WHO WE HELP · ECOMMERCE

An ecommerce SEO agency that treats your catalogue as the product.

An ecommerce SEO agency works at catalogue scale: controlling which faceted navigation URLs are crawlable, stopping product variants and category pages competing with each other, handling out-of-stock and discontinued products without losing their equity, and marking up products so they qualify for shopping and AI answer surfaces.

  • ✓ the founder reviews every domain personally
  • ✓ a reply within 48 hours
  • ✓ four reviews taken a month

THE NUMBERS THAT SHAPE THIS WORK

  1. 51%

    of ecommerce sites have "mediocre" or worse product page UX, across 344 benchmarked US and European retailers

    Baymard Institute, Product Page UX benchmark
  2. 1,300+

    distinct product page usability issues recorded in moderated testing for that benchmark

    Baymard Institute, Product Page UX benchmark
  3. Infinite

    URL spaces that parameter-based faceted navigation can generate, in Google’s own words, harming the site

    Google Search Central, faceted navigation
  4. 34.5%

    lower click-through rate at position one when an AI Overview is present, across 300,000 keywords

    Ahrefs, 300,000-keyword study, April 2025
  5. +115%

    AI visibility lift from citing sources, for pages starting around position five

    Aggarwal et al., GEO: Generative Engine Optimization, KDD 2024

Ecommerce SEO is an architecture problem that happens to involve content. A 4,000-product catalogue with colour, size, price and brand filters can generate hundreds of thousands of crawlable URLs, almost all near-duplicates. Google will spend your crawl budget on them and reach the category pages that matter late. No amount of blog content fixes that.

So the work starts at the template and URL layer, where one decision applies to ten thousand pages at once.

WHERE CATALOGUE SEO BREAKS

  1. Faceted navigation is generating URLs faster than Google can crawl them

    Four filters with ten values each, combinable and order-independent, is tens of thousands of URLs from a single category. Most stores expose all of them, then wonder why new products take three weeks to index. The fix is a deliberate policy per facet: which combinations are indexable because people search them, which are crawlable but not indexable, and which are blocked outright.

    Google attributes the problem to URL-parameter filtering, which it says can generate infinite URL spaces. Google Search Central, faceted navigation

  2. Your product variants are competing with each other

    The same shoe in seven colours, published as seven URLs with near-identical copy, splits the ranking signals seven ways and lets Google pick a winner on its own. Sometimes it picks the one that is out of stock. Canonical strategy, variant consolidation and which attributes genuinely deserve their own URL are decisions that change revenue, not hygiene tasks.

    Google’s product documentation treats variants as a distinct markup case, not as separate products. Google Search Central, product structured data

  3. Out-of-stock pages quietly destroy accumulated equity

    A product that has earned links and rankings over two years gets discontinued, and the platform 404s it, redirects it to the homepage, or leaves a live page with a dead buy button. All three waste the asset. Seasonal lines returning next quarter, permanent discontinuations and temporary stockouts each need a different answer, and most stores apply one rule to all three.

    Availability is a product property Google reads; marking it correctly is the difference between a stale listing and an honest one. Google Search Central, product structured data

  4. Your category pages are cannibalising each other

    "Running shoes", "mens running shoes", "running trainers" and a landing page built for a paid campaign all target the same intent with overlapping copy. Google consolidates, picks one, and it is rarely the one you optimised. Deciding which pages own which intent, then merging or differentiating the rest, usually recovers more than any new content would.

    Baymard found 51% of 344 benchmarked retailers have mediocre or worse product page UX, the layer category pages feed into. Baymard Institute, Product Page UX benchmark

  5. Crawl budget is spent on the wrong half of the catalogue

    Log files on a large store routinely show Googlebot spending most of its requests on filter permutations, paginated depth, internal search results and parameter noise, while bestsellers go weeks between crawls. Until that ratio is fixed, every content and merchandising change takes longer to register than it should.

