Ecommerce SEO that works on a million pages at once.

A large store is a few templates repeated across millions of URLs. We work on those templates, the product feed, and the filter URLs that waste Google's crawling, to help categories rank and get your products into shopping results and AI answers.

Get a free AI visibility audit All site types

By Robert McCrayUpdated 17 minute read

Key takeaways

  1. Online keeps taking share. Ecommerce was 17.1% of US retail sales in the second quarter of 2026, up from 14.2% in the same quarter of 2022.1

  2. Filters are the classic trap. Google says filter URLs can create infinite URL spaces, which leads to overcrawling and slower discovery of new pages.6

  3. Shoppers often choose before they reach your site. Google says free listings can show your products on the Shopping tab, in Search, Images, Lens, and YouTube, and in Gemini.12

  4. AI tools now send shoppers. Adobe, which sells analytics software, measured a 693.4% rise in traffic from generative AI tools to US retail sites in the 2025 holiday season, against the year before.2

  5. Product data and crawl fixes usually show within weeks. Changes to category and product templates usually take three to six months, because each one is tested before it rolls out.

The basics

What is ecommerce SEO?

Ecommerce SEO means making a large store's categories and products easy for search engines and AI tools to find, read, and recommend. It covers Google's regular results and the free product listings built from the data you send to Google Merchant Center. It also covers AI answers, shopping assistants, and the review sites and forums shoppers check before they buy.

Three things make it different

  1. You fix templates, not pages

    A million-URL store runs on a handful of templates. Change the product template and every product page changes with it. That scale is also why crawling needs managing. Google aims its guide on the subject at sites with roughly a million or more pages that change weekly, or 10,000 or more that change daily.7

  2. Filters can multiply URLs without limit

    Faceted navigation is the set of filters on a category page: size, color, price, brand. When each choice changes the URL, the combinations never end. Google says this can create infinite URL spaces, which causes overcrawling and slows the discovery of new pages.6

  3. The shopper can choose before they visit

    In May 2025 Google said its Shopping Graph held more than 50 billion product listings, each with details like price, reviews, and availability. More than 2 billion of them are refreshed every hour.4 Your product data competes there before your page ever loads.

  • 693.4%

    Growth in traffic to US retail sites from generative AI tools, 2025 holiday season2

    Against the 2024 season. Measured by Adobe Analytics

  • 50B+

    Product listings in Google's Shopping Graph, as of May 20254

    Google says more than 2 billion are refreshed every hour

  • 3.2%

    Of desktop searches in Ahrefs' shopping category showed an AI Overview, September 202520

    Ahrefs found one on 20.5% of all 146 million results pages it checked

  • 17.1%

    Ecommerce share of total US retail sales, second quarter of 20261

    Seasonally adjusted. It was 14.2% in the second quarter of 2022

    • 14.6% in 2021
    • 14.2% in 2022
    • 15.3% in 2023
    • 16.1% in 2024
    • 16.3% in 2025
    • 17.1% in 2026

How customers search

How shoppers search for products

Shoppers move between Google, marketplaces, review sites, and AI tools before they buy. Each kind of search is decided in a different place.

What they searchWhat they wantWhere it's decidedWhat wins it
women's waterproof hiking boots A shelf to browse OrganicShopping A category page with real stock, filters, and buying help
best hiking boots for wide feet A short list, with reasons AI answerReview sites A buying guide, real reviews, and review-site coverage
exact product name + size or color That item, in stock, at a fair price ShoppingMarketplaces A complete feed and a price that matches the page
leather vs synthetic hiking boots Help choosing between two options AI answerForums A plain comparison with honest trade-offs
black friday hiking boot deals This year's prices Organic One sale page that keeps its URL every year
your brand + return policy To check the terms before paying Your name A clear returns page and matching policy markup

Example searches. Part of our enterprise SEO work.

