In brief: Ecommerce SEO is not won with attractive copy — it is won in the architecture. A clear category structure, well-crafted product pages, and firm control of which URLs get indexed (filters, pagination, out-of-stock products) make the difference between ranking and disappearing. This guide covers how I approach it, step by step, with a focus on transactional intent, ecommerce-specific technical SEO, and product structured data. No fluff, all substance.
Why ecommerce SEO is different
When I work on the SEO of a corporate or services website, almost everything revolves around a handful of key pages. In an online store the problem is one of scale: you can have hundreds or thousands of URLs across categories, subcategories, product pages, filters, and combinations. Most of those URLs compete with each other, cannibalise each other, or simply pollute Google's index with pages that add nothing.
That is why the priority here is not to write more, but to decide what deserves to exist as an indexable page and what does not. If you come to SEO thinking of it as writing articles, shifting that mindset is the first step. Before touching anything it is worth having a serious diagnosis of the current state, and for that I usually start with an SEO audit that lays out what is being indexed, what is ranking, and what is broken.
Architecture and categories: the ecommerce foundation
If I could work on only one thing in an online store, it would be the architecture. Category pages are the ones that capture the most transactional traffic, because they respond to broad searches such as "running shoes" or "fabric sofas." Individual products capture long-tail; categories capture volume.
My approach is to build a logical, shallow tree. A user (and a crawler) should be able to reach any important product within three clicks or fewer from the homepage. Main categories at the top, subcategories below, and — when search volume justifies it — pages for specific attributes such as "red dresses" or "cordless drills."
Each category deserves a descriptive H1, a clean semantic URL, and a genuine block of text above or below that explains what the user will find there. Not 600-word filler repeating the keyword: two or three useful paragraphs that guide the purchase decision and answer catalogue questions. That page is the one that will compete for the main search query, so it is where I concentrate the copy and linking effort.
Transactional keyword research
Keyword research for an ecommerce store is nothing like research for a blog. Here commercial and transactional intent is paramount: "buy," "price," "deal," brand and model names, sizes, colours. My method is to map keywords to page types before writing anything.
Category-level searches ("women's hiking boots") go to category pages. Specific model searches ("Salomon X Ultra 4 size 40") go to product pages. Informational searches ("how to choose hiking boots") go to the blog. Confusing these levels is the number-one cause of the keyword cannibalisation I see in online stores: three pages fighting for the same query and none of them winning.
If you are weighing up whether to invest in organic ranking or paid campaigns for your most competitive terms, I recommend understanding the differences between SEO and SEM before splitting the budget. In ecommerce the norm is to combine both, not to choose one.
Product page optimisation
The product page is where the sale closes, but it is also where the most corners get cut. The classic mistake is using the manufacturer's description verbatim — identical to the one on 200 other stores. To Google that is duplicate, low-value content, and you are competing at a disadvantage against portals with far more authority.
What I work on in each product page that merits the effort:
- An H1 with the actual product name, including brand and model.
- An original description that answers purchase questions: who it is for, what problem it solves, what differentiates it.
- Technical specifications in a scannable format (list or table), because many buyers make decisions there.
- Images with descriptive file names and useful alt attributes.
- Genuine customer reviews, which naturally supply fresh, unique content.
There is no need to write a novel for every SKU. In large catalogues I prioritise: first, the pages for the products that sell or get searched for most; the rest, with a decent template and short but original descriptions.
Content: categories and blog
Content in an ecommerce store lives in two places. On one side, the category texts already mentioned, which have a direct transactional objective. On the other, the blog, which captures the research phase before the purchase.
A good store blog answers the questions someone asks before buying: sizing guides, comparisons, how to care for the product, which model to choose for a given use. Those articles rank for informational searches, attract users who are not yet buying, and I then push them towards categories and product pages with well-placed internal links. The blog does not sell on its own; it feeds the pages that do.
Ecommerce-specific technical SEO
This is the part that truly separates an optimised store from one swimming against the tide. These are problems that barely exist on small sites and multiply in ecommerce.
Facets and filters
Navigation filters (colour, size, price, brand) generate near-infinite URL combinations. If you let Google index all of them, you end up with thousands of near-duplicate pages devouring your crawl budget and diluting authority. My rule: I index only filter combinations with genuine search demand and commercial value — for example "women's running shoes." The rest I block from the index with a meta robots noindex tag, or prevent from generating crawlable URLs. The rule is simple: if nobody searches for that combination, it does not deserve to be an indexable page.
