Most eCommerce stores with large product catalogues share a common problem that quietly kills growth. It's not their products, pricing, or marketing budget—it's their site structure. Sam Wright, founder of Blink SEO and creator of Macalytics, reveals why taxonomy is the biggest drag on growth for stores doing £3-5 million annually, and exactly how to fix it using Search Console data.We explore why collection pages represent 35% of all search impressions (more than products and blogs combined), how to determine the right level of granularity for your categorisation, and why most stores aren't deep enough with their subcategories. Sam shares his framework for using Search Console impression data to identify exactly where to create new collection pages, and explains the critical difference between what works for user experience versus what search engines can actually index.Key Point Timestamps:06:30 - The Large Catalogue Challenge11:45 - Why Collection Pages Are Your Biggest SEO Opportunity16:20 - The Granularity Problem Most Stores Face22:15 - Using Search Console Data to Guide Taxonomy27:40 - Real-World Example: Redesigning for Better Structure35:10 - Future-Proofing for AI Search with Persona Data42:30 - The AI Shortcut and Critical WarningThe Large Catalogue Challenge (06:30)Sam defines large catalogue stores as those where the buying journey tips into a different mode—one based around comparison and filtering rather than simple browsing. This typically happens around 250 products, though it varies by category."With large catalogue stores, the buying journey is based around comparison and filtering," Sam explains. "A lot of the time these stores have grown up organically over a period of time and no one's taken ownership about how the store's organised."This organic growth creates a drag on everything—SEO, user experience, conversion rates, even email segmentation. Stores reach £3-5 million in annual revenue, so things are fundamentally working. But growth isn't happening as fast as it should because nobody stepped back to think strategically about organisation and purpose.Why Collection Pages Are Your Biggest SEO Opportunity (11:45)Sam shared compelling data from across all the Shopify stores his agency works with: "It's about 35% of all impressions come on collections, which is much more than products and blogs. It's basically the entry point for most people when they're doing actual new product discovery."More than a third of search visibility comes from collection pages—the pages where new customers first encounter the store. Yet most stores aren't categorised in a way that aligns with how people actually search for their products.This represents a massive untapped opportunity. If collection pages are already driving 35% of impressions without optimisation, imagine the potential when they're properly structured and aligned with search behaviour.The Granularity Problem Most Stores Face (16:20)The real opportunity for most stores lies in going deeper with categorisation. Much deeper."Most people are not granular enough with their categorisation," Sam emphasises. "A lot of stores will just have a t-shirts category. They won't subcategorise those t-shirts to the level that matches how people are actually searching."Sam uses sofas as an example: "So sofas as the parent category, like blue sofas, blue four seat sofas, blue four seat corduroy sofas. That filtering process, that is how people do search."The challenge on Shopify is that these filters aren't indexable for search engines. Google ads can't effectively target filters either. The solution is breaking out popular subcategories into actual collection pages."The real opportunity for a lot of stores is how deep you go in that categorisation because you've got products that other people don't have," Sam explains. "And that's the easiest way to capture new users."Using Search Console Data to Guide Taxonomy (22:15)Search Console data reveals exactly how people are actually searching, throwing up interesting patterns in the impression data.Sam's agency uses this practical approach: "If you look at the Search Console data for a collection page, what you might see is what we would call attribute searches on a collection. So the collection is blue sofas, but we're seeing impressions for eight-seater sofa, like blue four seat sofa, blue four seat corduroy."When these search terms appear on a collection page, it's a signal to drill down another level. "You can typically keep going until we probably say like three in stock products is probably the limit for how far down you can go," Sam notes. "And actually, only having a few products is actually quite a good user experience a lot of the time."Having three tightly related products for a specific search creates a focused, relevant experience that converts better than overwhelming customers with too many options.Real-World Example: Redesigning for Better Structure (27:40)Matt shared his experience redesigning the eCommerce Podcast website, which had grown organically over 200+ episodes with just a blog-style feed and search bar.The solution was creating a proper hierarchy: top-level categories like "Marketing & Growth," subcategories like "SEO & Content" and "Messaging & Automation," then specific topics like "Technical SEO" or "Email Automation.""Now you can quite quickly navigate using that menu," Matt explained. "If you wanted to find out about technical SEO, within one click really, you can see that link and you can go, well, that's marketing and growth or technical SEO. You can click that. And then all of the episodes we've done which are connected to technical SEO then come up on the page."The result was better user experience and better SEO, with Google now showing dedicated pages for specific topics, each with multiple pieces of related content.Sam's response captured the principle perfectly: "What's always interesting is the further you get into something, into the levels, the more interesting and more useful that information is. And I think that transfers directly to the e-commerce experience as well."Future-Proofing for AI Search with Persona Data (35:10)In his "saving the best till last" segment, Sam shared what he believes will set businesses apart as AI search evolves: enriching product data with persona and use case information."I think the real thing that's going to set a lot of businesses apart in the future is by bringing in attribute data that other people don't have, and that is based on things like personas and use cases," Sam explained. "So this is kind of more human data."He referenced ChatGPT's announcement example: "I want to see the best coffee machine under $200 that captures the taste of coffee in Italy."Traditional attributes—price, product type—are straightforward. But "captures the taste of coffee in Italy" represents persona-based, use case-focused, benefit-driven data that most stores don't systematically capture."Getting that into your product data, I think, is going to be something that really, really sets lots of people apart," Sam emphasised. "Especially when we're potentially moving into this world where someone might not touch the website."The immediate benefit doesn't require waiting for AI search. "If you've got your persona types clearly laid as a Metafield in Shopify, you could use that for custom tagging on a collection page, best for X," Sam explains. "That kind of curated experience we know works really, really well."The AI Shortcut and Critical Warning (42:30)Both Sam and Matt discussed using AI to accelerate taxonomy work. "I swear like 80% of my working days is talking to chat GPT," Sam shared. "Working through things step by step. That's probably how I spend most of my days now."Matt had similar success using Perplexity Labs to research menu structures, getting about 70% of the way there as a starting point for conversations.But Sam offers a critical caveat: "The thing about AI and automation is it's only going to accelerate what you're doing. If what you're doing is a mess, you're just accelerating a load of mess basically."This is the uncomfortable truth. AI won't fix a fundamentally broken site structure—it'll just help scale the confusion faster. The foundation must be right first, based on actual data about how people search. Then AI and automation become powerful accelerators rather than mess multipliers.Today's GuestToday's guest: Sam WrightCompany: Blink SEOWebsite: blinkseo.co.ukLinkedIn: Connect with Sam on LinkedIn