Technical SEO for eCommerce: The Checklist That Actually Matters

Crawl Budget and Index Bloat

eCommerce sites generate enormous numbers of URLs. Every filter combination, colour variant, size option and sort order can create a unique address. A site with 5,000 products and twenty filterable attributes can spin up millions of URLs, most of them duplicate or near duplicate.

Google gives your site a finite crawl budget. If Googlebot spends its visits crawling thousands of filter pages, it has less capacity left for the category and product pages you actually want indexed. This is not a rare edge case. It is the first thing we look for, and it is usually the biggest win.

The scale of the waste is easy to underestimate. Botify, analysing crawl data across many large sites, found that on unoptimised sites Google crawls only about 40% of known URLs in a given month, which means roughly 60% of pages are not being crawled regularly and may not be indexed at all. On one online marketplace with more than ten million pages, Google ignored 99% of them and crawled just 1%. If your important pages are sitting in that ignored majority, no amount of on-page work will save them.

The fix. Use robots.txt to block crawling of filter parameters and sort orders. Add canonical tags pointing filtered pages back to the main category. For faceted navigation specifically, consider AJAX based filtering that does not create new crawlable URLs at all. Keep your XML sitemap clean, indexable, canonical URLs only, so the sitemap reinforces what matters rather than feeding the bloat.

If crawl waste is eating your index, our technical SEO services start with a crawl budget and log file review.

Site Architecture and Internal Linking

Your architecture should follow one clear hierarchy: homepage, then category pages, then subcategory pages, then product pages. Every page that matters should be reachable within three or four clicks of the homepage.

We audit internal linking by crawl depth, the number of clicks it takes Googlebot to reach each page. If your best selling products are buried six clicks deep behind subcategories and pagination, they will not get the crawl priority or the internal PageRank they deserve. Depth is quietly one of the strongest signals of how important you think a page is, and Googlebot reads it that way.

Breadcrumbs, related product blocks, "customers also bought" widgets and HTML sitemaps all strengthen the internal structure. But the highest impact change is almost always the main navigation and category structure, making sure your highest value pages are linked prominently from the homepage and top categories rather than stranded deep in the tree.

While you are in there, hunt for orphan pages, pages that exist and may be indexed but have no internal links pointing at them. They cannot receive internal PageRank and are effectively invisible to your own structure. Either link to them properly or noindex and redirect them.

Product Page Optimisation

Thin content is the number one on-page issue on eCommerce sites. A product page with nothing but a title, a price and an "add to cart" button gives Google almost nothing to understand or to tell apart from ten thousand similar pages. Those pages do not rank, and the most common cause is manufacturer copy pasted across every retailer selling the same item, which Google sees as duplicate content.

Each product page needs a unique title tag built around the product name and a high intent modifier such as "buy", "shop" or "free delivery". The meta description needs a reason to click, price, a USP or availability. The page itself needs unique product descriptions, technical specifications in a structured format, customer reviews and an FAQ block that answers the questions people ask before they buy. Tighter copy is half the job, see our notes on on-page SEO for the rest.

For variants such as size and colour, use a single canonical URL with variant selectors rather than a separate URL for every combination. That consolidates ranking signals onto one strong page instead of splitting them across dozens of near identical ones.

Structured Data and Rich Results

eCommerce sites should run Product schema on every product page. It is what makes rich results possible, the price, availability, review stars and image showing directly in the search result. A result that takes up more space and carries a star rating earns more clicks than a plain blue link next to it, and the size of that effect is well documented by Google itself.

Google publishes case studies from sites that added structured data. Rotten Tomatoes added it to 100,000 pages and measured a 25% higher click-through rate on those pages versus the ones without. Nestlé found pages shown as rich results had an 82% higher click-through rate than non rich pages. Food Network enabled search features on 80% of its pages and saw a 35% increase in visits. These are not agency estimates, they are Google's own numbers.

At a minimum, implement Product schema (name, price, currency, availability, image, SKU), AggregateRating where you have reviews, BreadcrumbList for your breadcrumbs, and Organization schema site wide. Validate everything with Google's Rich Results Test. The common mistakes are wrong price formats, missing required properties, and markup that does not match what is visible on the page, which Google treats as spam.

Page Speed and Core Web Vitals

eCommerce sites are heavy by nature. Product images, review widgets, recommendation carousels, tracking scripts and payment integrations all add weight. But speed is both a ranking factor and a conversion factor, and on eCommerce the conversion side is where the money is.

The numbers are blunt. Google found that 53% of mobile site visits are abandoned if a page takes longer than three seconds to load. Google's research with SOASTA showed that as load time goes from one second to three, the probability of a bounce rises by 32%, and by 90% as it goes from one second to five. Portent, analysing 94 million pageviews across ten eCommerce sites, found conversion rate drops about 4.42% for every additional second of load time between zero and five seconds.

