How do you improve your ecommerce site's SEO?

The SEO of an ecommerce site rests on a clear site structure, unique product pages and content that renews itself; customer reviews are the cheapest source of fresh content a shop has. This guide covers which pages to prioritise, technical SEO, the product page, the role of reviews, stars in Google, visibility in AI answers and the traps specific to Shopify, PrestaShop and WooCommerce.

What is ecommerce SEO and why does it protect your margin?

Ecommerce SEO is the set of actions that get Google to rank your category pages and product pages for the searches your future customers type, without paying for each click. It differs from the SEO of an editorial site in what it deals with: thousands of product URLs that appear, change price, go out of stock and disappear, while a blog publishes ten stable articles a month. The difficulty is not writing. It is structuring.

The margin question comes up as soon as you look at your acquisition costs. A shop that makes 80% of its revenue through Google Ads and Meta Ads pays for each order at whatever the ad market charges that day. When a competitor outbids you, when a product launch saturates the auction or when the season pushes CPCs up, the order that cost £8 now costs £14, and the unit margin is gone. Organic traffic does not follow that curve: a category page ranking third for "women's walking boots" brings visits every day at zero marginal cost, campaign or no campaign.

That stability has a price: time. An ecommerce SEO strategy built on sound foundations usually takes three to six months to show visible results, and a year to weigh on revenue. In return, every pound invested in site structure, speed or reviews keeps working for years. It is the difference between renting your traffic and owning it. The shops that do best do not choose between the two: they use campaigns to test queries and SEO to make them profitable afterwards. The Review Collect for ecommerce page explains how reviews fit into that logic.

Which pages should you prioritise in an online shop's structure?

For an online shop's SEO, category and subcategory pages bring more qualified traffic than product pages, because they match the queries people actually type: "solid oak coffee table", not "Elsa coffee table 110 cm oiled oak". The first rule of a good site structure is therefore to create one category per search intent with volume, and only one.

The second rule concerns depth. Google crawls more often, and passes more authority to, pages close to the homepage; a product page reachable in three clicks (home, category, subcategory, product) is within the norm, while a page buried six clicks deep behind filters will be visited rarely and ranked poorly. Internal linking designed for SEO pushes strategic categories into the menu, the footer and "related categories" blocks, and links each product page to its parent category and to three or four similar products.

Third rule: every level carries its own content. A category without text is a list of thumbnails that Google cannot qualify. A paragraph of 150 to 300 words below the product grid, answering the buying questions of that category (materials, sizes, uses, care), is enough to set the page apart from the thousands of competing categories that have none.

Categories or product pages: which should you work on first?

Start with the ten categories that concentrate the most revenue, not with product pages. A well-built category pulls its products up; the reverse is not true. In Search Console, filter your category pages receiving impressions in positions 4 to 15: those are the ones where a rewritten title, an added category text and a few internal links gain positions within weeks. Product pages come next, and not all of them: only references with real search volume on their exact name (known brands, iconic products) deserve individual work. For the others, the quality of the product page template does the job at scale. A shop with 3,000 references that works on its 15 categories and its product template covers the essentials; one that rewrites 3,000 pages one by one exhausts its team for a marginal effect.

How do you handle filters and pagination without creating duplicate pages?

Navigation filters (size, colour, price, brand) generate an almost infinite number of URL combinations, and each combination is a page Google can crawl instead of your useful categories. The principle: index the combinations that match a real search ("white trainers men") by treating them as true subcategories with their own URL, and exclude all the others with a noindex tag or a robots.txt rule. For pagination, Google recommends a distinct URL per page, such as ?page=2, linked to the next one with an <a href> link and carrying its own canonical; it also states that it no longer uses rel="next" and rel="prev", and that a page number placed after a # is ignored. These rules are in Google's documentation on pagination and incremental page loading. Infinite scroll without crawlable links hides your products from Google beyond the first page.

How do you fix the technical SEO of an ecommerce site?

