Industries

Websites for online retailers, built around finding and deciding

An online store is a tool rather than a presentation. Its job is to let somebody find a thing, decide about it and buy it, and almost everything that determines whether it does that well is unglamorous: how filtering works, whether the product page answers the question a photograph cannot, how fast the pages load, and what the checkout asks for. Stores that invest in atmosphere and neglect those tend to look accomplished and convert poorly. Subscription presentation and tax obligations belong with your advisor.

Filtering and navigation carry more weight than anything visual

A customer with a catalog in front of them is trying to narrow it down, and that mechanism is the store's main interface.

Filters are how somebody turns a thousand products into five. If the attributes they want to filter by are missing, or the data behind them is incomplete, the store is unusable at scale.

Attribute coverage is the usual failure and it is a data problem rather than a design one. A size filter that only works on half the catalog is worse than none, because it silently hides products.

Filtering on a phone needs its own treatment rather than a compressed version of the desktop panel, since it is where most of the traffic actually is.

The interaction between filters and search matters too. Somebody who searched and then wants to narrow should be able to, and in many stores those are separate systems that do not cooperate.

This is also where the technical side meets the front end, since the same filter combinations that help customers can generate enormous numbers of indexable addresses if handled carelessly.

The product page answers what the photograph cannot

Somebody is buying something they cannot touch, and the page has to close that gap.

Scale is the most common unanswered question. Photographs on white backgrounds give no sense of size, and a product shown in context or beside something familiar prevents a return.

Materials, dimensions, what is included, compatibility and care are the specifications people actually look for, and they are frequently incomplete or buried.

Multiple images doing different jobs beat many images doing the same one. Scale, detail, in use, and what the packaging looks like.

Reviews and customer questions belong on the page rather than behind a tab nobody opens, since they answer things the description does not and they are the most trusted content there.

Availability and delivery expectation should be visible at the point of decision. Somebody who wants it by a date is making that decision here, not at checkout.

Speed is revenue here more directly than anywhere else

Every step of the path is a page load, and every load loses a share of the people on it.

A retail path involves many pages: category, filters, product, cart, checkout. The compounding effect of slowness across that sequence is larger than on a site where somebody views two pages.

Images are the usual weight and they are also the product, so the answer is proper handling rather than fewer of them: correct sizing, modern formats and loading only what is needed.

Third-party scripts accumulate quietly. Analytics, chat, reviews, personalization and tag managers each add weight, and stores acquire them over years without anybody removing any.

Mobile performance is the one that matters, since most retail traffic is on phones and frequently on connections nothing like the office network the store was built on.

Test the actual path on a real phone rather than the home page in a lab. The checkout on a slow connection is where the money is, and it is rarely what gets measured.

Checkout is a design problem with real constraints

It is the most consequential sequence on the store and the one with the least room for creativity.

Guest checkout, minimal fields and payment options people actually use are the established fundamentals, and stores still get them wrong regularly.

Address entry is worth particular attention, since it is where errors and frustration concentrate. Validation that rejects legitimate addresses is a real and common failure.

Progress and cost transparency reduce anxiety. Somebody should know how many steps remain and what the total is before the final click.

Error states deserve design rather than defaults. A form that clears entries on an error loses a customer who had already decided to buy.

Everything here is measurable at a step level, which means arguments about checkout can be settled with the store's own data instead of with opinion.

Subscriptions have to be presented, not slipped in

Where an offer creates a recurring charge, how it appears is subject to real requirements.

Subscription and auto-renewal disclosure is regulated, and negative-option billing has drawn particular enforcement attention.

The pattern under scrutiny is exactly the one that converts best in the short term: a subscription pre-selected, a renewal price shown less prominently than the introductory one, and cancellation that is harder than signing up.

Design choices are where this becomes concrete. What is selected by default, what appears at the point of agreement, and how the confirmation restates the terms.

Cancellation deserves the same care as signup. A path that is materially harder than the one in is the specific practice regulators have addressed.

This belongs with your advisor rather than with an agency's judgment. Our part is making sure whatever they establish is presented clearly and where people actually see it.

What the platform decides and what you decide

Much of a store's behavior comes from the platform it runs on, which is worth knowing before commissioning design work.

Checkout structure, address validation, search behavior and how variants work are frequently platform decisions rather than design ones, and some are not changeable at all.

That is worth establishing early, since a design that assumes control over the checkout on a platform that does not offer it produces a rebuild nobody budgeted for.

Where a platform's defaults are the constraint, the work moves to what surrounds them: the product page, the navigation, the content and the data quality.

Migration is a serious undertaking with real search consequences and should be a considered decision rather than a reaction to one limitation.

The parts you always control are the product data, the imagery and the information architecture, and those are usually where the largest available improvements sit anyway.

Questions we actually get

Where should we invest first in a store redesign?: Filtering and product data, usually. Filters are how somebody turns a large catalog into a short list, and incomplete attribute data makes them silently hide products. That is a data problem rather than a design one and it is where most stores are weakest, well ahead of anything visual.
What is missing from most product pages?: An answer to scale. Photographs on white backgrounds give no sense of size, which drives both hesitation and returns. Beyond that: complete specifications, what is included, compatibility, and reviews and customer questions on the page rather than behind a tab nobody opens.
How much does speed actually matter?: More than on most sites, because a retail path is many page loads and the effect compounds across the sequence. Images are the usual weight and also the product, so the answer is proper handling rather than fewer. Third-party scripts accumulate over years and nobody removes any. Test the checkout on a real phone.
Can we redesign our checkout?: Partly, and how much depends on your platform. Checkout structure, address validation and variant handling are frequently platform decisions rather than design ones, and some cannot be changed. Establish that before commissioning work, since a design assuming control you do not have produces an unbudgeted rebuild.
How should we present a subscription option?: Clearly, and with your advisor's input rather than a designer's judgment. The pattern that converts best in the short term, meaning pre-selection, a faint renewal price and cancellation that is harder than signup, is precisely what regulators have addressed. What is selected by default and how the confirmation restates terms are the concrete decisions.

What is different here

Two areas set the constraints. Advertising claims, subscription and auto-renewal disclosure, and how reviews are gathered and displayed are regulated federally and increasingly at state level too, with negative-option billing drawing particular enforcement attention. Separately, selling into many states raises sales tax collection questions that turn on where a business has nexus. Neither is a marketing decision, so both are worth settling with an advisor before spend scales.

Other disciplines for online retailers

Worth reading alongside this

Written by KC Thompson, Morgul Marketing. Updated September 6, 2026.

Talk to us about web design for online retailers

Tell us the market you work in and what the account looks like now. We will say where we would start, and what we would not touch yet.