It works in the other direction too. The Deloitte and Google study "Milliseconds Make Millions" found that a single 0.1 second improvement in mobile load time lifted retail conversions by 8.4% and average order value by 9.2%. For Core Web Vitals specifically, Google found that pages meeting all the thresholds saw visitors 24% less likely to abandon before the page finished loading.

Priority fixes. Serve images in WebP or AVIF with proper srcset for responsive sizing. Lazy load anything below the fold. Defer non-critical JavaScript, your review widget and recommendation carousel do not need to load before the product image and price. Put static assets on a CDN. Cache category pages server side. Largest Contentful Paint is usually the most important Core Web Vital for eCommerce, and the hero product image is normally the LCP element, so preloading it and serving it in an optimised format is often the single biggest win on the speed score.

HTTPS, Canonicals and Redirects

Every eCommerce site should be fully on HTTPS. It has been a ranking signal since 2014 and it is non negotiable for a site taking payments. We still find sites with mixed content warnings, HTTP pages that were never redirected, and canonical tags pointing at HTTP versions of pages. Each of those quietly undermines trust signals and confuses indexing.

Audit your redirect chains. eCommerce sites collect redirects over the years as products are discontinued, categories are reorganised and platforms are migrated. A chain where A redirects to B redirects to C redirects to D wastes crawl budget and bleeds PageRank at every hop. Flatten every chain so each old URL points straight to its final destination in a single step. When the chain comes from a platform migration, fix it as part of the migration plan, not after.

Then check canonicals are doing their job. Filtered and sorted versions of a category should canonical to the clean category URL. Variant pages should canonical to the main product. Paginated pages should reference themselves, not collapse to page one. Canonicals are how you tell Google which version of a near duplicate set actually deserves to rank, and on an eCommerce site there are a lot of near duplicates.

Prioritise by Impact

You cannot fix everything at once, so sequence by impact. Crawl budget and index bloat come first, because they affect the entire site and decide whether your other work is even seen. Product page content and schema come next, because they decide ranking ability on the pages that make money. Speed improvements come last in the sequence, not because they do not matter, they clearly do, but because the gains are usually incremental next to the structural fixes, and a fast page that Google never crawls still earns nothing.

This list is a subset of the 47-point technical SEO audit we run at the start of every engagement.

Product feeds and structured data overlap with paid as well. If Google Shopping is in the mix, the same clean product data feeds both channels.

If you are not sure where to start, we run technical SEO audits built specifically for eCommerce. We crawl the site, find every issue, and hand back a prioritised action plan with the expected impact of each fix. No fluff, just the roadmap and the order to work through it. You can see how we approach it on our eCommerce SEO and technical SEO pages.

JavaScript rendering - the hidden ranking blocker on modern eCommerce stacks

Modern eCommerce platforms (headless Shopify, Next.js commerce builds, Vue Storefront, custom React implementations) increasingly rely on client-side JavaScript rendering for critical content - product data, prices, availability, review counts, faceted navigation. Googlebot renders JavaScript but does so on a delayed schedule (typically 3-14 days after crawl) and with material rendering budget constraints on large sites. Sites relying on client-side rendering for indexation-critical content typically see 20-60% of pages fail to enter the index for weeks or months after publication.

The workable rendering approach for eCommerce at scale: server-side rendering (SSR) or static site generation (SSG) for all product, category and content pages, with client-side interactivity layered on top of pre-rendered HTML. Product prices, availability, canonical tags, structured data, and title/meta tags must all be present in the initial HTML response - not injected via JavaScript after page load. Sites migrating from client-side to server-side rendering typically see 30-80% organic traffic recovery within 60-120 days as previously-unindexed pages become discoverable.

The diagnostic tools that catch rendering issues: Google Search Console URL Inspection tool (compare "HTML" and "Rendered HTML" tabs to spot content that only appears after JS execution), Screaming Frog with JavaScript rendering enabled, and the mobile-friendly test tool for a quick smoke test on individual URLs. Any content visible in "Rendered HTML" but missing from "HTML" is at risk of delayed or failed indexation.

International SEO and hreflang for multi-region eCommerce

Multi-region eCommerce sites face specific technical SEO challenges around content duplication, currency and language switching, and search engine understanding of regional intent. Hreflang tags signal to search engines which language and regional variant of a page should be served to which audience - implemented correctly they materially improve regional targeting; implemented incorrectly they produce widespread indexation issues affecting every regional variant.

The workable hreflang implementation for UK eCommerce with international expansion: hreflang tags in HTML head for all indexed language/region variants including a self-referencing tag and an x-default fallback, bidirectional linking (every variant must reference every other variant, not just one-way references), consistent URL structure across variants (either subdomain, subdirectory, or ccTLD but consistently applied), and geo-targeting configuration in Search Console for each variant. Sites migrating from broken hreflang to correct implementation typically see 15-40% organic traffic recovery on regional variants within 60 days.

