center Help, products can now appear at no cost across Google Search, the Shopping tab, Images, Lens, YouTube, Maps and Gemini. Google also notes that products aren't guaranteed to show, because it relies on the data you provide.

What I implement and check

• Product with Offer on every product page: price, currency, availability, condition, GTIN or manufacturer code, brand, images. Values come from the same source as the visible content, so JSON-LD never shows a different price from the page.
• Variants as ProductGroup with hasVariant when the store sells colours and sizes, plus shipping and returns policies.
• A Merchant Center feed with the required attributes (including ID, title, link, image, price, description, availability, GTIN or MPN) and proper product titles rather than abbreviations from the stock system.
• Free listings switched on under marketing methods in Merchant Center, and automatic item updates based on website data, which Google suggests to cover feed sync delays.
• The merchant listings and product snippets reports in Search Console, plus Merchant Center diagnostics after every template change.

The markup itself, with an @id graph and validation, is described in detail in Schema.org implementation. In e-commerce SEO it's one part of the work, not a separate project.

SEO for Shopify, WooCommerce, PrestaShop and Next.js stores

The SEO rules are the same everywhere, but each platform breaks them in a different place. So I start an audit from the known issues of the platform and theme, and only then crawl the whole store.

On Shopify, be careful with robots.txt edits. The platform only allows changes through the robots.txt.liquid template, and Shopify Help warns plainly that incorrect use of the feature can result in loss of all traffic. So I test every rule against a list of crawled URLs before publishing.

I build and fix headless Next.js stores as a developer, so I know client-side rendering problems from the inside. I describe them in the article on JavaScript SEO, and the store build itself in the headless Shopify store service.

SXO for online stores – where SEO meets conversion

SXO combines SEO with user experience: a page should not only attract traffic but also let people find and buy a product quickly. In a store the same changes often serve both goals: a faster category, usable filters and visible delivery costs improve both Google's signals and sales.

A good example is the Japanese store Rakuten 24, documented on web.dev: in an A/B test, the version optimised for Core Web Vitals had 53.37% higher revenue per visitor and a 33.13% higher conversion rate than the unchanged version. So I treat the speed of category and product templates as part of SEO and sales at the same time.

What I check from the buyer's point of view

• Whether a visitor from Google lands on a page that matches their query, not the homepage or a generic category.
• Whether filters work quickly on a phone and it's easy to get back to the list after filtering.
• Whether price, availability, delivery cost and delivery time are visible without scrolling and before the cart.
• Whether the product page answers the questions customers ask support, and whether reviews are visible.
• Whether conversion changes (pop-ups, widgets, A/B tests) slow the page down or create duplicates.

When a store has traffic but sales are the problem, I move on to conversion rate optimization (CRO) with funnel analysis and A/B tests. If you don't know whether traffic, conversion or tech is the bottleneck, an e-commerce audit is a good place to start.

Supporting content and link building for online stores

In a store, guides and comparisons exist to catch the customer before the decision and pass authority to categories. I plan every piece together with the category page it should lead to, not as a standalone blog chasing traffic.

• A buying guide (such as “how to choose running shoes”) links to the categories and filters that match the criteria it describes.
• Comparisons and roundups link to specific product pages.
• Categories link back to the guides, so both shoppers and crawlers see the full topic.

I built this kind of cluster architecture for a Polish slow fashion brand: 14 topic clusters, 5 pillar pages and 62 articles connected by internal links, backed by 8 digital PR placements. Organic search became the brand's number one acquisition channel, ahead of paid ads. Details are in the Fashion Brand SEO/GEO case study, and the method in the article on topical authority.

Links to the store

I don't buy “X links per month” packages. For a store I look for placements on topically relevant sites with real traffic and readers: industry media, creators who test products, partners' guides. Links go to categories and guides, not only to the homepage. If a placement is paid, Google's policies require the link to carry rel="sponsored", which we agree on before buying. How I vet sites is described in link building.

Online stores in AI Mode and ChatGPT – what you can realistically do in 2026

In AI shopping, stores with complete, up-to-date product data and access for AI search crawlers win. There's no separate “ChatGPT SEO” for products: the same data a store should have for Google is what counts.

• Google AI Mode combines Gemini with the Shopping Graph, which according to Google (May 2025) holds more than 50 billion product listings, with more than 2 billion refreshed every hour. A complete Merchant Center feed and correct on-page data are your way into that database.
• ChatGPT, according to OpenAI's help centre, builds a product's merchant list from metadata supplied by third-party providers or by merchants directly, and ranks merchants on factors such as availability, price, quality and whether they are the maker or primary seller. OpenAI also stresses that product results are not ads.
• Crawler access: OpenAI recommends allowing OAI-SearchBot in robots.txt for sites that want to appear in ChatGPT search. Blocking GPTBot only signals that content shouldn't be used for model training.
• Feeds for agents: OpenAI publishes a product feed specification as part of the Agentic Commerce Protocol. It matters for stores that want to sell directly inside ChatGPT, not as a requirement for appearing in results.

I go deeper, including the ACP and UCP protocols, in E-commerce SEO & GEO and Agentic Commerce. Store content for citations in Google's AI summaries is handled in AI Overviews and AI Mode optimization.

Replatforming a store without losing traffic

A store migration is the easiest moment to lose years of SEO work, because URLs, templates and product data all change at once. So in e-commerce the redirect map is only part of the job: moving attributes, reviews, images and filter rules matters just as much.

• I match product URLs between platforms by SKU or EAN, not by name, because names often differ.
• I map product attributes to fields on the new platform, so the specs needed for filters, structured data and the feed don't disappear.
• Filters with demand get equivalents in the new structure, and the rest get rules that keep them out of the crawl from day one.
• I switch the Merchant Center feed together with the site, so free listings don't link to old URLs.

That's how I ran a WooCommerce to PrestaShop migration: every category, product and blog URL got its counterpart, product attributes were mapped to PrestaShop features, and the new structure was fully re-indexed within 3 months with no loss of visibility. The full migration process is described in SEO migration.

How I measure e-commerce SEO results

The main metric is organic revenue, not the number of keywords in the top 10. Rankings show direction, but a store can climb on keywords that don't sell and slip on the ones that do.

I always compare year over year, because seasonality in stores can hide both growth and decline. For larger template changes I annotate the release date in reports, so the effect of a specific fix is visible.

How much e-commerce SEO costs and what drives the quote

My e-commerce SEO starts from €850, and the final quote depends on store size, the number of products and categories, and the platform. Smaller stores with a simple structure sit closer to the starting price; the more templates, variants and technical issues, the larger the scope. I prepare the quote after a free consultation and a short review of the store.

If you only need one element, you can order it separately: a technical SEO audit from €800, Schema.org implementation from €600, crawl budget optimization from €900, an SEO migration from €1,200 or link building from €500 per month. E-commerce SEO combines these into one plan, ordered by impact on revenue.