Season

The challenge

Season makes a small range of home goods with genuinely unusual production stories — materials, makers, and constraints that customers care about once they know them. The storefront rendered all of it as a grid of thumbnails, which is the format that hides exactly that kind of difference.

Two problems compounded. The story that justified the price lived three clicks deep, so most visitors never reached it. And the merchandising team could not change a headline without filing a ticket, so the pages that did work were rarely updated and the ones that did not stayed broken for months.

What we found

We ran eight moderated sessions, reviewed twelve months of behavioural data, and sat with the merchandising team through a normal week. The picture was consistent:

• The story converted, and almost nobody saw it. Sessions that reached the production-story content converted at several times the rate of those that did not — but very few reached it.
• Checkout lost people between steps, not within them. Drop-off clustered on transitions rather than on any individual field.
• Merchandisers had no edit path. Every copy change required a developer ticket, so the team had stopped asking.
• Mobile was the majority and the worst experience. Most traffic was mobile, where the thumbnail grid compressed into an undifferentiated column.

The strategy

We reorganised the catalogue around narrative rather than taxonomy. A product page became a single scrolling argument — what it is, who made it, why it costs what it costs — with the purchase decision available at every point in that scroll rather than parked at the end.

In parallel we treated checkout as one continuous step instead of four discrete ones. The fields did not change much; the transitions between them disappeared, which is where the losses actually were.

Three decisions that shaped it

How it was built

Design and front-end ran in parallel with the same team rather than in sequence, so layout decisions were tested against real content and real load behaviour as they were made instead of after handover.

Product imagery is the heaviest part of a storefront, so media handling was decided first: explicit dimensions to prevent layout shift, responsive sources, and lazy loading below the fold. Performance targets were set at the start and checked on every build rather than audited at the end.

How the work ran

The new experience launched on a subset of the catalogue first, ran alongside the old templates for three weeks, then became the default once the merchandising team was comfortable editing without support.

The outcome

The storefront now presents each product as the argument that sells it, and the team that owns merchandising can act on what they learn without waiting for a release. Checkout is a single step, and the pages get better every week because the people closest to the products can finally change them.

What we handed over

• Catalogue and information architecture
• Narrative product page system
• Single-step checkout design and build
• Responsive front-end implementation
• Editable content model for merchandising
• Performance budget and training session

Work with us on something similar

Season drew on the services below. Tell us what you’re building, or see the rest of our work.

• Website Design & Build
• UI/UX & Product Design