How Does Ecommerce Site Structure Affect SEO?
Structure sets the ceiling. Everything else in this cluster works within a limit that structure has already fixed. Most shops have never examined theirs because it arrived with the catalogue rather than being decided.
Why Structure Sets The Ceiling
A shop can only rank for terms it has pages for. Structure decides which pages exist, so it decides which terms are even available before anybody writes a word.
A concrete example. A shop selling industrial fasteners, organised entirely by manufacturer.
Top level: eleven manufacturer names. Below each: that manufacturer's range. Every product correctly filed, the catalogue accurate, the site well built.
What customers actually search for. A size, a material and a type.
Somebody needs stainless steel socket screws in a particular gauge. They do not know or care which manufacturer makes them. They will never type one of those eleven names.
What content cannot fix. Anything.
You could write two thousand words on every manufacturer page and it would change nothing, because the terms those pages target are terms nobody searches. The pages are competing for demand that does not exist while the demand that does exist has no page at all.
What the fix actually is. New categories, cutting across the existing tree.
Pages for material, for type, for size. The products do not move. What changes is the set of routes into them, plus the set of terms the shop can compete for.
Why this is the highest leverage work available. Because it raises the ceiling rather than working under it.
Every other improvement in this cluster makes existing pages better. This one decides which pages exist to be improved.
The boundary worth stating now. This page is about a shop that already exists.
Structure as a decision made during a build sits in our ecommerce web design guides. Same subject, opposite end of the timeline.
Category Hierarchy
How categories nest and how deep products sit affects two audiences with different tolerances. A structure that suits one can fail the other.
What depth means to a shopper. Clicks between arriving and buying.
Each level is a decision. Each decision loses people. A shopper four levels from a product is being asked to commit to a path before seeing whether it leads anywhere useful.
What depth means to a crawler. Priority, effectively.
Pages further from the home page tend to be treated as less important and visited less often. On a large catalogue, products buried deep can be reached rarely enough that changes to them take a long time to register.
What a sound hierarchy looks like. Broad at the top, shallow in practice.
Enough top level categories to cover the range without a menu nobody reads. Two or three levels to any product by the main route. Every product reachable by more than one path, so it is not dependent on a single chain.
The failure at each extreme. Both are common.
Too flat means one enormous category containing eight hundred products, which serves nobody and competes for a term far too broad to win. Too deep means narrow categories with four products each, splitting demand into pieces too small for any of them to establish anything.
The test that resolves it. Block six gives the splitting test.
Whether a category deserves to exist is a question about demand rather than about tidiness, which is block three.
Building Categories From Demand, Not From The Feed
Most shops inherited their category tree from a supplier's product data. That tree reflects how an industry organises stock, which is frequently nothing like how customers look for it.
Why this happens to almost everybody. Because it is the path of least resistance.
The feed arrives with categories already attached. Importing it produces a working shop in an afternoon. Nobody decides to use the supplier's taxonomy. It simply is not questioned. Years later it is the shape of the business.
Why the two differ so reliably. They were built for different purposes.
A supplier organises by what is convenient to manufacture, stock and distribute: by brand, by range, by product code family. A customer organises by the problem they have: by what it is for, what it fits, what it is made of, who it is for.
Four tests for your own tree. All of them can be done this afternoon.
Look at your top level. If the first thing a visitor chooses between is a manufacturer name, the tree came from the feed.
Read your category names aloud. If they are not words a customer would use unprompted, they came from somewhere internal.
Check your own site search. What people type into your search box is the most direct evidence available of how they think about your range. If those words are not category names, you have found the gap.
Ask the people who answer the phone. They hear the customer's own words all day. They will tell you in five minutes what the analytics take a month to suggest.
What to do with the findings. Add rather than replace, initially.
New categories cutting across the existing tree cost little and risk nothing. Rebuilding the whole hierarchy changes addresses and belongs in block nine.