The technical SEO of a shop comes down to three areas that show up in almost every audit: loading speed, duplicate content and the handling of products leaving the catalogue. The other topics (hreflang tags for country versions, an up-to-date XML sitemap, chains of 404 errors) matter too, but those three explain most of the ranking losses seen on retail sites.

One point of method before opening the first tool: a shop does not need a perfect score, it needs a clean category template and a clean product page template, because those two templates account for 95% of its URLs. Fixing a problem at template level fixes it on every page at once. Mobile speed deserves particular attention: most ecommerce visits come from a phone, often on patchy 4G, and mobile is where Google evaluates your pages.

Finally, an ecommerce site lives: new collections, sales, withdrawn products, theme changes. Each change can break a canonical, create a redirect loop or put test pages back online. A monthly crawl of the site, even with a free tool, spots these regressions before they cost you positions.

Speed and Core Web Vitals

Core Web Vitals are the three measures Google uses to qualify a page's loading experience: LCP (display of the largest visible element) must stay under 2.5 seconds, INP (reaction delay after an interaction) under 200 milliseconds, and CLS (unexpected layout shifts) under 0.1. Google states that these measures are part of what its ranking systems seek to reward, without making them a dominant factor; the details are on the Google Search Central page on Core Web Vitals. On a shop, the usual causes are well known: uncompressed product images served at full size, third-party app scripts (chat, reviews, recommendations, tracking) loaded before the content, and promotional banners that slot in after loading and make the page jump. Converting images to WebP, deferring non-essential scripts and reserving space for banners solves most of it.

Duplicate content, canonicals and out-of-stock products

Duplicate content on an ecommerce site rarely comes from plagiarism; it comes from variants. The same t-shirt in five colours and six sizes can exist under thirty URLs, the same product can appear in three categories with three different paths, and the version with a sort or tracking parameter (?utm_source=, ?orderby=price) doubles the count again. The answer is the canonical tag: each variant points to the product's reference URL, and Google concentrates the signals on a single page. Check that your platform generates it correctly, and that you do not have a canonical pointing to a redirected or 404 page. For out-of-stock products, three cases: a product that will return keeps its page, with an "unavailable" notice and a date; a replaced product is 301-redirected to its successor or its category; a discontinued product with no equivalent can return a 410 once its traffic has faded. Deleting a well-ranked page for a three-week stock-out throws away months of SEO.

How do you optimise a product page for Google?

Product page SEO rests on a title that uses the name customers search for (brand, model, distinctive attribute), a description written for your shop and not copied from the manufacturer, and Product structured data describing price, availability and rating. These three elements are the ones Google reads first and the ones most shops neglect.

The description copied from the supplier is the most widespread flaw: hundreds of retailers publish the same text, and Google has no reason to rank yours over another. One hundred and fifty words written by you, answering the questions your customer service receives about this product (compatibility, true size, care, delivery time), are enough to stand out. Add a questions-and-answers section when the product calls for it, images named after the product with an alt attribute, and a link to the parent category and two or three complementary products.

Structured data completes the work. Product markup with offers (price, currency, availability) makes the page eligible for rich results and Google's merchant listings; aggregateRating and review markup, fed by real customer reviews, adds the stars. The template does most of it here: if your theme or your review app generates this markup correctly, every new product page benefits without intervention. A dedicated article on full product page optimisation will go deeper; for now, remember that the unique description and Product markup are the two actions with the best effort-to-result ratio.

What role do customer reviews play in ecommerce SEO?

Customer reviews act on an ecommerce site's SEO in three measurable ways: they add to each product page unique, renewed text that Google indexes, they feed the AggregateRating markup that shows stars in results, and they give AI assistants concrete elements to quote when a buyer asks for a recommendation. No other content source combines these three effects at such a low cost.

What reviews bring is the customer's vocabulary. Your product page says "breathable waterproof jacket"; your customers write "got caught in a twenty-minute downpour on my bike and stayed dry", "runs a bit large, go a size down", "the zip is solid". These sentences contain the long-tail queries you would never have thought of writing, and every new review adds more. A product page with 40 reviews has content several times richer than with its description alone, and that content changes every week, which is precisely the freshness signal Google looks for on a catalogue.

