Create+
Build+
Grow+
Systems+
Guides+
Updates+
Case Studies+
AboutContact
Start a project
Home/Case Studies/Luxury Marketplace
✦ Case study / Luxury fashion  —  Multi-brand marketplace · Global

The best prices online. Unfindable on Google.

A global luxury marketplace carrying 400+ designer brands — Gucci, Prada, Armani, Palm Angels and hundreds more — across 7+ connected stores, at prices few resellers could match. Everything was in place except the one thing that mattered: shoppers searching for those exact products couldn’t find the store. Two agencies had already tried and failed to say why. We found the cause in days — then rebuilt the engine behind it and took the brand into new regions and currencies. Shared with permission; the client stays anonymous.

Client: Confidential · under NDASector: Luxury multi-brand marketplaceMarkets: North America · Middle East · Turkey · EuropeReferred by: a happy client
Technical SEOGoogle Merchant CenterGoogle ShoppingShopify Grow · Dawn themeEcommerce developmentNext.js / NestJSMySQL · AWS · VercelInventory automationMulti-region · multi-currencyArabic RTL localizationTariff & tax automationAffiliate integrationsN8N automationAI content at scaleOngoing operations
§ 01 — Where they started

Everything right, except being found.

On paper, the store had every advantage. A catalogue of 20,000+ products (80,000+ variants) from 400+ high-value designer brands, live feeds from7+ connected stores, and referral traffic from luxury aggregators likeModeSens and Lyst. Their real edge was price: deals with manufacturers and resellers let them beat almost everyone on the same item.

And yet organic sales barely moved. Not because the offer was wrong — because of how luxury shoppers actually search. Nobody types the store’s name. They find a piece they love on a brand’s own site, copy its SKU, and search Google for that exact code at the best price. Win that moment and you win the sale. The marketplace should have owned it — but it was nowhere to be seen on Google Search orGoogle Shopping. Their in-house teams, based in Turkey and India, had followed the standard guidelines to the letter and still couldn’t say why.

01
Discover
Shopper spots a piece on a brand’s own site or a famous fashion destination.
02
Copy the SKU
They grab the product code — the exact identifier for that item.
03
Search the price
They paste the SKU into Google, hunting for the best price anywhere.
04 · the gap
Should’ve been here
Cheapest in the results — but invisible, so the sale went to someone else.
§ 02 — How they found us

A referral from a happy client.

They didn’t find us through an ad or a pitch. They asked one of our long-standing clients — the heritage couture house in our HSY case study — who they’d trust with a problem nobody else could crack. We got the introduction on the strength of work we’d already done, which is the only kind of referral that matters.

By the time they reached us they were understandably tired of it. Their own teams had followed every standard Google Merchant Center guideline and still weren’t listed. The brief to us was blunt: “Everyone says the catalogue is fine. So why can’t anyone find us?”

The buyers weren’t searching for the store. They were searching for a SKU — and the best price.

§ 03 — The diagnosis

A problem in the seam between two teams.

Our SEO team found it within days. Honestly, it was a small fix — the kind that feels obvious once you see it. The reason it had survived two in-house teams is that it didn’t live cleanly on either side of the usual line. The products weren’t qualifying for free listings on Google Shopping because the site wasn’t meeting the technical criteria Google needs to list them — a development problem dressed as an SEO problem.

Whoever was looking at the website wasn’t looking at the Merchant Center. Whoever was in the Merchant Center never looked at the build. Each side was certain the fault was the other’s. Because our search and development people sit on the same team, they read both at once and the cause was plain. No heroics — just looking in the place no one had thought to look.

It wasn’t an SEO problem or a dev problem. It was the seam between them — where no one was looking.

Before
A multi-million catalogue invisible on Google Shopping · two teams unable to name the cause · stock that couldn’t keep pace
After
Products qualifying & listed within days · operations rebuilt end to end · out-of-stock sales down to zero · three new regional markets
Honest note

We won’t dress this up: the listing fix itself was quick and small. The value wasn’t in the size of the change — it was in being the one team that could see both the code and the search side at once. We share the work, not the client’s name or internal numbers; those stay with them.

§ 04 — Taking over operations

From one fix to the whole engine.

Solving the thing two teams couldn’t earned a bigger remit: custom integrations, SEO,Shopify Grow maintenance, design, advertising and regional strategy — the whole digital operation. The deepest problem was still hiding, though. Once products gained Google visibility, sales surged — and so did out-of-stock orders. The legacy system fetched prices and inventory from 7+ connected stores with just enough delay that an item shown as “in stock” had often already sold out elsewhere. With no unified source of truth, store feeds were fighting each other.

We built a completely new automation layer from scratch, while keeping the live storefront running. We also replaced the site’s paid, poorly-supported theme with a custom build onShopify’s free Dawn theme — faster, fully in our control, and restyled to a true high-end fashion feel: editorial typography, generous space, and a look finally worthy of the brands inside.