Address Structure
What your page addresses look like matters less than most articles claim and is far harder to change than most articles admit. Both halves of that are worth understanding.
What the options actually do. Less than people expect, directly.
Whether a product sits under its category path or at a flat product path has a modest effect in itself. The words in the address help marginally. Readability helps when somebody shares a link.
Why it still matters. Because of what changing it costs later.
Once a shop is established and linked to, every address is an asset somebody else is pointing at. Changing them means redirecting all of them. Anything missed loses whatever that page had accumulated.
What that makes this decision. Close to irreversible in practice.
Not technically irreversible. Expensive and risky enough that it is rarely worth doing on its own, which our migration guides cover.
What to settle if you can still choose. Three things.
Consistency, so the same kind of page always looks the same. Stability, so an address does not contain anything likely to change such as a year or a price. And a single canonical route to each product, so the same item is not reachable at several different addresses that then compete.
If your platform fixes some of this. Several do. It is not usually the constraint people fear.
Our platform guide covers what each allows, plus why the limitation costs work rather than outcomes.
Filters And Faceted Navigation
Filters are excellent for shoppers and, left unmanaged, they are the single most destructive structural problem a large shop can have.
What happens. Every filter combination can become an address.
Colour, size, price band, brand, material, rating. Each selection produces a distinct page. The combinations multiply against each other rather than adding up.
The scale of the multiplication. This is why it gets away from people.
Six filters with a handful of options each do not produce a few dozen pages. They produce thousands. Each one is a near duplicate of a category page that already exists. A shop with four thousand products can present tens of thousands of findable addresses without anybody having created a single one deliberately.
At what point it becomes serious. Sooner than most shops notice.
On a small catalogue it is untidy. Once the generated pages substantially outnumber the real ones, attention that should reach products is being spent working through combinations instead, which is exactly the failure our complete guide uses as its example of wasted work.
The decision to be made, rather than the configuration. Three categories. Every facet belongs in one.
Facets that deserve their own page. A small number, where genuine demand exists for that combination.
A brand within a category may be something people actually search for. So may a material within one. Those are worth existing as proper pages with their own content, per our category page guide.
Facets that should never be findable. The large majority.
Price bands, sort orders, availability toggles and multi-select combinations. Nobody searches for them and each one dilutes the page it came from.
Facets in between. Useful to shoppers, not worth being found for, so reachable but not treated as separate pages.
The rule that decides which. Does anybody search for this combination?
If yes, build it properly. If no, it should not exist as a findable page. Almost every faceted navigation problem we see comes from that question never having been asked.
Pagination And Listing Pages
How a long category is broken into pages decides how much of your catalogue is reachable at all. It is one of the least examined decisions on a shop.
Why it matters more than it sounds. Products on later pages are further from the home page.
A category with two thousand products split into eighty pages puts the last few hundred products a long way down a chain. They are reachable in theory and visited rarely in practice.
The three common approaches. Each with a consequence.
Numbered pages. The traditional approach. Predictable, reliable, then creates the depth problem above on a long category.
Load more on demand. Comfortable for shoppers. Whether the additional products are discoverable depends entirely on how it is built. It is frequently built in a way that hides them.
Continuous loading. The same concern, more so, plus it removes the footer a shopper may be trying to reach.
What resolves most of it. Not the pagination method.
Better categories. A category needing eighty pages is usually a category that should have been several categories, which returns to block three. Splitting by something customers actually search for fixes the depth problem and creates targetable pages at the same time.
What to check on your own shop. Whether products on later pages are being reached at all.
If a large part of the catalogue is only reachable by paging deep into a listing, block seven is the answer rather than any pagination setting.
Internal Linking
Internal linking is how a shop distributes what it has earned across its own pages. Most shops rely entirely on what the theme produces automatically, which is rarely enough.
What the theme gives you. Navigation, breadcrumbs and product grids.