Volume matters as much as quality. A shop that waits for customers to leave a review spontaneously gets one per hundred to two hundred orders; a shop that asks by SMS or WhatsApp after delivery gets one per two or three. The guide on user-generated content in ecommerce shows how to organise that collection, and the Review collection page describes how Review Collect works.

Reviews: fresh, long-tail content on every product page

For reviews to work for your SEO, they must be published in the product page's HTML, readable by Google without running a script. A review widget that loads comments in JavaScript from a third-party domain, inside an iframe, brings nothing to your page: the content belongs to the platform hosting it. Check it by viewing the page source (not the inspector): if the review texts are not there, Google probably does not see them. Once reviews are in the page, leave them long: ten lines beat a rating alone, and paginating reviews with crawlable links lets the older ones get indexed. Reviews that mention a precise use ("for a 6-month-old baby", "with an Epson printer") are the ones that rank for queries your description does not target.

Why replying to reviews counts too

The merchant's reply adds a second layer of text under each review, written in your words, and it reassures the buyer who reads it: according to an Ifop study (January 2026), 58% of consumers prefer businesses that reply to reviews. On your site, replying to a review that flags a sizing problem lets you clarify the size guide and avoid the next complaint. On Google or Trustpilot, the public reply is read by every prospect comparing reviews before buying, and the pace of replies weighs on the brand's image. The constraint is time: a shop receiving twenty reviews a day cannot reply by hand. Review Collect generates a reply in your brand's tone on every published review, positive and negative alike, in under 60 seconds after publication.

AUTOMATED COLLECTION

Your reviews become SEO content, without writing a line

Review Collect asks every customer for a review after delivery, by SMS, WhatsApp or email, in your brand's name, then directs it to the platform of your choice. Every published review enriches your product pages and your Google, Trustpilot or Avis Vérifiés profiles.

• 40% response rate on average by SMS and WhatsApp
• Up to 30 times more reviews in 30 days, from the first month
• Up and running in 48 hours, no developer needed

Discover Review CollectBook a demo

×30 reviews collected in the first month

How do you show review stars in Google results?

The stars that appear under a Google result come from three distinct mechanisms, with three different sets of conditions: the Product rich result on your product pages, seller ratings on your ads and merchant listings, and aggregated product ratings in Google Shopping. Confusing them is the first reason stars never show up.

The rule many shops discover too late concerns reviews of the business itself. Google states in its documentation that the LocalBusiness and Organization types are not eligible for stars when "the entity being reviewed controls the reviews about itself", that is, when reviews are published on its own site, directly or through a third-party widget, and that ratings must come directly from users. So you cannot get stars on your homepage by marking up your own shop reviews there; you can get them on each product page with real product reviews. The details are in Google's documentation on review snippets.

In practice, it is the review platforms recognised by Google (Google Customer Reviews, Trustpilot, Avis Vérifiés and a few others) that feed the seller rating; your job is to send them a sufficient and regular volume of reviews. The technical guide on activating rich snippets with stars details the markup, and the article on Google Seller Rating covers the conditions on the ads side.

How do you make your shop visible in AI answers (GEO)?

GEO (Generative Engine Optimization) refers to the actions that get ChatGPT, Perplexity, Gemini or Google's AI Overviews to cite your shop when a buyer asks them "which mattress brand is best for back pain" or "where to buy a reliable cargo bike". Google states that its AI features rely on the same ranking systems as classic search; AI search optimisation is therefore not a separate discipline, it is SEO pushed towards more precise and more factual content.

What assistants cite becomes obvious when you read their answers: numbered characteristics (weight, dimensions, battery life, composition), comparisons between models, prices, delivery times and return conditions, and concrete customer feedback. A product page that lines up these elements in short, complete sentences, with specification tables, is far easier to pick up than a paragraph of mood. Category pages with a structured buying guide ("how to choose", criteria, price ranges) are the ones most often picked up for advice queries.

