Product schema, feeds and PDP structure that win in search and in-chat shopping.
Ecommerce SEO is a structural problem before it is a content problem. A catalogue of thousands of products, each with variants, spread across faceted category pages, with items going out of stock and returning, generates duplication and sprawl at a scale that editorial sites never encounter. Solve the structure — unique product content, real category pages, controlled facets and variants — and the rest follows, including the clean product data that increasingly feeds in-chat shopping answers.
The defining challenge of a large catalogue is that scale multiplies every structural decision. A faceting system generating combinations produces thousands of near-identical URLs; a variant model creates many pages describing almost the same product; manufacturer descriptions used verbatim mean your pages match every competitor selling the same item. None of these are content-quality problems in the usual sense — they are architecture problems producing content symptoms.
This is why ecommerce work looks different from editorial SEO. The highest-value interventions are usually rules and templates applied across thousands of pages rather than improvements to individual ones, and the most damaging failures are systematic rather than isolated. Understanding ecommerce as structural is what directs effort correctly: fixing the pattern that generates the problem rather than the individual pages exhibiting it.
The most common and most damaging ecommerce content failure is using manufacturer-supplied descriptions unchanged. Every retailer selling that product has the same text, which gives an engine no reason to prefer your page and produces duplication across the web at scale. It also gives the customer nothing they could not get anywhere else.
Writing genuinely distinct product content — what it is for, who it suits, how it compares to alternatives you stock, practical detail from actually handling it — is expensive across a large catalogue, which is why prioritisation matters: invest in the products that drive revenue and let the long tail carry lighter treatment. Understanding that unique descriptions are a competitive necessity rather than a nicety is why catalogue content deserves real budget rather than being treated as a data-population task.
Product structured data declares the facts of a product explicitly: name, description, images, price, availability, condition, identifiers, and aggregate review data. This makes the page eligible for enhanced search treatments and, increasingly importantly, supplies clean machine-readable facts to the systems that answer shopping questions.
Accuracy is essential and continuous: price and availability change constantly, and structured data that misstates them is worse than none, risking both guideline problems and customer frustration. The practical requirement is to generate schema from the same source that populates the page, so it updates automatically. Understanding Product schema as a live data feed rather than static markup is why its implementation belongs in the platform rather than in a content process.
Category pages are frequently the most commercially valuable pages on an ecommerce site, targeting the terms buyers actually search — and they are frequently left as bare grids of product links with no substantive content. This wastes the opportunity, because a grid of links tells an engine little about what the category is or why someone should choose from it.
Adding genuine content — what the category covers, how to choose between options, what distinguishes the types, answers to the questions buyers actually ask — makes these pages substantive. The practical caution is to place it where it serves shoppers rather than obstructing the products, which is a design problem with known solutions. Understanding that category pages need real content is why they deserve editorial investment comparable to major landing pages.
Faceted navigation is essential for usability and destructive for crawlability if uncontrolled. Each combination of filters can generate a distinct URL, and a moderate number of facets produces an enormous number of near-identical pages, consuming crawl budget and flooding the index with variations of the same content.
Control comes from deciding which facet combinations deserve to be indexed — typically those matching real search demand, such as a category plus a common attribute — and preventing the rest through robots directives, canonical tags, or link handling, applied as a deliberate rule rather than ad hoc. Understanding facet control as a design decision made once and enforced systematically is why it belongs in the technical architecture rather than being addressed reactively when index bloat appears.
A catalogue multiplies every structural decision by thousands. Fix the rules that generate duplication rather than the pages exhibiting it — and the clean product data that results is what feeds in-chat shopping answers.
Products with variants — sizes, colours, configurations — raise the question of whether each variant deserves its own indexed page. The answer depends on whether people search for the variant specifically: where a colour or size is genuinely searched, a distinct page can be justified; where it is not, consolidating variants onto one page with selectable options is cleaner.
The mechanism for consolidation is canonicalisation, pointing variant URLs at the primary product page so signals concentrate rather than fragmenting across near-identical pages. Getting this wrong in either direction costs: fragmenting authority across dozens of variant pages, or hiding genuinely-searched variants from the index. Understanding variants as a demand question answered through canonicalisation is why the decision should follow research rather than platform defaults.
Products going out of stock are a recurring ecommerce decision with no single right answer. Removing the page discards accumulated authority and produces errors for anyone arriving from a link. Leaving it unchanged frustrates customers. The workable approach depends on whether the product is returning: temporarily unavailable items usually keep their page with clear availability status and alternatives offered.
Permanently discontinued products are better redirected to the closest equivalent or to their category, preserving authority and delivering the customer somewhere useful. What matters most is having a defined rule applied consistently rather than case-by-case handling that produces inconsistent outcomes across a catalogue. Understanding out-of-stock handling as a policy decision is why it should be specified once and automated.
Product feeds supply structured catalogue data to shopping platforms, marketplaces, and increasingly to the systems answering shopping questions. Feed quality — accurate titles, complete attributes, correct identifiers, current pricing and availability, good images — determines how well products are represented wherever the feed is consumed.