All useful, all automatic and all identical on every page, which means they distribute evenly rather than deliberately. Nothing in that arrangement says which categories matter.
What a flat shop does with what it has. Leaks it.
A shop where every page links to every other page in the same way spreads its authority thinly and uniformly. The home page passes value to two hundred equally weighted destinations. Nothing accumulates anywhere in particular.
What deliberate linking looks like. Four things a theme will not do.
Priority categories linked from the home page directly, rather than buried in a menu that technically contains everything.
Related categories linked to each other, reflecting how people shop rather than how the catalogue is filed. Somebody looking at one thing is frequently considering an adjacent thing.
Content linking down into categories. A buying guide that mentions products should link to the category pages rather than only to individual items.
Products linking upward and sideways, so a product page is not a dead end that only leads back to a grid.
The orphan problem. Common at scale and invisible without looking.
On any large shop some products are reachable only through search or a direct link. They are effectively not part of the site. Finding them is one of the first things block eight looks for.
Auditing An Existing Structure
Six things to look at, in this order, because each one changes how you read the next. This is the practical core of the page for a shop that already exists.
One. Count the pages. How many exist against how many are eligible to be found.
A large gap in either direction is the headline finding. Far more findable than real means block five. Far fewer means something is blocking, which our technical guides cover.
Two. Map the tree. Write down the actual hierarchy, level by level.
Most shop owners are surprised by their own structure once it is written out rather than navigated. Categories nobody knew existed, duplicated branches then orphaned sections all turn up at this stage.
Three. Test it against demand. Per block three, using your own site search and the words your staff hear.
This produces two lists: categories that exist and should not. Categories that should exist and do not. The second list is where the value is.
Four. Measure depth. How many clicks from the home page to a typical product, then to your worst case product.
The worst case matters more than the average, because that is where the catalogue is being lost.
Five. Examine the filters. Per block five. What can be generated, plus whether it is findable.
Six. Find the orphans. Products no internal link reaches.
What the findings mean together. They resolve into one of two situations.
Either the structure is sound and underused, which is a content and linking problem. Or the structure itself is wrong, which is block nine.
What Can Be Fixed And What Needs A Rebuild
The practical divide, so you know what you are facing before spending anything. The line falls in a place that surprises people: it is not about how bad the structure is.
Fixable without changing addresses. More than most shops expect.
Adding new categories that cut across the existing tree. Writing content on categories that have none. Rebuilding internal linking. Bringing orphans into the site. Governing what the filters generate. All of that leaves existing addresses untouched, which means no redirects and no risk to what the shop already earns.
Needs addresses to change. Fewer things than people assume.
Moving products between categories where the address contains the category. Renaming categories. Flattening or deepening the hierarchy itself. Each of those breaks existing addresses, so each needs redirects planned before the change rather than discovered afterwards.
Where the line actually falls. On whether addresses move, not on severity.
A badly wrong tree can frequently be improved substantially by addition alone, because adding the categories customers search for delivers most of the value without touching anything existing. That is the single most useful thing on this page for a shop that feels stuck.
When a rebuild is genuinely warranted. Three cases.
Where the platform cannot support the structure the business needs. Where the existing tree is so wrong that additions would create a confusing hybrid. Or where a replatform is happening anyway, in which case the structure work should shape it rather than follow it.
The order either way. Fix what needs no addresses to move first.
It is cheaper, faster and it tells you how much of the problem was really structural, which our diagnostics guide covers.
Your categories came
from your supplier.
A feed organises by what is convenient to manufacture and distribute. A customer organises by the problem they have. If the first thing a visitor chooses between is a manufacturer name, no amount of content fixes it, because those pages target terms nobody searches.
What is included every month:
£350 per month, one target area. No setup fee, nothing billed separately.
Twenty-two guides.
One subject.
This guide covers structure. The rest of the series covers the sequence, category and product pages, technical health, measurement, the marketplaces and everything a store owner has to decide.
