3linez is a Saudi abaya label. Its storefront ran on an off-the-shelf Salla theme that sold clothes but said nothing about the brand. The rebuild adds no production budget — same phone, same locations, same faceless content — and fixes the direction, the checkout and the community instead. This is what changed, and why.
The live site is a default Salla build. It works as a transaction pipe, but three things undercut a brand charging premium prices for considered garments:
Saturated pink field, system Arabic UI font, rounded theme components, SKU-coded product names.
Warm paper ground, ink text, one soft-pink accent, square corners, serif display against a geometric sans.
The brief was never "spend more on production". Everything the brand already does stays exactly as it is: shot on a phone, in the cafés and streets its customers actually sit in, faceless by choice. No studio, no photographer, no model day rate, no agency.
That's worth stating plainly, because it defines what the redesign is allowed to solve. Three levers cost nothing but attention:
Same phone, same locations — but consistent framing, natural light, a fixed palette, and a rule for what a 3LINEZ image is.
Creators already produce this content. Giving each one a trackable edit turns scattered tagging into a measurable channel.
A store that feels like a brand — and a wall the customer appears on — earns the repeat visit that a promo code only rents.
Faces stay out of frame by preference and convention. Treated as a problem to work around, that produces awkward crops and reads as amateur. Treated as a rule, it becomes a recognisable visual language: the drape, the fabric, the doorway and the light carry the image, and the shopper projects herself into it rather than comparing herself to a model.
It is also the cheapest possible route to looking editorial, because it forces composition to do the work that a face would otherwise do for free. The design system is built to support it — full-bleed imagery, no decorative frames, type that sits on the photograph rather than beside it.
Rather than copy one site wholesale, the redesign borrows structure from one reference and surface from another.
Rhode's homepage is a masterclass in low-friction DTC flow: a bestseller rail before the full grid, always-visible add-to-bag, ratings on the card, and a cart that tells you how close you are to free shipping. That skeleton was lifted and adapted.
Beauty of Joseon proves a brand can feel premium and warm without colour noise: off-white paper, near-black type, square corners, hairline rules, and photography doing the emotional work. That restraint became the design system.
The tension to resolve: the brand's own personality is young, Saudi and pink. The job was keeping that warmth as an accent inside a restrained system — not deleting it, and not letting it run the page.
A saturated background competes with every garment photographed against it. Moving to a warm off-white lets the clothes carry the colour, and reserves pink for the logo, price-adjacent accents and the WhatsApp button — so it signals the brand instead of shouting.
The name means "three lines". Replacing the numeral with three slashes makes the mark say its own name — legible at any size, works as a favicon, and gives the brand a repeatable graphic device.
Carousels split attention and most visitors never see slide two. A single full-bleed photograph with a place line and a single call to action sets the tone in one beat — the way a fashion house opens, not a catalogue.
Hover doesn't exist on touch, and most GCC traffic is mobile. Making the action permanently visible removes a discovery step, and surfacing star ratings on the card answers the trust question before the click rather than after.
In this market a large share of orders close in WhatsApp, not a checkout form. Every surface can hand off to a pre-filled message containing the item, size and total. The cart also shows how far the basket is from free shipping — a nudge that raises order value without a discount.
Abstract mood buckets ask the shopper to self-diagnose. Shopping through named creators with distinct styles is how this audience already discovers clothes, and it doubles as the affiliate structure — each edit is attributable and trackable.
Modest fashion is bought on drape and fit, which studio shots flatten. Letting customers post their own photos — on the homepage and under the specific product — shows the garment on real bodies and gives the brand a content engine it doesn't have to produce.
Try-on matters for a garment sold on fit. Each boutique card carries opening hours, a maps link and its own WhatsApp thread, closing the gap between browsing and walking in.
Saudi provenance is the brand's real differentiator against imported fast fashion. Stating it in the hero, and repeating it as a quiet "Made in Saudi Arabia" line, turns a fact into positioning.
A single pink does the logo, the chat button and the live counter. Restraint is what makes the accent readable as a brand signal.
Square corners, hairline rules and flat grounds exist so nothing frames or decorates the imagery. The clothes are the interface.
From any screen — card, product, bag, store — the shopper is one tap from a pre-filled conversation, because that's where this market closes.
This is a working front-end prototype, not a deployed store. Being specific about that matters more than a tidier story:
| Area | In the prototype | Production would need |
|---|---|---|
| Catalogue | Six products in a local array | Salla or Shopify catalogue API |
| Cart | In-memory React state | Persisted cart + inventory checks |
| Checkout | Composed WhatsApp message | Mada, Apple Pay, Tabby, COD |
| Photo uploads | Local preview only | Storage plus a moderation queue |
| Photography | Placeholder shots, faces in frame | Re-framed to the brand's faceless rule |
| Language | English | Arabic RTL as the primary locale |
| Performance | JSX compiled in the browser | Pre-built bundle, image CDN |
No performance claims are made here. Nothing in this case study is backed by live traffic or A/B testing — the reasoning is drawn from category convention and market behaviour, and would need validation against real analytics.