    Google: crawling spent on useless URLs leaves less time for new, useful ones. Google Search Central, faceted navigation

  6. Your products are invisible to AI shopping surfaces

    Assistants that recommend products read structured data: price, availability, currency, GTIN, condition, review counts, return policy. Incomplete or stale Product and Offer markup means you are omitted from answers that your competitors appear in, and the omission is silent. It is also one of the few places where a small retailer can be chosen on data quality rather than domain authority.

    Ahrefs measured a 34.5% lower position-one CTR where an AI Overview appears, across 300,000 keywords. Ahrefs, 300,000-keyword study, April 2025

WHAT WE DO ABOUT IT

The ecommerce playbook, in the order it pays.

  1. A facet and parameter policy, written down

    Every filter gets an explicit decision: indexable, crawlable-not-indexable, or blocked, driven by whether real people search that combination. Then canonicals, robots rules and internal linking are made to agree with it, because conflicting signals are worse than either choice alone. That is template-level work, which is technical SEO.

    When
    Weeks 1–3
    Who
    Senior technical strategist, with your developer for the ship
    Deliverable
    A written policy, one line per facet, plus the robots, canonical and internal-link changes that enforce it
  2. Variant and canonical consolidation

    We decide which attributes deserve a URL and which belong on one page behind a selector, then consolidate the rest so ranking signals stop splitting. Done once at the template, it applies across the whole catalogue, which is why it usually outperforms months of content work.

    When
    Weeks 2–6
    Who
    Technical strategist, implemented in your template layer
    Deliverable
    A variant decision table and the canonical rules, applied catalogue-wide rather than page by page
  3. Category pages that are worth ranking

    Most category pages are a grid and nothing else. The ones that win carry genuine buying guidance above or below the grid, internal links to the sub-intents they should own, and copy that answers the comparison question the shopper actually has. That is where the full-stack programme.

    When
    Month 2 onwards, weekly
    Who
    Senior editor, briefs approved by a strategist
    Deliverable
    An intent map naming which page owns which query, then merged, differentiated or removed pages and the buying guidance that goes above the grid
  4. Product and Offer markup built for answer surfaces

    Complete, accurate, automatically maintained structured data: price, availability, shipping, returns, reviews and identifiers. It is the entry ticket for rich results today and for assistant-driven shopping answers now being built, which is the territory of answer engine optimisation.

    When
    Weeks 3–8
    Who
    Technical strategist with whoever owns the product feed
    Deliverable
    Feed-generated structured data that stays accurate through price and stock changes, validated against Google’s rich result tests
  5. Being named when a shopper asks an assistant

    Shoppers increasingly ask for a recommendation before they visit a store. Appearing in that answer depends on corroborating third-party mentions, clean product data and claims a model can quote, which is the work of generative engine optimisation.

    When
    Month 2 onwards, continuous
    Who
    Strategist plus digital PR
    Deliverable
    A tracked set of buying prompts, measured monthly across the major assistants, with the corroborating mentions that move them
SIDE BY SIDE

What changes when the agency has worked a catalogue before

CapabilityGeneralist SEO agencyEcommerce specialist
Faceted navigationBlanket noindex, or left alonePer-facet policy driven by actual search demand
Product variantsTreated as separate pagesConsolidated unless the attribute has its own demand
Out-of-stock productsOne rule for every caseDifferent handling for temporary, seasonal and discontinued
Crawl budgetDiscussed in theoryMeasured from server logs before any plan is written
Category pagesMore content addedIntent assigned per page, the rest merged or removed
Product schemaAdded once at launchGenerated from the feed so it survives a price change
ReportingSessions and rankingsRevenue by category, plus indexation rate for new SKUs
ReplatformingNot offeredRedirect map and recovery plan built before migration

if crawling is spent on useless URLs, the crawlers have less time to spend on new, useful URLs.

Google Search Centralofficial documentation on faceted navigationGoogle Search Central, faceted navigation
WHAT LANDS WHEN

The first six months, honestly.

What is delivered at each point, and what it is realistic to see. These assume the technical work is not sitting in your engineering queue.

  1. Week 1

    Crawl of the full catalogue, server log sample, facet and parameter inventory, variant and canonical audit.

