The Oleum Crete case study tells the story. This piece is the mechanics: the changes, in the order we made them, with what each one was for. It is written so a founder with a good product and an underperforming store can go down the list and tick off what theirs is missing.
The headline, for context: a single-estate Cretan olive oil sold in Dubai from a generic Shopify template; after the rebuild, monthly sales on the new storefront reached eighteen times the old template's, with no paid advertising. Every order came from search, AI assistants, WhatsApp, Instagram or people typing the address.
1. Decide what the product is worth before touching the site
The store was priced at entry-premium for an oil that belonged next to the best in the market. The whole project was scoped to earn a higher price, not to sell more of a cheap one. Everything below traces back to that decision. If your product is under-priced, a redesign that keeps the price is a nicer shelf for the same margin.
2. Go headless so the storefront can tell the truth
A better theme would have been faster and cheaper, and it would still have been a theme. The pieces that justify a premium bottle, a film, an origin story, verifiable certification, an editorial product page, do not fit in theme blocks. The storefront was built as a custom headless app on Shopify Hydrogen, on Shopify's edge network, with Shopify still owning the catalogue, cart, checkout and payments. The client's operations did not change by a click.
3. Put the proof where the decision happens
PDO, EU Organic, FSSC 22000 and Kosher certificates open as the real documents in a same-page overlay, placed directly above the product grid. An interactive origin map draws the estate's real export reach. The family's legacy film was re-mastered from the original 4K source rather than reshot. Proof is the last thing a visitor sees before shopping, and it is verifiable without leaving the page.
4. Make the choice obvious, then get out of the way
Five bottles is enough to confuse a first-time buyer, so each carries a role: for the table, the connoisseur's pick, the gift, for the kitchen. The product page puts the buy moment above the fold and folds tasting notes and sourcing into accordions. From there a wallet-pay customer places an order in three taps: add, checkout, Apple Pay or Google Pay. Guest checkout is the default, the welcome code carries through automatically, and the cart shows progress toward free delivery to earn a second bottle. Checkout runs on the brand's own domain, in AED.
5. Meet customers where they actually talk
A WhatsApp concierge button sits on every page with the number pre-loaded. In the UAE that is the channel; email forms are where enquiries go to die.
6. Make it fast on the phone it will be bought from
Before launch the mobile hero went from 209KB to 42KB, an 80% cut, and the mobile page from eighteen screens of scrolling to thirteen. Lighthouse best practices went from 92 to 100. Premium buyers in the UAE shop on phones; a beautiful slow page is a slow page.
7. Be readable by search engines and by AI assistants
Every page is server-rendered with full structured data: Organization, Product, FAQ, Breadcrumb, Article. The crawlers behind ChatGPT and the other assistants are allowed in. Per-product titles, canonicals and Open Graph shipped with the build, not after it. This is the change that keeps producing new customers months later: when someone asks an assistant where to buy premium olive oil in Dubai, the store is a brand the machines can read.
8. Collect reviews automatically and show them
Reviews are requested after every order and shown on the product pages, and the ratings feed structured data so they appear as stars in search. Five reviews arrived in the launch month, all five stars. Reviews are the cheapest conversion lift there is, and most stores never ask.
9. Launch with the customers you already have
The launch went out with a film on Instagram and a 30% welcome offer with a hard deadline. But the result that surprised everyone came from a quieter move: the roughly thirty people who had bought before the relaunch were messaged one to one on WhatsApp, with the offer as a thank-you. Six of the seven launch-month orders came from that list, and the overwhelming majority of launch-month revenue came from people who already trusted the brand. Before you pay for a single cold click, talk to the people who already bought.
What happened after
The launch spike settled, the reviews accumulated, search and AI assistants started sending buyers, and WhatsApp kept the past customers close. Two months on, monthly sales were eighteen times the old template's, before the first paid campaign was switched on. Three things did most of the work: the higher price point the storefront now earns, the buying experience that converts the visitors who arrive, and discoverability, which is where the new customers keep coming from. Paid search was scoped for after that foundation, on a modest monthly media budget, targeting a return the store could now support.
The checklist for your store
- Is the price right for the product, or is the site apologising for it?
- Can the storefront show the proof a premium buyer needs, or is it a theme with your logo?
- Is the buy moment above the fold, and is checkout three taps for a wallet-pay customer?
- Is WhatsApp one tap away on every page?
- How heavy is the mobile hero, and how many screens is the page?
- Is every page server-rendered with structured data, and are AI crawlers allowed?
- Are reviews requested after every order and shown where the decision happens?
- Have you messaged your existing customers personally before buying any traffic?
- Is paid search planned for after the store converts, not before?
If your site undersells what you make, the first call is free and the written point of view that follows will say which of these nine to pull first.
