What Makes a Good Ecommerce Website?
One test runs through this entire page. Can a stranger find what they want, decide on it and buy it without needing to ask anybody a question. Everything that does not serve that is decoration.
The Test
Can a stranger find what they want, decide on it and buy it without needing to ask anybody a question? Three stages, each of which a shop can fail independently. Every block below applies the same test to a different part of the site.
Two words in that sentence do most of the work. Stranger matters because the people judging a shop are the people who built it. Everything is findable to somebody who already knows where it is. The owner knows the catalogue. The designer has seen every page. Neither can experience the site as somebody arriving for the first time from a search result, halfway down a page they did not choose.
Without asking matters more, because it is where the measurement problem lives. If a shopper cannot work something out, they do not send an email. They leave, leaving no trace beyond one more session that ended without an order. The enquiry you never receive is the failure you never see, which is why shop owners are so often confident about problems their customers experience daily.
What follows works through finding, choosing, believing, paying and what happens afterwards, judging each against that one question rather than against taste.
Can They Find It
Three mechanisms do this work. They fail in different ways.
Navigation. The commonest fault is that a menu reflects how the business is organised rather than how customers think. Departments, suppliers and brand hierarchies make sense internally. A customer is thinking about a problem, an occasion or a room. They look for the word they would use rather than the word on your purchase orders.
Category structure. The largest single lever on a shop and the one taken apart in how to design ecommerce category pages.
On-site search. Neglected almost everywhere, because it looks like a small box rather than a feature. That is a mistake for a specific reason: somebody using the search box has already decided what they want. They are the most committed visitor on the site. A search that returns nothing sends the most decided customer away.
There is a second reason to care. It is the one nobody acts on. Your search logs are free research and almost no shop reads them. They contain the words customers actually use, which are rarely your product names. They show things people expect you to stock and you do not. They show misspellings worth accommodating. And they show demand for categories you have never created, which is a shortcut to knowing what to build next.
Can They Choose
Finding the product is not choosing it. Choosing requires information most shops do not publish, because they publish what the manufacturer supplied instead.
What actually decides a purchase. Dimensions and weight, since the real question is whether it will fit and whether one person can carry it. Materials and construction. Compatibility, meaning whether it works with what the customer already owns. What is in the box, which is the source of a great many returns. Size and fit, where a general chart is not the same as knowing whether this item runs small. Care and maintenance.
Why manufacturer copy fails. It was written to describe a product to a retailer rather than to answer a buyer's question. It is marketing language. Marketing language is designed to make something sound appealing rather than to resolve a doubt. There is a second problem: every competitor stocking that product has the identical words, so nobody has said anything the others have not.
The test that fixes this is available to every shop and costs nothing. What do customers ask you before they buy? Those questions, answered plainly on the page, are the most valuable content a shop can carry, because each one is a doubt that was stopping somebody. This is developed in how to design ecommerce product pages that convert.
Do They Believe You
A first time visitor entering card details is answering one question, whether or not they would put it this way. Will these people still exist in three weeks when the parcel has not arrived?
Every trust signal on a shop is an answer to that question, which is why the useful ones are the mundane ones.
What makes somebody hesitate. No real address anywhere, only a form. No telephone number, perhaps one that is clearly never answered. A returns position that is either absent or written to discourage. Delivery cost revealed at the final step, which reads as concealment whether or not it was. Reviews that are all five stars, all short and all phrased the same way. A payment step that looks unfamiliar or throws the shopper onto a page that does not resemble the shop. Small signs of neglect, such as a copyright line several years out of date.
None of that is about design in the visual sense. A beautifully presented shop with no address and vague returns loses to a plain one that states where it is and what happens if the item is wrong.
The pattern worth noticing is that trust is built by specificity. A named town, a stated number of days, an actual person in a photograph, a review that mentions a detail nobody would invent. Vagueness is what reads as risk. Detail is cheap.
Is It Fast Enough
Speed belongs to the test twice over. A shopper who is waiting has not yet found anything. A shopper waiting on a phone with a weak signal frequently stops waiting.
The part worth knowing here is that an ecommerce site is slow because of decisions rather than bad luck. Images uploaded at full size. A theme carrying features nobody uses. Third party scripts. Apps, each adding its own code, accumulating quietly until nobody remembers what any of them do.
That matters for the test because it is a failure that hides. Nobody who left because a page took too long is in your enquiries. Testing your own shop on an office connection will not reproduce what a customer experiences on a train.