Shopping results

Win the shelf Google builds from your data

There are two ways to give Google your product data: a feed sent to Merchant Center and structured data on your product pages. Google recommends doing both.10 New Merchant Center accounts are opted in to free listings by default.12

  1. Feed

    Does Google have every product, with every detail?

    A feed is the file of product data you send to Merchant Center: IDs, titles, images, prices, stock. Google notes that crawling isn't guaranteed to find every product on a site, and that a feed can update as often as hourly.10

  2. Markup

    Does the page state the same facts in code?

    Product structured data is code that gives the name, price, and stock in a form machines can read. Google recommends putting it in the initial HTML, because markup added later by JavaScript can make its Shopping crawls less frequent and less reliable.9

  3. Match

    Do the feed, the page, and the checkout agree?

    Google asks that the price in your feed match the landing page and the checkout.13 If the page says out of stock while the feed says in stock, or the reverse, Google disapproves the product.14

Pages

Build the store around its category pages

Someone who types "women's waterproof hiking boots" wants a shelf to browse, not one product. That search lands on a category page. Each category needs real content, plain links to its products, and an address that stays the same from one year to the next.

One category, built the way shoppers search
Site planEcommerce and retail
Hiking bootsHub
  • Women's hiking bootsCited
  • Men's hiking bootsCited
  • Waterproof hiking bootsPage 1
  • Wide-fit hiking bootsPage 1
  • Lightweight hiking bootsDrafting
  • Hiking boot buying guideDrafting
  • Kids' hiking bootsPlanned
  • Hiking boot salePlanned

One main category, with subcategories, a few filter pages worth indexing, a buying guide, and a sale page under it.

What goes on a category page

  1. A plain heading and a two-line introWhat's on this shelf and who it's for, written so an AI tool can quote it whole.
  2. Products in the first HTMLThe product grid's links sit in the page source. Google's crawlers don't click "load more."11
  3. Links down the treeMenus link to categories, categories to subcategories, and those to every product.
  4. Pagination with real linksGoogle advises giving each results page its own URL and its own canonical (main version) tag.11
  5. Filters kept out of the crawlSort orders and most filter combinations are blocked. A chosen few become real pages.
  6. Buying help under the gridFit, materials, and sizing in short answers from someone who knows the products.
  7. Breadcrumbs and list markupSo search engines can see where the page sits in the store.
  8. One address, kept every yearSeasonal pages reuse the same URL, so links and history carry over.

Filter pages, done right

A few filter combinations are real searches, like waterproof boots in wide sizes. Most are noise. For the ones you want indexed, Google advises the standard & separator and a fixed filter order in URL paths. An empty combination should return a 404, the page-not-found code.6

Make it a real page when

  • Shoppers search for that exact combination
  • It has enough products to be worth the click
  • It has its own title, heading, and fixed URL

Keep it out when

  • It's only a sort order or a price range
  • The same filters in another order make a new URL
  • The combination returns no products

Content

Answer what shoppers ask before they buy

Shoppers ask these on Google, in AI tools, and on the product page itself. A clear answer keeps the sale and hands AI tools a line worth repeating, with your store's name attached.

  • Question 01

    What will shipping cost, and when will it arrive?

    A good answer covers

    The cost and the delivery window, shown before checkout. In Baymard's survey of US online shoppers, leaving out people who were only browsing, 40% of cart abandoners cited extra costs like shipping, tax, and fees.5

  • Question 02

    Can I return it, and who pays?

    A good answer covers

    The return window, the fee, and how to send it back, on the product page. Google's merchant listing markup can carry return and shipping details into the results.9

  • Question 03

    Which size or model fits me?

    A good answer covers

    A size chart or fit guide on the product page, with what buyers say about fit. For parts and accessories, a plain list of what each one works with.

  • Question 04

    Is it in stock in my size?

    A good answer covers

    Live stock for each size and color. If it's sold out, say when it's due back and show the closest match.

  • Question 05

    Which is better, this one or that one?

