How Does Web Design Affect Ecommerce SEO?
A shop's search ceiling is set during the build. Structure, addresses, filters and linking are design decisions. Undoing them later means a rebuild rather than a fix. That is why the design work and the search work cannot sensibly be bought separately.
The Ceiling Is Set At Build Time
The maximum search visibility a shop can reach is largely determined in the first fortnight of its build, by people who may not have been thinking about search at all. Everything done afterwards operates underneath that ceiling.
Take one decision as an example. A shop puts every product one level below the homepage, with no subcategories, because the catalogue was small when it was built and the structure looked clean. That shop has decided it will compete for a handful of broad terms and nothing more specific, because it has no pages for anything more specific to be about.
Changing it later is not an adjustment. Every new category needs an address, every old address needs a redirect, internal links have to be reworked and there is a period during which visibility falls while the change settles. That is a migration. It costs more than building it correctly would have.
Now compare that with a design somebody dislikes. A colour, a layout, a product template that turned out awkward. All changeable in an afternoon, on a live shop, with nobody outside the business noticing.
The asymmetry is the whole argument. The decisions taken first, fastest and with least discussion are the permanent ones. The decisions everybody argues about are the reversible ones. And by the time anybody is engaged to improve a shop's search performance, the permanent decisions have already been made by somebody who was never asked about them.
Site Structure
Structure is how categories nest inside each other and how far a product sits from the front door. It is the first decision made and the most expensive to reverse.
Why depth matters twice. To anything crawling the site, pages far from the entry points are reached less often and treated as less central. To a shopper, every additional level is another decision and another opportunity to give up. Both pressures point the same way, which is unusual and convenient.
What a sound structure looks like. Broad and shallow rather than narrow and deep. Categories named as customers describe things rather than as the business files them or as a supplier's catalogue divides. Enough subcategories that each one corresponds to something people actually search for. Few enough that each contains a worthwhile number of products.
How to judge one you already have. Four questions, answerable this afternoon.
How many clicks does it take to reach a typical product from the homepage? More than three or four and the lower catalogue is being neglected by everybody.
Does every category correspond to something somebody would type? A category existing because the business is organised that way has no demand behind it.
Does every category contain enough products to be worth landing on? An almost empty category converts a search into a disappointment.
And the most revealing: are there products reachable only through the search box or a filter? Those products are effectively invisible. They are usually numerous, yet almost no shop owner knows they exist.
URL Structure
The pattern your web addresses follow is decided once, usually by accepting whatever the platform proposed. It is close to irreversible afterwards.
The options and what each does over time. A flat pattern, where a product sits directly beneath the domain, is simple and survives a product being recategorised without any address changing. A category path, where the address reflects the hierarchy, expresses context and breaks the moment a product moves between categories. An identifier based pattern is robust against everything and means nothing to a human reading it.
Why the decision is close to permanent. Once a shop is live and other people have linked to it, every address change requires a redirect. Redirects accumulate, some external links are never updated and each change carries a period of instability. A shop that changes its address pattern twice in five years is carrying the cost of both.
What actually decides it at build time. One question that nobody asks: will products move between categories? If they will, because the catalogue is seasonal or the ranges get reorganised, a path based pattern will generate address changes every time somebody tidies the catalogue. If they will not, the context a path provides is worth having.
Most builds settle this by accepting a default, which means the answer was chosen by whoever wrote the platform rather than by anybody who knows how your catalogue behaves.
Filters And Faceted Navigation
Row by row this is covered from the category page's point of view in how to design ecommerce category pages. What belongs here is the shop wide arithmetic, because that is what makes it a build decision rather than a page decision.
The count is unrelated to your catalogue size. This is the part that surprises people. The number of addresses a shop can generate has almost nothing to do with how many products it sells. A shop with five hundred products, forty categories and generous filtering can produce more addresses than a shop with fifty thousand products and none.
It does not require a large catalogue to become serious. As soon as more than two filter dimensions can be combined freely, the total climbs beyond anything a business would have chosen deliberately. Shops assume this is a problem for large retailers. It is a problem for anybody who offers size, colour and brand together.
What has to be decided before anything is built. Which filtered views are worth existing as pages in their own right, because they match how people genuinely search. Which must work perfectly for a shopper while never becoming pages. And what happens to the combinations nobody anticipated, which is most of them.
That first decision needs demand data, which means somebody has to look at how customers actually search before the filters are configured. Doing it in that order is the difference between a filter set that earns pages and one that generates them.
Pagination And Product Listings
A catalogue is discovered by following links. For most shops the only links to most products are on category listings. That makes how those listings work a discovery decision rather than a presentational one.