The whole subject, including what can be fixed afterwards and what needs a rebuild, is in why ecommerce website speed affects sales and rankings.
Can They Buy Without Friction
By this point the shopper has decided. They want the thing, they believe you and they are trying to give you money. Everything that happens now can only lose the sale, which makes checkout the least forgiving part of a shop.
Where it goes wrong is well established and rarely fixed. Being required to create an account before buying, which asks for commitment from somebody who has not yet received anything. Forms longer than the transaction warrants. Costs appearing at the last step that were not visible earlier. Delivery options that do not say when anything will arrive. And error handling that rejects an entry, then clears what was typed, which is the point at which a tired person on a phone gives up.
One framing is worth carrying into any checkout review. Every field is a request. Every request needs a reason the shopper would accept. A delivery address is obviously reasonable. A telephone number may be, if you say why. A question about how they heard of you, positioned between them and their purchase, is not.
The detail, including what is recoverable and what to look at first, sits in how website design affects ecommerce conversion rates.
What Happens Afterwards
The test ends at buying, which is where most shops also stop thinking. That is a commercial mistake rather than a courtesy one.
What a customer needs after paying. Immediate confirmation stating what was ordered and when it should arrive. A dispatch notice when it leaves. Tracking that works. And a returns process a customer can start without composing an email to explain themselves.
A shop that goes silent after payment produces a specific and expensive sequence. The customer becomes uncertain. Uncertainty produces a where is my order enquiry, which costs somebody's time to answer. The order arrives eventually and the experience is remembered as anxious rather than as good. That customer does not return. The second sale is the one that pays for the cost of getting them in the first place.
What makes this worth attention is the economics. Post purchase communication is the cheapest marketing a shop has, because the recipient is already paying attention. They are waiting for a parcel and they will open a message about it, which is not true of anything else you send them.
Applied to the test: a customer who had to ask where their order was is a customer who needed to ask a question. The shop failed, slightly later than usual.
Accessibility
Accessibility means the shop works for people who cannot use it in the standard way. It belongs in this page rather than in an appendix, because it is the same test applied to more customers.
What it means in practice for a shop. Everything reachable and usable with a keyboard rather than only a mouse. Text with enough contrast against its background to be read comfortably. Form fields with labels that stay visible rather than disappearing once typing starts. Meaningful descriptions on product images. Information never conveyed by colour alone, which matters for stock status and error messages. Captions on video.
Who it affects. A much larger group than people assume, because it is not only permanent impairment. It includes somebody with a broken arm using one hand, somebody reading a screen in bright sunlight, somebody on a train without sound and anybody whose eyesight has changed with age, which is eventually everybody.
The commercial argument follows from that. Almost every accessibility fix improves the shop for people who were not the reason it was made. Clear labels, visible focus, sufficient contrast and descriptive text all help somebody buying in a hurry. The same descriptive text is what allows any automated system to understand what a page contains.
Accessibility also carries obligations under UK law. This page is general information rather than legal advice. Anybody wanting to know where their business stands should take proper advice on their own position.
What Good Design Gets Confused With
Five things that look like quality and work against the test. None of them is forbidden. Each has to earn its place against the question of whether a stranger can find, decide and buy.
Awards. Judged by designers on appearance, by people who are not trying to buy anything. A shop can win one and sell less than the plain competitor beside it.
Animation and scroll effects. Content that arrives as you scroll delays the thing somebody came for. On a phone it frequently arrives late or not at all.
Novelty navigation. Hidden menus, unlabelled icons and horizontal scrolling. Every one of them asks a stranger to learn something before they can look for a product.
Full screen video and enormous hero images. They push everything you sell below the fold, they carry weight and they answer nothing.
Minimalism past the point of usefulness. Removing labels, prices, stock information or delivery detail because the layout looked cleaner without them. The clean version is the one that made somebody ask a question.
The common thread is that all five are judged by looking at the site rather than by using it. What separates appearance from experience is covered in what UX design is and why it matters for ecommerce.
Judged on whether
strangers can buy.
We design against the test on this page rather than against taste. We will also tell you where an existing shop fails it. Structure, category planning and search architecture are decided before anybody opens a design tool.
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 covers the standard. The rest of the series covers what a build involves, what it costs, which platform suits which business, where conversion is lost and why the search ceiling is set during the build.