    A good answer covers

    An honest side-by-side of two products you sell: who each one suits, and what you give up with the cheaper one.

  • Question 06

    Is this store safe to buy from?

    A good answer covers

    Real reviews, a clear returns page, and contact details that are easy to find. In the same Baymard survey, 19% of cart abandoners said they didn't trust the site with their card details.5

Each answer sits where the question comes up: on the product page, the category page, or a guide. Someone who knows the products checks it before it goes live.

AI search

Be the store AI tools send shoppers to

Shoppers ask AI tools what to buy, and they click through. Adobe, which sells analytics software, measured a 693.4% rise in traffic from generative AI tools to US retail sites in the 2025 holiday season, against the year before.2 It also found that AI referrals converted 31% more often than other traffic that season.3

  1. Keep product data complete

    Google says free listings can show in Gemini, its AI assistant.12 Shopping in Google's AI Mode pairs Gemini with the Shopping Graph.4

  2. State product facts in plain text

    Size, materials, fit, and what it works with, written out on the page. Google lists having important content in text form among the basics for its AI features.18

  3. Earn reviews where AI reads

    OpenAI says ChatGPT writes its review summaries from reviews on public websites.19 Real reviews on public pages are worth earning.

  4. Check the answers every month

    Each month we put the same 40 shopper questions to the main AI tools and log which stores and products come back.

Google says its AI features need no special files or markup.18 AI Overviews, its AI-written summaries, are a small part of this. In Ahrefs' September 2025 desktop data, they showed on 3.2% of results pages for shopping searches, against 20.5% overall.20

Foundation

Reviews, links, and a site Google can crawl

A store's reviews, its links, and its crawl health each need an owner and a backlog. None of them is a one-time project.

Reviews

Reviews have to be real

Stars in search results come from review markup. Google's guidelines say not to put fake reviews on the page or in that markup.15 The FTC's 2024 rule bans fake reviews too.16

  • Ask every buyer for a review, not only the happy ones
  • Never reward a review on condition that it's positive or negative. The FTC rule bans that16
  • Mark up only reviews that shoppers can plainly see on that page15
  • Don't pull in ratings from other websites. Google's rules bar it15
  • If a reviewer got a free product or a discount, disclose that clearly15
Authority

Links you earn, not links you trade for

For a store, the easiest links are the risky ones. Google counts sending someone a product in return for a write-up with a link as link spam, unless the link is marked as sponsored or nofollow.17

  • Gift guides and product reviews in real publications
  • Where-to-buy links from the brands you stock
  • Affiliate and gifted-product links tagged as sponsored, so Google knows they were paid for
  • Buying guides good enough that forums link to them
  • Nothing bought in bulk from a link seller
Technical

A site Google can crawl without waste

Crawl budget is Google's term for the URLs it can and wants to fetch from a site. If its crawlers spend too much time on URLs they shouldn't be crawling, they may not get to the rest of your site.7

  • Sort and filter URLs blocked in robots.txt, the file that tells crawlers what to skip
  • Sitemaps, the URL lists you hand Google, with only live pages and a last-changed date
  • Products removed for good return a 404 or 410, the page-gone codes. A noindex tag only hides a page and still costs a crawl7
  • Product markup in the first HTML, not added by script9
  • Fast category pages on phones. Adobe says 56.4% of US online transactions came through a smartphone in the 2025 holiday season2

Budget

Paid media, in-house, or a partner?

Most large retailers need all three. Paid media rents the shelf. The in-house team keeps the site healthy. The open question is who has the hours to test and ship template changes.

Paid search and retail mediaIn-house team aloneSEO and AI search
What you pay forEach click or impression, on Google or on another retailer's siteSalaries, plus a place in the engineering queueA fixed scope of work, not a fee per click
How soon it showsDaysAs fast as tickets clear the queueWeeks for feed and crawl fixes, months for templates
What limits itBudget. The placements stop when it doesHours. SEO competes with every other ticketYour release cycle, which we work inside
What's left if you stopNothingEverything, if the people stayFixed templates and pages that go on earning

Free listings can appear on the Shopping tab at no cost, while Shopping ads keep the ad slots there.12 A clean feed is worth having whether or not you buy ads.