Pagination
Long categories split into pages (1, 2, 3…). The correct way to handle them today is for each page in the series to be self-canonical — that is, it points its canonical to itself, not to page 1. That way Google understands that page 3 is a real page with different products. What you should not do is point the canonical of every page to the first one, because then the products on subsequent pages can fall out of the index.
Canonical tags
The canonical tag is your tool for telling Google which is the authoritative version of a URL when several similar ones exist. I use it to consolidate sorting parameters (sort by price, by newest) towards the clean category URL, and to prevent the same product accessible via multiple paths from being counted as a duplicate. Misused, however, it hides pages you actually wanted to rank, so it is worth reviewing carefully.
Out-of-stock or discontinued products
This is badly neglected and has a direct impact. My approach depends on the case:
- If the product will come back into stock, I keep the page live, flagging the lack of availability and offering alternatives.
- If it disappears permanently but there is a clear successor, I do a 301 redirect to that equivalent product.
- If there is no logical replacement, I redirect to the parent category so as not to lose the link equity or the trail of authority.
What I avoid at all costs is leaving hundreds of pages returning 404 errors unaddressed, or deleting them silently. Every poorly managed discontinued product is traffic and authority thrown away.
Speed
A slow store loses sales and rankings. Product pages with many images and third-party scripts tend to load poorly on mobile, which is where most people shop. I work on image compression, lazy loading for off-screen content, and stripping scripts that add nothing, keeping an eye on Core Web Vitals. I am not chasing a 100 in tools; I am chasing a page that is usable quickly on a normal phone with a normal connection.
Product structured data
Here ecommerce has an enormous advantage over other sectors. The Product schema markup allows Google to show price, availability, and star ratings directly in the search results. Those rich results attract far more attention and typically improve click-through rates.
On every product page I implement the Product schema with its offer property (price and currency), availability (in stock or out of stock), and — if you have genuine reviews — the aggregate rating block. An important note: ratings must be authentic and visible on the page. Marking up invented stars or ones the user cannot see is exactly what Google penalises, and you also risk losing the rich result entirely. Mark up only what genuinely exists on the page.
| Technical element | Problem if ignored | My approach |
|---|---|---|
| Filters and facets | Thousands of near-duplicate URLs in the index | Index only those with real demand; noindex the rest |
| Pagination | Products on pages 2+ outside the index | Self-referencing canonical on each page in the series |
| Out-of-stock product | 404 errors and lost authority | Keep, 301 redirect, or send to category depending on the case |
| Schema Product | Results without price or stars, fewer clicks | Mark up with price, availability, and real reviews |
| Mobile speed | Abandonment and lower rankings | Compressed images, lazy loading, script clean-up |
Internal linking
Internal linking distributes authority and tells Google which pages matter. In a store I work it in several directions: from the homepage and menus to main categories; between related categories; from product pages to their parent category and to complementary products; and from blog articles to the categories and product pages that resolve purchase intent. "Related products" and "customers also bought" blocks are not just cross-selling tools — they are internal links that help crawling and indexation.
Common ecommerce mistakes
- Product descriptions copied from the manufacturer, identical to those of 200 other stores.
- Letting all filter combinations be indexed and blowing the crawl budget.
- Pagination canonicals always pointing to page 1.
- Deleting discontinued products without redirecting, generating 404s in bulk.
- Non-existent category texts or, at the other extreme, endless filler paragraphs with no value.
- Schema Product with ratings the user cannot see on the page.
- Forgetting local SEO when there is a physical store. If you sell online but also have a retail location, working on local SEO captures nearby purchase-intent searches that a purely online competitor cannot touch.
Measurement
Without measurement you do not know if things are working. In ecommerce I do not stop at rankings and traffic. I monitor performance by page type in Search Console (how categories perform vs. product pages vs. blog), indexation coverage to catch junk URLs early, and actual behaviour in web analytics through to conversion. A keyword that climbs in position but generates neither visits nor sales is not a win — it is a vanity metric.
Conclusion
Ecommerce SEO is won in the unglamorous work: a clean architecture, firm control of what gets indexed, and product pages that genuinely help people buy. The visible work (blog, copy) adds value, but it does not save a poorly built structure. If you have an online store that is not gaining traction on Google, the problem is almost always in these foundations, not in a lack of content.
If you want me to review your store and tell you where the leaks are, I work on this from Valladolid and Las Palmas through my SEO and SEM consultancy service. Tell me your situation and let's put some numbers to it.