E-commerce SEO

E-commerce is technical SEO on hard mode: thousands of URLs, faceted navigation minting millions more, inventory that appears and vanishes, and the most commercially contested SERPs on the web. It is also where SEO connects most directly to revenue. In 2026 there is a second layer on top: AI Mode, ChatGPT and agents now compare and sometimes buy products for people, and they rely on your product data at least as much as your pages. This guide assembles the curriculum's pieces into a store-shaped strategy.

The page-type hierarchy#

A store ranks through three page types, each with its own intent slot:

The strategic insight most stores miss: category pages are your SEO workhorses. They target the high-volume commercial heads, and they are where internal authority should concentrate.

Category pages#

• One category per meaningful search demand. Your keyword research decides the taxonomy, not just merchandising logic.
• Add real content: a short buying-guide intro and, where shoppers really ask them, a few questions and answers. That gives engines text to rank and gives AI answers something to quote. Do not add FAQPage markup expecting a rich result: Google stopped showing FAQ rich results on 2026-05-07.
• Crawlable product links: real <a href> grids, paginated with self-canonical pages, never JS-only infinite scroll (site architecture).
• Subcategory landing pages for demanded facet combos ("waterproof trail running shoes"), and only those. Everything else stays out of the index (see faceted navigation below).

Product pages#

• Unique descriptions. Manufacturer copy is duplicated across every retailer selling the item, so rewriting it is the cheapest differentiation available. Prioritize by margin and traffic. If you use AI to draft at scale, Google's guidance (updated 2026-10-01) says to fact-check all of it before publishing, and Merchant Center requires AI-generated titles and descriptions to be labeled. See AI-Assisted Content Without Penalties.
• Reviews on-page. User reviews are unique text, long-tail coverage and conversion proof in one. Since 2026-07-24 Google's review snippet guidelines explicitly prohibit fake and undisclosed incentivized reviews.
• Variant handling: color and size variants on one canonical URL beat indexable near-duplicates per variant, unless a variant has its own search demand ("pegasus 41 white"). Google supports ProductGroup markup with hasVariant for this.

Merchant listing vs product snippet markup#

Google splits Product structured data into two experiences, and picking the wrong one is a common mistake:

A merchant listing for a store page looks like this:

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Product",
  "name": "Pegasus 41 Trail",
  "gtin13": "1234567890123",
  "brand": { "@type": "Brand", "name": "Nike" },
  "category": "Apparel & Accessories > Shoes",
  "aggregateRating": { "@type": "AggregateRating", "ratingValue": 4.6, "reviewCount": 218 },
  "offers": {
    "@type": "Offer",
    "price": 139.99,
    "priceCurrency": "EUR",
    "availability": "https://schema.org/InStock",
    "priceSpecification": [
      {
        "@type": "UnitPriceSpecification",
        "priceType": "https://schema.org/StrikethroughPrice",
        "price": 159.99,
        "priceCurrency": "EUR"
      },
      {
        "@type": "UnitPriceSpecification",
        "price": 124.99,
        "priceCurrency": "EUR",
        "validForMemberTier": { "@type": "MemberProgramTier", "@id": "https://example.com/members#gold" }
      }
    ]
  }
}
</script>

What changed in 2025-2026 (all from Google's Search Central changelog):

• Member pricing: validForMemberTier on a UnitPriceSpecification (February 2025, documented as beta) shows a loyalty price.
• Loyalty programs: MemberProgram with MemberProgramTier tiers, nested under your Organization or OnlineStore (June 2025). Available in Australia, Brazil, Canada, France, Germany, Mexico, the UK and the US.
• Shipping and returns at the organization level: Google now prefers one store-wide MerchantReturnPolicy (with the required returnPolicyCountry) and a store-wide ShippingService (November 2025) under Organization. Offer-level shippingDetails and hasMerchantReturnPolicy still work and override the defaults for specific products.
• Category and sale dates: Product.category accepts text or a category code that matches Merchant Center (July 2026), and Google documents validFrom, validThrough and priceValidUntil for sale prices.

Validate every template with the Rich Results Test. The structured data guide covers the rest.

Out-of-stock and discontinued#

Temporarily out      → keep 200, mark OutOfStock in schema, suggest alternatives
Discontinued, has    → 301 to the successor product
  a successor
Discontinued, none   → 410 (or 301 to category if the page has links/traffic)
Seasonal             → keep the URL alive year-round; it re-ranks each season

Mass-404ing products is the classic self-inflicted e-commerce wound: equity that took years to build, deleted with one inventory sync.

Faceted navigation, revisited#

Stores live or die on crawl-trap containment. Google's faceted navigation guidance moved to its crawling documentation site in December 2025, and the advice is the same: block filter combinations you never want crawled, keep the ones with demand clean and linkable. The e-commerce specifics:

• Parameter order must be canonicalized (?color=red&size=m and ?size=m&color=red resolve to one URL).
• Sort and view parameters are never indexable.
• Run log-file checks quarterly (audit playbook) to confirm Googlebot is not drowning in filter URLs.

Merchant Center: your second index#

Your product feed is no longer a side channel for ads. Google's own AI optimization guide (May 2026) names Merchant Center feeds and Business Profiles as the way to make products visible "in both AI responses and other Google Search results". The feed feeds the Shopping tab, free product listings, product panels, Lens and the product results inside AI Mode.

• Keep feed and page in sync. Price, availability and GTIN in the feed must match the landing page and its markup. Mismatches cause disapprovals, and they confuse every surface that reads both.
• Fill the optional attributes. Product highlights, detailed sizes and materials, shipping and return settings: AI answers compare products on exactly these facts.
• Label AI-generated content. Merchant Center requires AI-generated images to carry IPTC DigitalSourceType metadata (TrainedAlgorithmicMedia) and AI-written titles and descriptions to be labeled.
• Business Agent, launched in January 2026 and managed in Merchant Center, lets shoppers chat with your brand on Search. Consider it once the basics are clean.

Agentic commerce: what is real in 2026#

AI agents now research, compare and in some cases buy on a user's behalf. Three protocols define how:

The market is still settling. In March 2026 OpenAI said "Instant Checkout is moving to Apps": purchases in ChatGPT now mostly hand off to retailer apps or the retailer's own site, after only a small number of merchants went live (reported by Search Engine Land and others). Google's agentic checkout through UCP is live but limited to eligible US merchants.

What to do as a store in 2026:

1. Product data first. Clean feeds, complete merchant listing markup and accurate stock are the inputs every agent and AI answer uses. This is work you need for Shopping anyway.
2. Join programs through your platform. Shopify and the large platforms handle much of the protocol work. Check what your platform supports before building anything custom.
3. Keep checkout on your site agent-friendly. Browser agents read the DOM and the accessibility tree, so labelled form fields, real buttons and no surprise interstitials help them complete a purchase.
4. Do not block the bots you want. Shopping and AI search crawlers blocked at the CDN cannot see your prices. Check the AI Crawlers & Bot Access guide.

Beyond the organic SERP#

• Product review content ("best trail shoes 2026") is judged on first-hand testing evidence and original insight (E-E-A-T). Self-promotional "best X" lists that rank your own products first drew scrutiny in 2026.
• AI shopping answers recommend specific products. The GEO playbook (entity clarity, review presence, comparison-page inclusion) and SEO for AI Mode & AI Overviews apply with money attached.
• Visual search: Lens and Circle to Search lean on product images and merchant data. Search Console added a multimodal search filter in September 2026 so you can measure it.

E-commerce ties together nearly every guide in this curriculum: if you can run a store's SEO, you can run anything's. Next: Analytics & Forecasting.