Timeline

What to expect, and when

Timelines depend on how fast changes clear your release process. A store that ships every week sees results sooner than one with a long code freeze before the holidays.

  1. Weeks 1 to 4

    Find the waste

    • Audit of what Google crawls and indexes, by page type
    • Merchant Center feed and markup review
    • Baseline of AI answers for 40 shopper questions

    Feed errors are usually the first thing fixed.

  2. Months 2 to 3

    Fix the crawl and the feed

    • Sort, filter, and site search URLs brought under control
    • Feed disapprovals and mismatches cleared
    • First template tests scoped with your engineers

    Crawling usually shifts toward pages that can sell.

  3. Months 4 to 6

    Test the templates

    • Category page changes tested against a control group
    • Product page content and markup rolled out
    • First monthly report on AI answers

    Winning tests roll out to the full template.

  4. Months 7 to 12

    Build what's missing

    • Buying guides and comparison pages
    • Filter pages promoted where demand is proven
    • Seasonal pages set up on permanent URLs

    Watch revenue from searches without your store's name.

Reporting

Measure revenue by page type, not rankings

A ranking doesn't pay for inventory. We report organic revenue and visits for each page type, split out non-brand searches, the ones without your store's name, and tie every template test to the sales it moved.

See how we report

What goes in your monthly report

  1. Non-brand organic revenueBy page type: category, product, guide
  2. Free listing clicksFrom Merchant Center, by category and brand
  3. Crawl spent on real pagesFrom server logs: Google's requests that hit pages you want indexed
  4. Index health by page typePages indexed, and pages Google found but hasn't crawled yet
  5. Template test resultsEach test with its control group, its result, and its rollout date
  6. AI citations and referralsTimes AI tools name your store, plus the visits and orders they send

What to avoid

Six mistakes we see on large online stores

  • Letting every filter make a URL

    Each filter choice adds another address, and most are near copies of a page you already have. Google says filter URLs can lead to overcrawling and slower discovery of new pages.6

  • Deleting pages when stock runs out

    A sold-out product that's coming back should keep its page and say so. Google's feed rules have values for out of stock and backorder. Removal is for products you no longer sell.14

  • A new URL for every sale

    A fresh Black Friday page each November starts with no links and no history. Keep one address and update it. US shoppers spent $257.8 billion online from November 1 to December 31, 2025, by Adobe's count.2

  • Variants that compete with each other

    When every color and size is a separate page with nothing tying them together, links and history get split between them. Google's ProductGroup markup ties variants to one parent product.8

  • Site search pages in the index

    Internal search results are endless and thin. Google's spam policies count stacks of similar pages that sit closer to search results than to a browsable category tree as doorway abuse.17

  • A feed that disagrees with the page

    If the feed says in stock and the page says sold out, Google disapproves the product.14 It asks for the same match on price, between the feed, the page, and the checkout.13

Try it yourself

Check your own store in ten minutes

Six checks you can run from your phone right now, with no tools and no logins.

Found gaps? That's what the free audit covers, in detail.

Get a free AI visibility audit

Your checklist

0 of 6 checked

Careful by default

Changes tested before they reach every page

  • Testing

    Tested on a sample first

    Template changes run on a set of pages against a control group before they reach the whole store.

  • Rollback

    Logged and reversible

    Every change has a ticket, an owner, and a way back. Legal and brand review what shoppers will read.

  • Shortcuts

    Nothing that risks the domain

    No bought links, fake reviews, or mass-made pages. Your whole catalog sits on one domain.

How we work inside large organizations

FAQ

Ecommerce SEO questions

Something else on your mind? Ask it on your audit call.

How long does ecommerce SEO take?

