E-Commerce Technical SEO

Technical SEO for e-commerce is making the catalog graph honest: one preferred URL per product and collection you care about, a crawl path that reaches them, and a sitemap that matches. Copy and keyword mapping still matter; they cannot rank URLs the engine never selects.

The conceptual loop is in How E-Commerce SEO Works. This article is the plumbing.

Duplicates: where they come from

Stores manufacture duplicates by default:

• HTTP vs HTTPS, www vs apex, trailing slash vs none.
• The same product at /products/x and /collections/sale/products/x.
• Variants as separate URLs with identical copy.
• Session IDs, utm on internal links, sort parameters.
• Printer-friendly and “quick view” URLs.
• Staging and the live host both reachable.

Pick a host and path convention and 301 everything else. Then pick a product URL pattern and canonical the rest to it. Shopify’s collection-product paths are a classic example—see Shopify SEO Fundamentals. WooCommerce permalinks and plugin extras are another—see WooCommerce SEO Fundamentals.

Canonicals

rel=canonical should point to the URL you put in the sitemap and in internal links.

Rules that stay out of trouble:

• Self-canonical on the preferred PDP and category page 1.
• Variant URLs you do not want indexed canonical to the parent (or use noindex if they should not consolidate).
• Filter URLs canonical to the clean category unless you intentionally index that facet (category page SEO).
• Do not canonical all products to the homepage. That is how you hide the catalog.

Canonicals do not replace redirects for URLs you have deleted. Use 301s when a SKU is gone for good.

Robots.txt and meta robots

Use robots.txt to keep crawlers out of:

• Cart, checkout, account, wishlist (unless you have a reason).
• Internal site-search result URLs (/search?q=).
• Facet systems you have decided are uncrawlable and that are not the only path to products.

Do not block CSS/JS the renderer needs if you expect modern crawlers to see the designed page. Do not block collection paths that are the only HTML links to new SKUs.

noindex is for URLs that may be crawled but should not appear in results (thank-you pages, pure sorts, thin duplicates). noindex in robots.txt is not the same as a meta robots noindex—crawlers that are disallowed never see the meta tag.

Sitemaps

An XML sitemap is a list of candidates, not a ranking switch. Include:

• Canonical product URLs in stock or that you still want found (policy choice for OOS).
• Canonical collection URLs (page 1).
• Guides and other indexable articles.

Exclude noindexed URLs, filtered faceted URLs, and cart. Lastmod should reflect real changes or omit it. Generate URL lists with care using something like the sitemap URL generator for small sets; large catalogs need an automated feed from the live canonicals.

Submit in Search Console and watch coverage: “Duplicate, Google chose different canonical than user” is a signal your hints and links disagree.

JavaScript rendering

Themes and headless storefronts often render product grids in the client. If the initial HTML is an empty <div id="root">, some crawlers will eventually render it—and some tools and some bots will not.

Verify:

• Product and collection links exist in view-source or a fetch-as-Google equivalent, not only after clicking “Load more.”
• Pagination or a crawlable “page 2” exists.
• Critical content (title, price, availability) is in the HTML or reliably rendered.

Headless commerce makes this your problem explicitly. Coupled platforms still fail it with app-laden themes.

Parameter URLs

Treat parameters as a list:

Parameter	Usual handling
sort, orderby	Canonical to clean URL; noindex if needed
page	Index as a series or canonical policy—be consistent
color / size if not unique pages	Canonical to parent PDP/category
utm_* on internal links	Stop adding them internally; 301 or canonical if they leak
Session / sid	Never index; fix the app that appends them

Search Console URL Parameters (where available) is a last resort. Prefer not generating the URLs.

Page experience

Core Web Vitals for e-commerce is the catalog version: LCP on hero images, INP on filters, CLS on injected apps. Technical SEO includes not shipping three tag managers on the PDP.

Schema is adjacent: E-Commerce Schema Markup. It does not fix duplicates.

Hreflang, staging, and mixed signals

If you sell in multiple languages or domains, each locale needs a self-canonical and a coherent hreflang set. Do not hreflang a thin auto-translated collection to a rich original. Staging and preview hosts must be noindexed and not in the production sitemap.

Google Search Console “Duplicate without user-selected canonical” usually means internal links, ads, or the sitemap still point at the URL you thought you demoted. Fix the links; do not add a second canonical plugin.

HTTP status: soft 404s (200 on empty collections) waste crawl. Use 404 or 410 for gone SKUs you will not replace; 301 when there is a clear successor.

Pagination parameters must not create an indexable duplicate of page 1 with a different canonical. Be consistent: either ?page=2 is a series you want crawled, or it canonicals to page 1 (usually the wrong choice if products on page 2 are otherwise orphan).

Work in this order on a messy store: host/redirects → product URL canonicals → robots/sitemap agreement → renderability of grids → then copy. The e-commerce SEO guide sequences that with keyword and content work.