Customer reviews count twice here. First because they provide concrete, dated sentences that models readily reuse ("delivered in two days", "keeps you warm at minus 5 degrees"). Second because buyers check: according to an Ifop study (January 2026), 85% of consumers cross-check an AI's recommendations against other sources, and 61% trust customer reviews first when deciding, against 9% for AI alone. Being recommended by an assistant is worth nothing if your Google or Trustpilot profiles are empty or poorly rated when the buyer goes to check.

Three actions at this stage: make your pages accessible to the assistants' crawlers (check your robots.txt), provide verifiable facts rather than adjectives on every product page, and maintain a volume of recent reviews on the platforms AI tools consult. What Google says has no effect on its systems: creating an llms.txt file, artificially chopping up content or chasing mentions on third-party sites. The guide on getting recommended by ChatGPT and Perplexity in ecommerce goes further, and the SEO, AEO and GEO page shows how Review Collect tracks what AI tools say about your brand.

What are the SEO specifics of Shopify, PrestaShop and WooCommerce?

Each ecommerce platform imposes its own URL structure, canonical handling and performance constraints; the same catalogue does not have the same problems on Shopify, PrestaShop and WooCommerce. The table below sums up what each does well, what traps shops most often, and where to start. These observations reflect each platform's default behaviour; a theme or a module can change it, so check on your own site.

Beyond these differences, the three platforms share one weakness: they produce product pages empty of original content if you do not feed them. Customer reviews are the simplest answer to that weakness, because they integrate on all three (through native markup or an app) and they require no writing from your team. Review Collect connects to Shopify, WooCommerce and PrestaShop to trigger the review request from your orders.

How do you measure your ecommerce SEO progress?

Measure a shop's SEO on four indicators, in this order: revenue attributed to organic traffic, the number of category and product pages receiving at least one click a month, the average positions of priority categories, and indexing coverage (indexed pages over submitted pages). Raw traffic comes last: a rise in visits on pages that do not sell brings nothing.

Search Console provides the last three indicators for free; your analytics tool or your back office provides the first, as long as channel attribution is set up correctly. A monthly dashboard is enough: record the positions of your twenty main categories, the number of product pages that received a click, the ratio of indexed pages to sitemap pages, and organic revenue. Three months without movement on a category you have worked on signals a content or internal linking problem; a sudden drop in indexed pages signals a technical problem (canonical, noindex, robots.txt) to deal with the same day.

Finally, connect that traffic to what it earns. The guide on calculating your conversion rate explains how to isolate the conversion rate of organic traffic; on many shops it beats paid traffic, because a visitor arriving from a precise search already knows what they are looking for.

How Review Collect feeds your ecommerce SEO

Review Collect turns every delivered order into a review request, then into a review published where it counts for your SEO: on your product page, on your Google profile, on Trustpilot or on Avis Vérifiés. The request goes out by SMS, WhatsApp or email, in your brand's name, from Review Collect's Meta Business account, with a configurable sequence and a follow-up channel if the first one gets no reply. On average, 40% of customers contacted by SMS or WhatsApp respond.

Satisfied customers are directed to one platform at a time, the one you choose, with a configurable split: for example 60% to Google for the seller rating, 40% to Trustpilot for the merchant listing. Review Collect is not a review platform; it feeds the ones that already exist and that Google and AI tools consult. Customers who rate 1, 2 or 3 stars land on a feedback page in your colours, and your customer service receives an email alert with the text, the order number and the contact details, to call back before the problem settles in; the customer always keeps the choice to publish their review. On every published review, positive or negative, a reply generated by AI in your brand's tone is proposed in under 60 seconds.

With Review Collect, shops multiply their review volume by 30 in 30 days, from the first month.

For your SEO, the effect shows on three pages at once: product pages receiving fresh text every week, a Google profile and merchant listings reaching the volume needed for stars, and review platforms that are up to date when an AI assistant or a buyer goes to check your reputation. Setup takes 48 hours, no developer needed, on Shopify, WooCommerce or PrestaShop. The Review collection page describes the sequences, and a 20-minute demo is enough to see what your catalogue would produce in review volume.