    You see the real size of the indexable surface, which is usually several times what anyone expected.

  2. Month 1

    Facet policy agreed and shipped, canonical and variant consolidation live, out-of-stock rules implemented, product schema wired to the feed.

    Crawl distribution starts shifting towards pages that matter. New products begin indexing faster.

  3. Month 3

    Category intent map applied: pages merged, differentiated or removed. Buying-guide content shipping weekly against the mapped gaps.

    Consolidated category pages start to recover positions they were splitting between them.

  4. Month 6

    Authority and citation work on the categories that convert, refreshed product markup, measurement on revenue by category.

    The compounding curve, if the architecture work landed. This is where catalogue SEO pays or does not.

HOW AN ENGAGEMENT WORKS

What happens after you send a domain.

  1. The free founder review

    You send a domain. The founder goes through the catalogue personally: how many crawlable URLs your filters can generate, which variants are splitting signals, what happens to a discontinued product today, and whether your product markup survives a stock change. A verdict within 48 hours.

  2. Read the logs before proposing anything

    Crawl budget on a catalogue is an empirical question. Before scoping we want to see where Googlebot actually spends its requests, because a plan written without that is written from assumptions. If logs are unavailable we say so and scope around it.

  3. Scope and price, after the review

    No rate card. A 300-product store and a 40,000-SKU catalogue with four filter dimensions are not the same job, and quoting before seeing the facet behaviour is guesswork with a currency symbol on it.

  4. Template decisions first, pages second

    Facet policy, variant consolidation, out-of-stock rules and automated product markup come before any content work, because each decision applies to thousands of URLs at once. Content follows, once the architecture under it is worth ranking.

  5. Handover, not hostage-taking

    The facet policy, the redirect rules and the schema templates are documented and live in your stack. If you replatform in two years, the decisions travel with you instead of leaving with us.

Still building a shortlist?

Then this page is the wrong one, and we would rather say so. We publish a researched comparison of the 11 best ecommerce SEO agencies, with published pricing where it exists and who each one does not suit. We are on it, positioned where we honestly belong. Read the comparison.

WHO THIS IS NOT FOR

Four reasons to walk away.

Written plainly, because a wasted discovery call costs us both more than an honest paragraph does.

  • Stores with fewer than about 50 products. The architectural work that makes this valuable barely exists at that size, and you would be paying for a specialism you do not need.
  • Anyone who needs the store rebuilt or replatformed. We do not build or migrate stores. Several agencies on our ecommerce shortlist do, and that is a better fit.
  • Retailers whose growth is decided on Amazon and retail media rather than their own domain. Organic search on your own site is not the lever there.
  • Buyers who need an ecommerce reference before signing. Our one measured result is B2B SaaS. That is a fair requirement and today we would not meet it.
THE AWKWARD QUESTIONS

What you are probably thinking.

  1. How will I know it is actually working before revenue moves?

    Because the leading indicators move first and we agree them up front. Indexation rate for new products, the share of crawl requests hitting commercial pages rather than filter permutations, and the count of category pages competing for the same intent are all measurable within weeks. Revenue by category follows. If those three are flat at week six, something is wrong and we say so rather than waiting for month four to have the conversation.

  2. What happens if we replatform halfway through?

    The work is documented so it travels. Facet policy, canonical rules, out-of-stock handling and schema templates are written down as decisions, not left as undocumented config in one platform. We do not run migrations ourselves, so you would bring in a build partner, and we would supply the redirect map and the pre- and post-migration crawl comparison. Replatforming without that is where stores lose years of accumulated links.

  3. What if you get busy with other clients?

    Four founder reviews a month is the cap, and it exists for this reason. We are a small senior team by design, which means you get the people who did the audit rather than an account manager relaying their notes. It also means we turn work down. If we cannot give a catalogue the attention it needs, the honest answer at the review stage is no, and you get that answer in 48 hours rather than discovering it in month three.