The strategic point is that feed data and on-site product data should come from the same well-maintained source, since divergence between them creates inconsistency about basic facts. Investing in catalogue data quality therefore pays across every surface simultaneously. Understanding feeds as an expression of underlying data quality is why the durable fix for feed problems is usually upstream in product information management rather than in feed configuration.
AI shopping answers pull from clean product data and review signals. DUNkē tracks whether your products are the ones being recommended across eight AI engines — per prompt, against competitors.
Product reviews serve ecommerce SEO in several ways at once: they add unique content to pages that would otherwise carry only manufacturer copy, they supply the aggregate rating data that structured markup exposes, and they provide the evidence customers use to decide.
Their newest role is feeding AI shopping answers, which draw heavily on review content and sentiment when recommending products — not merely the score but what reviewers actually say about durability, fit, and suitability. This makes genuine review volume and substance a direct input to whether your products are recommended. Understanding reviews as both content and recommendation evidence is why systematic, legitimate review collection is among the highest-value ecommerce investments.
No team can give individual attention to every page in a large catalogue, so prioritisation is the practical constraint. Revenue-driving products and high-value category pages justify individual treatment; the long tail is served by templates, rules, and data quality that raise the floor without bespoke work.
This produces a two-tier approach: systematic quality applied everywhere through structure, schema, and clean data, with editorial investment concentrated where it changes outcomes. Attempting bespoke treatment across a whole catalogue exhausts the team and finishes nothing. Understanding the two-tier model is what makes ecommerce SEO tractable at scale, and it is why the structural work matters so much — it is what carries the pages nobody will ever write individually.
The recurring failures are structural. Manufacturer descriptions used verbatim make pages indistinguishable from every competitor. Category pages left as bare grids waste the most valuable terms. Uncontrolled facets flood the index with near-duplicates. Variant handling that fragments authority across dozens of near-identical pages weakens all of them. Deleting out-of-stock pages discards earned authority. And inaccurate structured data on price or availability creates problems worse than having none.
The remedies follow: write distinct content for products that matter and raise the floor with templates elsewhere, give category pages genuine substance, control facet indexation by deliberate rule, canonicalise variants according to real demand, define an out-of-stock policy and automate it, and generate schema from live data. Understanding these failure modes matters because each is systematic, which means fixing the rule corrects thousands of pages at once.
An ecommerce site’s own search logs are among the most underused research sources available. They record exactly what visitors want in their own words, including products you do not stock, terms you do not use, and questions your category pages fail to answer — all from people already on your site with commercial intent.
Mining them regularly reveals gaps in the catalogue, vocabulary mismatches between how you label products and how customers describe them, and category structures that do not match how buyers think. Each finding is directly actionable. Understanding internal search as first-party demand data is why it should feed both the SEO plan and merchandising decisions, since it is closer to real purchase intent than any external keyword tool.
Categories containing hundreds of products raise the question of how paginated listings should be handled. The considerations are that deep pagination buries products many clicks from entry points, that paginated URLs can consume crawl budget without adding value, and that visitors rarely go beyond the first pages.
Practical approaches include ensuring products are reachable through multiple routes rather than only through deep pagination, keeping page one genuinely representative, and ensuring paginated pages remain crawlable so products are discoverable. What fails is burying substantial parts of the catalogue where neither crawlers nor customers realistically reach. Understanding pagination as a discoverability problem is why it should be designed around getting products found rather than around listing conventions.
Duplication in ecommerce arises from several directions at once: manufacturer descriptions shared with competitors, variant pages nearly identical to each other, faceted URLs producing the same products in different orders, and syndicated product data appearing across many retailers.
Each has its own remedy — original descriptions, canonicalisation, facet rules, and added unique content respectively — but they compound if addressed piecemeal. The practical approach is to audit where duplication originates structurally rather than treating individual duplicate pages. Understanding duplication as several distinct mechanisms is why a single fix rarely resolves it, and why the audit should identify which mechanisms are active before remedies are chosen.
Ecommerce sites are unusually prone to performance problems, because product imagery is heavy, catalogue pages load many items, and commercial platforms accumulate third-party scripts for analytics, personalisation, chat, and marketing. The result is frequently slow pages on exactly the templates that matter commercially.
The remedies are known — image optimisation, deferring non-critical scripts, auditing third-party additions for genuine value — but they require ongoing discipline, since each new marketing tool adds weight. Performance also affects conversion directly, which usually makes the commercial argument easier than the visibility one. Understanding why ecommerce accumulates performance debt is why script auditing deserves a standing place in the process rather than occasional attention.
Product schema is the priority, but several other types earn their place on an ecommerce site. Breadcrumb markup reinforces the category structure and can improve how listings display. Organisation schema anchors the brand entity. Review and rating markup exposes the aggregate evidence customers and engines use. FAQ markup suits genuine product and delivery questions.
The consistent requirement is accuracy: markup must reflect what is actually on the page and remain correct as data changes, which for a catalogue means generating it from the same source that populates the page. Understanding the broader schema set is why an ecommerce implementation should be planned as a data-driven system rather than added page-type by page-type as needs arise.
