E-commerce SEO: the guide for online stores that want to scale

Most online stores do not lose search traffic because they published the wrong blog posts. They lose it because of decisions taken once, often by a developer or by a platform default, that then repeat across ten thousand URLs. A category template that renders no text, a filter system that generates infinite combinations, an import that copies the manufacturer description verbatim: each is one choice, multiplied by the size of your catalogue. That multiplication is what makes e-commerce SEO a different job.

The consequence is that leverage sits in the template, not in the page. Fixing one category page is worth almost nothing; fixing the category template is worth the entire tree. Practitioners from a content background often miss this and start writing, because writing is visible and reassuring. Teams that have actually scaled a store start with the crawl instead: what Google can reach, what it wastes its time on, and which of your commercially valuable URLs it has simply never seen.

This guide covers the structural decisions that decide whether a store can rank at all: category depth and category content, product uniqueness at catalogue scale, faceted navigation, the lifecycle of products that go out of stock, structured data, template weight and internal linking. It assumes the general hygiene work described in the technical SEO audit checklist is already on your roadmap.

Buyers rarely search for a reference number. They search for a class of product with a qualifier: waterproof hiking boots, standing desk for a small room. Those queries carry both the volume and the commercial intent, and the page that should answer them is a category or collection page, not a product page. Most stores treat categories as filtered lists with a title and forty thumbnails, then wonder why a marketplace outranks them. This is the largest untapped surface in most retail SEO programmes.

Making a category substantial does not mean pasting six hundred words of keyword soup under the product grid. It means answering the questions a buyer has before choosing: how to pick a size, what the main types are and who each one suits, what a realistic price range looks like. Put a short orientation paragraph above the grid, forty to eighty words, and the longer guidance below it, written as answers rather than as filler.

The second decision is how many category pages should exist at all. Create one wherever there is genuine repeated demand for a combination, and only then: brand plus category, material plus category, use case plus category. Each needs its own reason to exist and its own copy, which makes category content a content workload as much as a technical one. Categories created speculatively, with three products and recycled text, dilute the tree and compete with the page you wanted to rank.

Most catalogues are imported. The description, the specifications and often the photography come from the manufacturer, which means several hundred retailers publish the same paragraphs. Google does not penalise that, but it has no reason to prefer your copy to the brand's own site or to a marketplace with more authority. The page is not bad, it is interchangeable, and interchangeable pages sit on page three. Uniqueness on a product page is not a rewriting exercise, it is a question of what you know that the manufacturer does not.

What you know includes sizing behaviour observed from returns, compatibility with the other products you stock, honest notes on who a product does not suit, real photographs, customer questions and their answers. None of that can be scraped from a feed. The constraint is volume: nobody rewrites forty thousand pages. Tier the catalogue by margin and by demand, treat the top few hundred as editorial pages, give the middle band a template with two or three genuinely specific fields, and leave the long tail alone. Serious candidates in any directory of SEO agencies will propose that triage rather than a blanket rewrite.

Variants are the other trap. A colour or size with its own search demand deserves its own indexable URL; the other thirty do not, and should either canonicalise to the parent or never generate a URL. Watch for near duplicates created by the platform: the same product reachable through two category paths, print views, and combinations nobody sells. Each consumes crawl attention and splits the signals that should be accumulating on one single address.

Filters are the most useful feature of an online store and the most destructive one for search. Four facets with six values each, combinable, produce tens of thousands of URLs from a single category. Add price sliders and sort parameters and the number becomes effectively infinite. If those URLs are linked and crawlable, Google spends its crawl budget enumerating blue medium cotton shirts under fifty euros, and discovers your new products late or not at all.

The fix is a deliberate policy, decided once and enforced in the template. Choose the small set of facet combinations that have real search demand and give them clean static URLs with their own titles and copy. Everything else should not become a crawlable link at all: render those filters so they do not produce an anchor tag, or block the parameter pattern in robots.txt. Noindex alone is a weak remedy, because the page still has to be crawled before the directive is seen.

Measure before and after. The crawl stats report in Search Console shows how many requests hit parameter URLs; server logs show it precisely. Compare the number of URLs a crawler finds on your site with the number of products you actually sell: a ratio above ten to one usually means the facets are leaking. Most crawlers and log analysers are inexpensive, and you can compare them in the SEO tools directory before committing to a stack.

Retail catalogues churn, and the way a store handles that churn decides how much of its accumulated authority survives. The default reflex, deleting the page and letting it return a 404, is almost always wrong: a product that ranked has external links, internal links, reviews and a history of impressions, and all of it evaporates the moment the URL disappears. The first rule of catalogue maintenance is that you never delete a URL that has traffic, links or seasonal demand.

Temporary out of stock means the page stays live, stays indexable and keeps its content. Say plainly that the item is unavailable, give an expected restock date, offer a notification signup and link to the closest alternatives you stock. Keep the availability field in the structured data accurate, because a page that claims to be in stock when it is not will eventually lose its rich result.

Permanent discontinuation is a redirect decision. If a genuine successor exists, redirect to it; if not, redirect to the parent category rather than to the homepage, which tells nobody anything. When nothing relevant remains, a 410 is honest and clears the index faster than a 404. Seasonal pages deserve the opposite treatment: keep the same URL year after year and refresh it, instead of creating a new page each season and starting again from zero authority.

Product structured data turns a plain listing into a result carrying a price, an availability status and a rating. The required fields are unglamorous: name, image, description, brand, an identifier such as GTIN or MPN, and an offer with price, currency and availability. Accuracy matters far more than completeness. Markup that says forty euros while the page displays fifty is a fast route to losing the enhancement entirely.

Free merchant listings extend the same data into shopping surfaces, fed by a product feed rather than by your pages. That split causes a recurring organisational problem: the feed belongs to the commerce or paid media team, the pages belong to SEO, and the two drift apart until titles, prices and availability disagree. Treat them as one dataset with one owner. Shipping and return policies now carry weight in those surfaces too, so the operational detail belongs in the feed, not buried in a policy page.

Template weight is the other structural decision. Store themes accumulate sliders, review widgets, chat bubbles and tag managers, and a category grid of sixty images is heavy before any of that loads. The realistic wins are boring: modern image formats at correct dimensions, lazy loading below the fold, removing third-party scripts nobody uses, and rendering the first screen server side. This is usually a web development mandate rather than an SEO one. Judge it on field data, and compare measurement tools on the tools leaderboard rather than by reputation.

Authority arrives at your homepage and has to travel to the pages that make money. On most stores it does not: products link sideways to random related items, and the guides published by the content team link nowhere useful. The structure that works is a hub. A top category holds the strategic copy and links down to its sub-categories, which link to products; guides and comparisons link back up into the categories they support, using descriptive anchors instead of a generic read more.

The mechanics are mundane and rarely done. Breadcrumbs on every page, related products chosen by genuine relevance rather than by a random query, editorial links inserted inside category copy, and a periodic hunt for orphan pages. Pagination should keep deep products reachable within three or four clicks of the homepage. If you sell across borders, the same hub has to exist per market, and mapping them together is a separate discipline covered in the guide to international SEO and hreflang.

Briefing an agency on an e-commerce mandate differs from briefing one on a blog. Give them the catalogue size, the platform, who owns the theme, the product feed and its owner, your margin structure by category, and read access to Search Console. Ask what they would do about facets and product uniqueness at your volume: the answer separates teams who have scaled a store from teams who have not. Market shortlists such as SEO agencies in France or agencies in the United Kingdom are a reasonable starting point, and the selection process itself is covered in the guide to choosing an SEO agency.