Common hreflang mistakes that produce silent traffic loss: missing self-reference (invalidates the entire hreflang cluster), one-way references between variants (Google requires bidirectional confirmation), incorrect language-region codes (en-uk is invalid; en-gb is correct), and hreflang references to non-canonical URLs. The Search Console International Targeting report surfaces most of these issues but requires active monitoring - many teams don't check it monthly and miss issues that compound over quarters.

Log file analysis - the diagnostic technique most eCommerce SEO ignores

Server log file analysis reveals what Googlebot actually crawls versus what SEO tools assume Googlebot crawls. On sites over 5,000 pages the divergence is often material - 20-60% of pages the site owner considers priority receive minimal Googlebot attention, while low-priority URLs (search result pages, parameter variants, admin routes accidentally exposed) absorb crawl budget that could be redirected to commercial pages.

The workable log file analysis cadence for eCommerce over 10,000 URLs: monthly analysis of the last 30 days of access logs filtered to verified Googlebot user agents, focused on identifying (1) high-value pages receiving low crawl frequency, (2) low-value pages absorbing disproportionate crawl budget, (3) 4xx and 5xx errors Googlebot has encountered that need remediation, and (4) redirect chains wasting crawl budget. Sites acting on log file analysis findings typically see 15-35% indexation improvement on previously-under-crawled priority pages within 60-90 days.

Core Web Vitals for eCommerce - the metrics that move rankings and revenue

Core Web Vitals in 2026 are LCP (Largest Contentful Paint, target under 2.5s), INP (Interaction to Next Paint, target under 200ms, replaced FID in March 2024), and CLS (Cumulative Layout Shift, target under 0.1). For eCommerce, the workable targets are stricter than Google's minimums: LCP under 1.8s on 4G mobile, INP under 150ms, CLS under 0.05. Sites hitting stricter targets consistently outperform sites at Google's minimum thresholds on both ranking stability and conversion rate.

The LCP optimisation stack for eCommerce PDPs: hero image served in AVIF or WebP with responsive srcset, preload directive for the LCP image, CDN edge caching with regional POPs, HTTP/2 or HTTP/3 protocol, and server-side rendering of the LCP element (not lazy-hydrated). The INP optimisation stack: minimal third-party JavaScript on critical interaction paths, code-splitting so cart and checkout logic doesn't block product-browse interactions, and web workers for non-critical processing. The CLS optimisation stack: explicit width and height on every image and video, reserved space for above-fold ad slots, and font loading strategies (font-display: optional or swap with matched fallback metrics) preventing layout shift on font swap.

The revenue impact quantified: a Deloitte study of major eCommerce sites found a 0.1s LCP improvement produced measurable conversion rate improvements averaging 8% for retail sites. A separate Google-published case study of a major UK retailer found INP improvement from 320ms to 180ms produced 12% conversion rate uplift on mobile traffic. Core Web Vitals are not a ranking tiebreaker to be optimised for SEO - they're a revenue lever that also happens to affect rankings.

Pagination, crawl budget and infinite scroll for large catalogs

Pagination handling in 2026 has settled on a stable pattern after years of rel=next/prev deprecation confusion. The workable approach: paginated category pages use self-referential canonicals (page 2 canonicalises to page 2, not page 1), each paginated page uses distinct meta titles ("Category - Page 2 of 12"), and product URLs are linked directly from each paginated page (not just page 1). Infinite scroll implementations must include a linked pagination fallback for crawlers - without it, Googlebot only sees the initial page of products regardless of how much content lazy-loads for users.

Crawl budget optimisation matters materially above ~50,000 URLs. The workable stack: robots.txt disallowing faceted URL patterns that produce infinite crawl surfaces, canonicalisation of filter combinations back to primary category URLs, XML sitemap segmentation by product freshness (recently-updated products in a priority sitemap that Google crawls more aggressively), and log-file monitoring of Googlebot behaviour identifying wasted crawl on low-value URL patterns for targeted remediation.

Product structured data - the schema fields that unlock SERP features in 2026

Product schema in 2026 requires more than name, price, and availability to unlock the SERP features that drive CTR. The workable Product schema payload includes: brand (as Brand entity, not string), aggregateRating with reviewCount matching visible reviews, review entities with author and reviewBody, offers with priceValidUntil and priceSpecification, shippingDetails with shippingRate and deliveryTime, hasMerchantReturnPolicy, and productID matching GTIN. Merchants publishing complete Product schema consistently capture rich product results (price, review stars, availability, shipping) that improve organic SERP CTR by 15-35% versus name-and-price-only implementations.

Working with Visionary on technical SEO

Technical SEO engagements at Visionary are delivered directly by Chris - a top proven expert with 15+ years of hands-on commercial experience across Shopify, Magento, headless builds and custom stacks, not junior staff coordinating from templates. Fees between £850 and £2,500/month depending on scope, with performance-linked terms available for proven brands with meaningful organic revenue at stake. Typical engagements start with a 2-week technical audit, followed by a prioritised remediation sprint, transitioning into ongoing monitoring and iterative improvement from month two onwards.

FAQ

Frequently Asked Questions