Product data and crawl fixes usually show within weeks. Changes to category and product templates usually take three to six months, because each one is tested on a sample of pages before it rolls out. You get a report by page type every month.

How much does ecommerce SEO cost?

Three things drive it: the size of the catalog, how many templates need work, and how much your own engineers can ship. We scope it after the free audit, so the plan and the price are based on your actual gaps.

Can AI tools recommend a store like ours?

Yes. OpenAI says ChatGPT's product results aren't ads, and that it ranks merchants on things like availability, price, and quality. Merchants can also apply to send it a product feed directly.19 We track what the main AI tools answer for a fixed set of shopper questions.

Do we need Merchant Center if we don't run Shopping ads?

Yes. Google says a free listing can show your products at no cost in Search, on the Shopping tab, in Images, Lens, YouTube, and Gemini.12 The feed is worth fixing whether or not you buy ads.

Should filter pages be indexed?

Only the few that match real searches. For the rest, Google suggests blocking them in robots.txt or putting filters in URL fragments, the part after the # sign. It calls canonical tags and nofollow generally less effective than those two over the long term.6

What should we do with out-of-stock products?

If it's coming back, keep the page live and say when. In the feed, use out of stock if you're not taking orders, or backorder with a ship date if you still are.14 If it's gone for good, redirect to its replacement, or return a 404, a strong signal to Google not to crawl it again.7

Do we need to move to a new ecommerce platform?

Rarely. Most problems sit in templates, feed rules, and crawl settings, and those can change on the platform you have. If the platform truly can't do something that matters, we'll say so and show what it's costing you.

How do you work with our engineers?

Through your own process. Every change is a ticket with a spec, a test plan, and a way to roll it back. We test on a sample of pages against a control group, and your team ships the winners in its normal release cycle.

Related insights

Sources

  1. U.S. Census Bureau. E-commerce retail sales as a percent of total sales. Quarterly, seasonally adjusted. The chart uses each year's second-quarter row, dated April 1. Via FRED.
  2. Adobe. Holiday shopping season drove a record $257.8 billion online with consumers embracing generative AI tools. January 2026. Adobe Analytics data on US online shopping, November 1 to December 31, 2025.
  3. Adobe. AI traffic surges across industries, retail sees biggest gains. January 2026. Adobe Analytics data on visits to US retail sites.
  4. Google. Shop with AI Mode, use AI to buy and try clothes on yourself virtually. May 2025.
  5. Baymard Institute. 50 cart abandonment rate statistics 2026. Baymard's own survey of US online shoppers on why they abandon carts. Page last updated September 2025.
  6. Google Crawling Infrastructure. Managing crawling of faceted navigation URLs.
  7. Google Crawling Infrastructure. Optimize your crawl budget.
  8. Google Search Central. Product variant structured data.
  9. Google Search Central. Merchant listing structured data.
  10. Google Search Central. Share your product data with Google.
  11. Google Search Central. Pagination, incremental page loading, and their impact on Google Search.
  12. Google Merchant Center Help. Free listings for products.
  13. Google Merchant Center Help. Product data specification.
  14. Google Merchant Center Help. Availability [availability].
  15. Google Search Central. Review snippet structured data.
  16. Federal Trade Commission. Federal Trade Commission Announces Final Rule Banning Fake Reviews and Testimonials. August 2024.
  17. Google Search Central. Spam policies for Google web search.
  18. Google Search Central. AI features and your website.
  19. OpenAI Help Center. Shopping with ChatGPT Search. Checked October 2026.
  20. Ahrefs. What Triggers AI Overviews? 86 Factors and 146 Million SERPs Analyzed. November 2025, updated May 2026. Desktop results pages from September 2025. Ahrefs sells SEO tools.

Free audit

Find out if AI recommends your brand.

We'll check how your brand shows up on Google, ChatGPT, Perplexity, and AI Overviews, and show you what to fix first. Free, with no pitch deck.

https://

We only use these to send your audit. Privacy