Retailers create pages for seasonal events and promotions, which raises a recurring decision: whether to build new pages each cycle or maintain permanent ones. Permanent pages that are updated each season accumulate authority across years, while newly-created pages start from nothing each time and leave a trail of expired URLs behind them.
The workable approach is a durable URL per recurring event, refreshed with current offers and content as each cycle approaches, with sufficient lead time for the update to be indexed before demand peaks. Understanding the accumulate-versus-restart trade-off is why seasonal strategy should default to maintained permanent pages, and why the timing of the refresh matters as much as its content.
Editorial content earns its place on an ecommerce site when it serves the questions buyers ask before choosing: how to select between types, what specifications actually mean in practice, how to size or configure, how products compare on the criteria that matter. This content reaches people earlier in the decision than product pages can.
It also supplies the substantive material that both topical authority and AI citation reward, which product pages alone rarely provide. The discipline is to build it around genuine purchase questions rather than publishing generic content adjacent to the category. Understanding why ecommerce sites need editorial content is that the buying decision starts before the product page, and being present for that stage is what brings buyers to it.
As assistants answer product questions and make recommendations, the quality of your underlying product data becomes a visibility factor rather than merely an operational one. Complete attributes, accurate specifications, correct identifiers, current pricing and availability, and genuine review content are what allow a product to be represented accurately in a recommendation.
Incomplete or inconsistent data produces products that cannot be confidently recommended, regardless of how good they are. The practical implication is that product information management — usually treated as an operations concern — has become part of the visibility stack. Understanding this convergence is why the durable ecommerce investment is clean, complete, well-maintained catalogue data, which serves search, feeds, and AI recommendations simultaneously.
Ecommerce sites replatform relatively often, and migrations are where hard-won visibility is most commonly destroyed. The risks are concentrated and known: URL structures change without complete redirect mapping, product data is lost or truncated in transfer, structured data is not reimplemented, and category structures are reorganised in ways that abandon established pages.
The protections are equally known — a complete URL inventory mapped to destinations before launch, verification that product data and markup survive, and close monitoring in the weeks following — but they require planning time that migration schedules rarely allocate. Understanding migration as the highest-risk event in an ecommerce site’s life is why the SEO requirements belong in the project plan from the outset rather than being raised late.
Ecommerce teams face a genuine tension between merchandising priorities — promoting particular products, controlling presentation, driving specific journeys — and discoverability, which favours substantive content, crawlable structures, and pages built around what customers search for rather than what the business wants to push.
The resolution is usually design rather than compromise: category content placed where it informs without obstructing products, promotional emphasis achieved through layout rather than by removing substance, and journeys that accommodate both entry from search and internal browsing. Understanding this tension as a design problem with known solutions is why it should be worked through collaboratively rather than settled by whichever team has more influence, since both requirements are legitimate.
Across a large catalogue, returns concentrate unevenly. The highest-return work is usually category page content, since these target the terms with real commercial volume and are frequently left as bare grids; structural fixes to facets and duplication, which affect thousands of pages at once; and product data quality, which serves search, feeds, and AI recommendations simultaneously.
Individual product page optimisation pays well for the small number of items driving disproportionate revenue and poorly across the long tail, where template quality and clean data do the work. Understanding where effort pays is why an ecommerce plan should lead with categories, structure, and data rather than with the product-by-product optimisation that intuition suggests and that exhausts teams without moving much.
Almost every ecommerce recommendation in this piece ultimately depends on catalogue data quality: unique descriptions, accurate schema, reliable feeds, correct availability, complete attributes, and genuine review content all rest on the underlying product information being complete and current.
This means the highest-leverage ecommerce investment is frequently upstream of anything recognisably SEO — in the systems and processes that maintain product data. A site with excellent structure and poor data will underperform one with adequate structure and excellent data, because the data is what fills every surface: pages, markup, feeds, and increasingly AI recommendations. Understanding the data foundation as primary is why an ecommerce audit should start with information quality rather than with page-level optimisation.
Ecommerce SEO is structural: scale and duplication are the core challenge, and the highest-value interventions are rules applied across thousands of pages rather than improvements to individual ones. Products need genuinely unique descriptions rather than manufacturer copy, category pages need real buying guidance rather than bare grids, facet combinations need deliberate indexation rules, variants need canonicalisation based on actual search demand, and out-of-stock handling needs a defined policy applied consistently.
Product schema and feeds should be generated from the same well-maintained catalogue data so they stay accurate on price and availability, since clean product data now feeds in-chat shopping answers as well as traditional search. Reviews matter doubly, supplying unique content and the sentiment evidence AI shopping recommendations draw on. And because no team can treat every page individually, the model that works is systematic quality everywhere plus editorial investment where revenue justifies it.
“A catalogue multiplies every decision by ten thousand. Ecommerce SEO is mostly about fixing the rule that generates the problem, not the pages that display it.” The Age’X Research Team
Get a free GEO audit — the same analysis behind every article here.