E-Commerce SEO Guide

E-commerce SEO is the work of making the right catalog pages discoverable without asking search engines to crawl every filter, sort and session URL your platform can generate. It is not a separate marketing channel that you switch on with a plugin. It is how your information architecture, templates and operations show up in search.

This guide is the map. Supporting articles go deeper on each layer. Use it to decide what to fix first, then follow the linked pieces for the mechanics.

What this guide covers

Layer	Job	Start here
Foundations	How store SEO actually works	How e-commerce SEO works
Research	Queries mapped to page types	E-commerce keyword research
Product pages	Buy-intent templates	Product page SEO
Category pages	Browse-intent hubs	Category page SEO
Technical	Crawl, index, canonicals	E-commerce technical SEO
Speed	Core Web Vitals on templates	Core Web Vitals for e-commerce
Links	Crawl paths and equity	Internal linking for online stores
Structured data	Product, breadcrumb, FAQ where true	E-commerce schema markup
Shopify	Hosted constraints	Shopify SEO fundamentals
WooCommerce	Permalinks, plugins, hosting	WooCommerce SEO fundamentals

Platform deep dives live in the Shopify SEO guide and the WooCommerce SEO guide. Product-page conversion and merchandising sit in the e-commerce product page guide.

How e-commerce SEO is different

A brochure site has tens of URLs. A store has hundreds or tens of thousands, many of them near-duplicates. Search systems have a finite crawl budget. If you spend it on ?color=navy&sort=price pages that add no unique value, category and product URLs get crawled later or less often.

Three consequences follow:

1. Page type matters more than word count. A category that explains a buying decision can rank for a head term. A product that repeats the category copy will not.
2. Indexation is a design choice. You decide which URL patterns are canonical, which are noindex, which are blocked in robots.txt, and which exist only as UI state.
3. Operations leak into SEO. Out-of-stock handling, seasonal SKUs, marketplace syndication and app-generated landing pages all create URLs. If nobody owns them, they rot in the index.

Read how e-commerce SEO works for the crawl–index–rank loop applied to catalogs. The rest of this guide assumes that model.

Information architecture

Information architecture (IA) is the tree shoppers and crawlers share. A workable store IA usually looks like this:

• Home
• A small set of department or collection hubs
• Category / collection pages that match how people search
• Product pages
• Supporting pages: about, shipping, returns, size guides, editorial

The failure mode is a flat dump of products, or a tree that matches your warehouse rather than search language. “Men / Footwear / Running” is useful if people search that way. “Season 24 / Bin A3” is not.

Design rules that keep IA searchable

• One primary URL per product. Variants (size, color) should not each be a full indexable page unless the variant is a genuinely different product with its own demand (for example a distinct model name).
• Categories match query clusters, not SKU counts. A category with three products can still be the right page if the query is real and the page helps the shopper choose. A category with 400 products and no filters, copy or merchandising is a dump.
• Do not create a new collection for every campaign. Sale and “new in” collections are fine as navigational tools. They become a problem when they duplicate the same products under thin, indexable URLs.
• Keep URL paths stable. Changing a slug without a redirect throws away accumulated signals. Draft paths with the URL slug generator before you publish.

Breadcrumbs should reflect the primary path, not every possible path. If a product sits in two collections, pick a canonical parent for the breadcrumb and the canonical tag.

Keyword research for a catalog

Keyword research for stores is assignment work: which page type owns which query. It is not a list of phrases to paste into every meta title.

Query types you will actually use

Query type	Example intent	Default page
Category / class	“wireless headphones”	Collection or category
Product / model	“sony wh-1000xm5”	Product
Attribute + class	“noise cancelling earbuds under 5000”	Filtered category or a dedicated landing page if the combination has demand and unique copy
Problem / use	“headphones for commuting”	Guide or curated collection with real editorial
Brand	“sony headphones”	Brand collection
Transactional modifiers	“buy”, “price”, “in stock”	Usually the same product or category; do not clone pages for modifiers

Work top-down. Start with the categories you will stock for the next year, not with a keyword tool export. Then check demand, overlap and current ranking URLs. The method is in e-commerce keyword research.