Built on
Next.js (frontend) and NestJS (backend), with MySQL, deployed across AWS and Vercel — fast to read, robust to run.
Real-time sync
Scheduled cron jobs keep stock, prices and new products synchronised across every connected store — no more conflicting feeds.
Auto-hide & auto-import
Out-of-stock items are hidden automatically the moment they sell elsewhere; new products from any connected store import themselves, no manual handoff.
Data governance
Each connected store arrived in its own raw, inconsistent format. We built a cleansing & standardization layer so every feed lands in one uniform structure — the foundation everything else runs on.
Zero-downtime
The new engine went live without taking the storefront down — and now absorbs a new connected store roughly every 2 months.
§ 05 — The content problem

Thousands of products. No two the same.

Because the feeds came from stores that supplied the same products to other resellers, the marketplace’s content wasn’t unique — and duplicate copy is poison for search. Over10,000 product descriptions needed rewriting, but products sold faster than any team could rewrite them by hand; by the time a description was updated, the item was often already sold out. A team big enough to keep up would cost a fortune and still be too slow.

So we built the rewrite into the pipeline: a custom AI application on OpenAI’s GPT API that takes a bulk CSV of products — 20,000+ at once — and rewrites every description in minutes, in the marketplace’s own brand voice, in multiple languages for the regional sites. Thousands of listings a week, each one unique, on brand, and live in time to actually sell.

In
Raw supplier copy for every incoming product — the same generic text dozens of resellers share.
Rewrite
A GPT-powered pipeline rewrites each description, tuned to the brand’s voice and each regional language, at feed speed.
Out
100% unique, SEO-optimised content across the entire catalogue, published on time.
§ 06 — Growing into new markets

One storefront, three languages.

With search finally working, we took the store into the Middle East and Turkey — building fully localised regional sites rather than one generic storefront for everyone, with region-specific pricing and payment options.

English
Primary market
North America, priced in USD — the original storefront, now actually found.
Arabic · RTL
Middle East
A complete right-to-left rebuild — native language, Dinar pricing, and regional payment methods.
Turkish
Turkey
Localised content and pricing in Lira, matching the market that already knew the brand.

Two problems came with the growth. First: shifting international tariff rules made landed prices unpredictable for shoppers buying from Europe, China and beyond. We built an automated tariff-calculation layer keyed to country of origin, so the final price — tariffs included — shows up front, with no surprise at checkout. Second: one traffic source was flooding the analytics with volume and never a single sale, dragging the conversion rate down. It turned out to be bot traffic, not shoppers; we blocked it and the numbers went back to reflecting real buyers.

We also expanded the marketplace’s reach through affiliate network integrations — custom API connections into Lyst, ModeSens and Rakuten — and are now finishing EU expansion, including a fresh VAT registration for Germany and new European stores coming onto the platform.

§ 07 — What changed

Findable, honest, and growing.

The store went from invisible to eligible and listed, exactly where its SKU-searching buyers were already looking — on a storefront that finally looks the part. The operation behind it now runs itself: stock stays true to the feeds, out-of-stock sales are gone, new suppliers slot in automatically, pricing accounts for tariffs up front, and every product earns its own copy in its own language. The numbers below describe the scale of the work, not promises of performance — those depend on the business and the market.

0%
Out-of-stock sales, down from a persistent problem
400+
Designer brands, across 7+ connected stores
20k+
Products rewritten & kept unique, in minutes
3
Regional sites & currencies — EN, Arabic, Turkish
§ 08 — In their words
After struggling with basic Google Merchant Center issues for months with other teams, this agency solved it immediately. But that was just the beginning. They’ve transformed our entire operation — from eliminating out-of-stock sales to solving impossible SEO challenges at scale. They don’t just fix problems; they build solutions that grow with us across continents and currencies.
— Operations lead, luxury multi-brand marketplace (client name withheld · NDA)
§ 09 — Who & what it took

One team that reads both sides.

The whole story turns on a single fact: our search anddevelopment people work side by side, not in separate silos. That’s what let one diagnosis cross a line two in-house teams couldn’t — and what keeps the system honest, in three languages, now.

SEO expertTechnical SEODevelopersGoogle Merchant CenterShopify Grow · DawnNext.jsNestJSMySQLAWSVercelOpenAI GPT APILyst · ModeSens · Rakuten
§ 10 — Why it worked

A rare problem, for a rare fit.

Plenty of agencies do SEO. Plenty build stores. Very few do both under one roof well enough to spot a fault that hides in the gap between them — and then go on to run the whole machine across regions and currencies. This marketplace had spent real time and money proving how rare that combination is. When they found it, they handed us everything, and it keeps growing: a new store roughly every two months, and Europe next.

The fix was small. Being the only team who could find it was the whole point.

§ The work behind it

How we’d do it for you.

One marketplace, one team — search and build working together. Each piece is a service in its own right.

§ Start a project

Got a store that should be found — and isn’t? Let’s look in the right place.

Tell us where the sales should be coming from and aren’t — the first reply comes from a partner, not a form.