  4. Why not just hire someone in-house?

    For a large catalogue, often you should, eventually. The problem is the first year: the architectural decisions that matter most are made once, early, and getting them wrong is expensive to undo. An in-house hire learning facet policy on your live store is a costly way to learn it. The sensible shape is usually specialist help for the foundation and the first compounding cycle, then an in-house owner to run it, with the documentation handed over so they inherit decisions rather than guesswork.

FAQ

Ecommerce SEO, asked and answered.

Not sure you are a fit?

Apply anyway. The founder reviews every domain personally and tells you what you actually need, including when the honest answer is "not us, not yet".

Apply for review ↗

It works on the architecture that governs thousands of pages at once: which faceted URLs are crawlable, how product variants are consolidated, how out-of-stock products are handled, which category page owns which intent, and whether product markup is complete enough for rich results. Content matters, but on a catalogue the template decisions move far more revenue.

Decide per facet rather than globally. Combinations people genuinely search, such as brand plus category, usually deserve indexable pages with their own copy. Combinations that only produce near-duplicates should be crawlable but not indexable, or blocked entirely. Then make canonicals, robots directives and internal links agree, because mixed signals cost more than either decision.

It depends on why it is out of stock. Temporary stockouts keep the page live with availability marked correctly in schema and alternatives offered. Seasonal products returning later stay live rather than being redirected. Permanent discontinuations redirect to the closest equivalent product or the parent category, never the homepage, so the accumulated links keep working.

Only when people search for them. Colour and size usually do not deserve their own URLs and split ranking signals when they get them, so they belong on one page behind a selector with a single canonical. Attributes with genuine standalone demand, such as capacity on a phone, can justify separate pages, provided each has its own differentiated content.

Technical and architectural fixes often show first, because they change how a large number of existing pages are crawled and consolidated rather than waiting on new content to earn authority. Category and buying-guide work follows the ordinary curve of several months. On a large catalogue the first measurable change is usually how quickly new products get indexed.

No, but it has known constraints you design around rather than fight: a forced /collections/ and /products/ URL structure, product pages reachable from multiple collection paths, and limited control over how filters generate URLs. These are manageable with the right canonical and internal-linking approach. The platform is rarely the reason a store is not ranking.

We do not publish a rate, because a 300-product store and a 40,000-SKU catalogue with four filter dimensions are not the same job. Apply for the free founder review and you get a scoped figure within 48 hours, after someone has looked at your actual catalogue, your facet behaviour and your crawl data.

Not yet, and we will not borrow one. Our one measured, cross-checkable result is sendr.ai in B2B SaaS: zero to 1.05M impressions in six months. The technical craft on this page transfers directly; the ecommerce results do not exist until we have them. If you need a vertical-specific reference before signing, we are not the right choice today.

More than most teams believe. Google states that parameter-based faceted navigation "can generate infinite URL spaces which harms the website", because filters combine and reorder freely. Four filters with ten values each is already tens of thousands of combinations from one category. The practical question is never how many exist, but which ones real people search for, because those are the only ones worth letting Google index.

Yes, and increasingly for a different reason. Rich results were the original payoff; the current one is eligibility for assistant-driven shopping answers, which read structured data rather than page copy. Google notes that "the more properties you can add, the more enhancements your page can be eligible for". The catch is accuracy: markup generated once at launch goes stale on the first price change, which is why it has to come from the feed.

Enough to change how you measure. Ahrefs studied 300,000 keywords and found the presence of an AI Overview correlated with a 34.5% lower click-through rate at position one. For ecommerce that pressure lands hardest on research and comparison queries rather than on branded or transactional ones, which is an argument for strengthening the commercial pages rather than abandoning informational content.

It becomes one. Baymard benchmarked 344 US and European retailers and found 51% have "mediocre" or worse product page UX, with more than 1,300 distinct usability issues recorded in testing. Those issues show up in engagement and return-to-SERP behaviour long before they show up in a crawl report, and they cap what any amount of ranking work can convert.
FOUR SLOTS A MONTH

Four founder slots a month.
One is yours if it fits.

Drop your domain. The founder runs the 52-check review personally and comes back within 48 hours with what your facets, variants and product markup are really doing, even when the answer is no.