Cannibalization is an IA bug

If two collections, a blog post and a product all target “running shoes”, search engines guess. Fix the guess:

• The category owns the class term.
• The product owns the model or SKU term.
• An article owns the “how to choose” term and links to the category and a short list of products.

Do not write a 2,000-word product description to “win” a category query. Change which URL you ask to rank.

Product page SEO

A product page has two jobs that must not fight: rank for the product people search, and help someone buy. SEO copy that hides specs below a keyword essay fails the second job. Conversion copy that never names the product fails the first.

Minimum on-page set:

• Unique title and H1 that lead with the product, not a slogan
• Short, accurate description of what it is and who it is for
• Specs in HTML, not only in an image
• Variants visible without a separate URL unless required
• Images with descriptive filenames and alt text that describe the image, not a keyword list
• Availability, shipping and returns in text, not only in a popup
• Internal links to the parent category, complementary products and any genuine size or care guide

Draft titles and descriptions with the SERP snippet preview. Use the keyword density checker only to catch stuffing — there is no target percentage.

Full template and pitfalls: product page SEO and the product page guide.

Thin and duplicate products

Manufacturer copy syndicated to every retailer is a duplicate problem you cannot “unique” with a synonym spinner. Add what only you know: pack contents, fit notes, compatibility, shipping reality, what is in the box, and how returns work. If two SKUs differ only by a letter in the SKU, consolidate or canonicalise.

Out of stock is not automatically a delete. If demand remains, keep the URL, state availability honestly, and offer a substitute or waitlist. If the product is gone for good, 301 to the closest useful category or successor — not to home.

Category page SEO

Category pages are where most stores under-invest. They are the natural home for class queries, and they are often a grid of products with a two-line intro copied from the brand manifesto.

A category that can rank usually has:

• A title and H1 aligned to the query cluster
• A short opener that defines the range (what is in, what is not)
• Filters that do not spawn thousands of indexable URLs
• Unique introductory copy where it helps a shopper choose (materials, sizing, regulations) — not a 900-word essay above the fold
• Merchandising: featured products, comparison cues, links to size guides
• Pagination that search engines can follow, with a consistent canonical strategy

Details and filter strategy: category page SEO.

Faceted navigation

Facets are good for shoppers and dangerous for indexes. A practical default:

• Index the plain category and a small set of facet combinations that match real queries and have enough products and unique copy.
• noindex, follow or parameter handling for the rest.
• Never block CSS/JS in robots.txt in an attempt to hide facets.
• Do not let sort orders (?sort=price) into the index.

This is technical SEO as much as copy. Pair it with e-commerce technical SEO.

Technical SEO for stores

Technical SEO on a store is mostly URL hygiene.

Inventory of URL types

Walk your own site as a crawler would:

Pattern	Typical decision
Product and category templates	Index, unique canonical
Tag / vendor / vendor+tag combos	Often noindex unless they are real landing pages
Internal search results	noindex
Cart, checkout, account	Disallow in robots.txt
Filter and sort parameters	Canonical to clean URL or noindex
Print, quick-view, sharing URLs	Canonical to product
Paginated series	Self-canonical or rel strategy applied consistently

Generate a conservative robots.txt with the robots.txt generator. Do not disallow theme or asset folders. Modern crawlers need CSS and JavaScript to understand the page.

Sitemaps should list indexable URLs only. Including cart, filters and noindex paths teaches tools nothing useful.

Canonicals, hreflang, HTTPS

Canonical tags must point at the URL you want indexed, including the preferred host and protocol. Self-referencing canonicals on products and categories are the default. Do not point every product at the homepage.

International SEO is a separate project: one locale per URL, consistent hreflang, and a real translation or market difference — not a currency switcher that keeps the same English indexable URL. Shopify and WooCommerce handle this differently; see the platform guides rather than bolting hreflang onto a single-store setup.

Step-by-step crawl diagnostics live in e-commerce technical SEO.