The consequence of getting it wrong. Where products beyond the first view are only reachable by pressing something or by scrolling, a portion of the catalogue may not be reachable by anything that does not press or scroll. On a large category that portion is most of it. The products exist, they are in the database, they are purchasable. Nothing knows they are there.
The requirement rather than the method. Every product should be reachable by following ordinary links from somewhere, whatever the shopper facing experience does on top of that. Both things can be true at once. Naming that as a requirement at build time is the entire job here.
The structural answer nobody reaches for. If a category is large enough that products sit hundreds of positions down a listing, the real problem is the category rather than the pagination. Dividing it properly puts those products within reach of both shoppers and crawlers, which is a structure fix rather than a technical one. It takes you back to block two.
Internal Linking
Three kinds of internal link exist on a shop and most shops have only one of them working.
Navigational links. Menus, breadcrumbs and footers. They arrive with the theme and connect upward and downward. They treat every category identically because a template cannot know which of your categories makes money.
Contextual links. Links inside written content, pointing at the categories and products the content is about. These are the ones that carry meaning, because a human decided this page relates to that one.
Merchandising links. Related and complementary products, plus connections between categories that customers associate even though the catalogue does not.
Why automatic linking is not enough. A theme produces a structure in which everything connects to the homepage and the homepage connects to everything. That expresses no judgement at all. If every page is linked the same way, nothing indicates which pages the business considers important. Importance is precisely what internal linking exists to communicate.
How a flat structure leaks. Value arriving from outside lands somewhere, usually the homepage or an article. It then either flows onward to the pages that need it or stops. A shop whose content links nowhere in particular has built a structure where everything arrives and nothing moves, which is covered from the other direction in what pages every ecommerce website needs.
Why Most Shops Are Built Badly For This
Almost every shop we are asked to look at has at least two of the problems above. That is too consistent to be explained by individual failings. The actual explanation is a sequence that nobody designed.
Design agencies are judged on how it looks. Portfolios are visual. Pitches are visual. The client reaction that wins the work is a reaction to a picture. No agency has ever been appointed because of its address structures, so no agency is under pressure to be good at them.
Platforms default to what suits the platform. Defaults have to work acceptably for every shop using the software, which means they are chosen for the average case rather than for yours. They also favour ease of setup over structural fit. Accepting them is not laziness. It is what defaults are for. It is also why they need a deliberate decision to override.
Nobody is asked about search until after launch. The usual sequence is design, build, launch, then engage somebody to improve visibility. That person arrives to find the structure fixed, the addresses live and linked to, plus the filters already generating whatever they generate.
Notice that nobody in that story behaved badly. The designer designed, the developer built, the platform did what platforms do and the search supplier was engaged at the point businesses normally engage one. The failure is in the order, not in the people, which is also why it repeats so reliably across otherwise well run projects.
What Can Be Fixed Later And What Needs A Rebuild
If you have an existing shop, this is the block that tells you which situation you are in.
Fixable without disruption. Page titles and descriptions. Headings. Content on category pages. Images and their descriptions. Most internal linking. Structured data. Most speed problems. All of these can be improved on a live shop, incrementally, without anybody outside noticing anything except that things get better.
Requires a migration. Category structure. Address patterns. How filters generate addresses. The architecture of page templates. Each of these changes what exists at which address, which means redirects, a settling period and a genuine risk of losing ground before regaining it.
The test that separates them, which takes one second. Does the change alter web addresses? If no, it is maintenance and it can start on Monday. If yes, it is a migration and it needs planning, a reason and a tolerance for temporary disruption.
That test is worth having when somebody proposes work on your shop, because it converts a vague conversation about improvements into a clear one about risk. It also tells you when a rebuild is genuinely warranted rather than being the easiest thing to sell, since a shop needing three items from the second list is a shop where doing them separately costs more than doing them together.
Where The Search Work Picks Up
Everything above is a decision taken during a build. None of it is ongoing work. None of it is what most people mean when they talk about doing SEO.
What follows the build is a different discipline with a different rhythm: understanding what customers search for and how that changes, writing and improving the content on category pages, earning references from other sites, watching what competitors do, then responding to how demand moves through a year.
That work belongs in our ecommerce SEO guides. It is continuous where this is structural. It is also considerably more effective on a shop built with the decisions above taken deliberately, which is the entire reason this page exists.
The practical version of doing it in the right order is set out in how to build an ecommerce website with SEO built in from the start.
Structure first.
Then design.
Category structure, address patterns and filter handling are decided before anybody opens a design tool, because those are the decisions that cost a migration to reverse. Design and search structure are one piece of work here rather than two purchases.
What a build covers:
Ongoing management afterwards sits inside our fixed monthly fee. Build work is quoted per project.
Twenty guides.
One subject.
This page carries the central argument of the series. The rest covers what a build costs, which platform suits which business, how category and product pages should work and where conversion is lost.