Internal linking

Stores leak PageRank and crawl paths into footer tag clouds and “recently viewed” modules. Useful linking is boring and structural:

• Every product links to its primary category.
• Categories link to child categories and a curated set of products (not only whatever the sort order shows today).
• Size guides, shipping and returns are linked from PDPs where they answer a real question.
• Editorial content links to categories with descriptive anchors, not “click here”.
• Related products should be merchandised, not a random-walk of the catalog.

A crawler that cannot reach a product without a sitemap has a discovery problem. Sitemaps are a hint, not a substitute for links.

See internal linking for online stores.

Schema markup

Structured data does not rank pages by itself. It can make eligible rich results possible when the page already qualifies, and it can help systems understand product identity.

For a typical store, the useful types are:

• Organization on the site
• BreadcrumbList on category and product templates
• Product (with Offer, availability, currency) on PDPs — values must match the visible page
• FAQPage only when the FAQs are visible, unique and not stuffed

Do not mark up reviews you do not display. Do not invent aggregate ratings. Do not copy a competitor’s JSON-LD. Invalid or misleading schema is worse than none.

Implementation notes: e-commerce schema markup.

Core Web Vitals

Lab scores on an empty homepage are not the metric. Field data on product and category templates is, because that is where organic landing happens.

Typical storefront weight comes from:

• Hero and product image galleries without modern formats or dimensions
• App scripts (reviews, live chat, personalization, pixels)
• Theme sections that load on every template
• Third-party fonts and video embeds

Fix the template, then the apps, then the extras. A faster PDP also helps conversion; treat speed as shared work with the merchandising team.

Read Core Web Vitals for e-commerce.

Platform constraints

Hosted and self-hosted stacks fail in different ways.

Shopify gives you a solid baseline (HTTPS, CDN, checkout) and less control over URL structure, some duplicate collection routes, and app script load. Work inside those rails: theme.liquid discipline, metafields for unique copy, careful app audit. Start with Shopify SEO fundamentals and the Shopify SEO guide.

WooCommerce gives you permalinks, robots, hosting and plugins — which means you can build a clean index or a slow, plugin-heavy duplicate farm. Hosting and a restrained plugin list are SEO decisions. Start with WooCommerce SEO fundamentals and the WooCommerce SEO guide.

Neither platform “does SEO for you”. Both can rank. The work is still IA, templates and measurement.

Measurement without folklore

You cannot manage e-commerce SEO with a single keyword rank tracker. Use a small set of views:

1. Index coverage — which URL patterns are indexed, excluded, or erroring.
2. Query → URL — do class queries land on categories and model queries on products?
3. Landing page organic revenue — contribution after returns if you can join the data, not sessions alone.
4. Template performance — Core Web Vitals by page type.
5. Crawl stats — are you spending crawl on filters?

Rankings move. Coverage and mapping tell you whether the system is even pointed at the right pages.

Pair SEO reporting with the e-commerce analytics guide so organic is judged with the same conversion and AOV language as paid.

A practical sequence

If you are starting from a live store, this order wastes less time than “write more product copy”:

1. Inventory URLs. Crawl or export. Group by pattern. Decide index / noindex / redirect / delete.
2. Fix robots, sitemap and canonicals so the index can only grow in the directions you want.
3. Lock IA. Primary category tree, product-to-category links, breadcrumbs.
4. Assign queries to existing templates. Create new landing pages only when no current URL can own the cluster.
5. Upgrade category templates for the terms that already have impressions but weak copy or weak merchandising.
6. Upgrade product templates for models and SKUs with demand.
7. Cut app and image weight on those templates.
8. Add schema that matches the visible page.
9. Review monthly: coverage, query mapping, organic revenue, CWV.

Tools for drafts: SERP snippet preview, URL slug generator, robots.txt generator, keyword density checker.

What this guide does not claim

Search visibility depends on competition, demand, site history, and whether the catalog deserves the query. This guide does not promise rankings, traffic, or revenue. It does not treat AdSense or any ads product as a store strategy. Do the architecture work, measure coverage and contribution, and iterate on templates — that is the job.