01 The Situation
Areve is a Belgrade women's atelier. Silk-forward, handmade, sized and finished in-house, with a small showroom on Stefana Prvovenčanog. The product was already there: a designed collection, a working atelier, a brand identity that customers who had handled the pieces in person did not need convincing about. The infrastructure to sell it online was not.
When the engagement began, Areve was operating on a WordPress site that had been built around the brand's editorial side, not its commerce. There was no functioning Meta advertising program. There were no email flows. There was no measurement layer worth the name. Customer acquisition was ambient: referral, walk-in, the quiet gravity of a brand that people in the right Belgrade circles had heard of but most people anywhere else had not.
The brief was unusually clean and unusually total: build a direct-to-consumer business from zero. Site, store, ad account, creative system, email. The entire DTC apparatus, end to end, with the atelier as the only existing asset.
The Market
Areve did not enter a category that was already configured for it. The Serbian women's fashion market is densely populated. Ballary, MYKA, Di Valgoni, dozens of regional players. But most of it operates at price points well below where a handmade silk piece sits, and most of it competes on volume, social proof, and Instagram density rather than on craft. Where silk does exist as a category, it exists at the cheap end. Silkeep, the most visible Serbian silk pyjama brand, retails its hero piece at roughly one-sixth of Areve's equivalent price.
This was not a positioning problem solvable by clearer messaging. It was a structural fact about the market. Premium handmade silk is not a category in Serbia in the way that affordable polyester sleepwear is. The customer who would pay €120 for a handmade silk shirt exists, but she is not browsing a saturated category looking for the best option. She is being convinced, one purchase at a time, that the category is worth entering.
The Constraints
Every decision made across nineteen months was shaped by a set of real constraints, not hypothetical ones.
Production
Areve's atelier produces handmade pieces in small quantities. Inventory is not a
tap that can be opened. Every scaling decision had to respect what the atelier could actually
finish in a given week, which meant that the right answer was rarely 'spend more on Meta'. It
was 'spend better on the units we have.'
Channel
The starting site was WordPress, configured editorially rather than commercially. Migrating to Shopify in May 2024 was not a cosmetic decision. It was the moment the business actually became measurable, and the entire performance dataset that follows in this document is dated from it.
Market
Serbia is a small country with high COD penetration, a customer base that requires more trust per euro of price than a comparable customer in Western Europe, and a Meta ad market where the same audience is being repeatedly targeted by every fashion brand operating in the territory. Acquisition cost compounds quickly. Creative fatigue arrives early. The discipline required to keep MER above 5x in this environment is non-trivial.
Brand
Areve's price point is a feature, not a bug. The strategic instruction was to never compete downward. To never run the discount calendar that the rest of the market runs, never erode the premium signal that the atelier was built on. Growth had to come from the right customer at the right price, or it did not count.
02 The Challenge
The mandate had three parts. Build the channel. Sell at the brand's actual price point, without discounting. Generate enough volume, against the atelier's production ceiling, to make the channel a meaningful contributor to the business, not a marketing expense.
That third requirement was the one that needed the most careful handling. A Belgrade atelier with handmade silk pieces is not a brand that benefits from being scaled into oblivion. The right ceiling for the DTC channel was whatever the atelier could absorb, plus a defensible margin for production headroom. Pushing past that ceiling would have produced backorders, customer disappointment, and the kind of brand damage that takes years to repair. Falling short of it would have left genuine demand unconverted.
The challenge was therefore not a media buying problem in isolation. It was a coordinated build across three surfaces, site, ads, email, that together had to produce a saleable customer experience for a brand that, until May 2024, had no saleable customer experience at all.
03 The Strategic Insight
Two insights, taken together, defined the engagement.
The first was about audience. Premium handmade silk is not a category Serbian Meta users are searching for, but the customer who buys it exists in clearly identifiable patterns. Urban, educated, cosmopolitan, between 30 and 55, employed in fields where presentation matters and disposable income permits a €120 purchase that is not strictly necessary. Meta's targeting infrastructure can find this customer, but only if the creative and the offer can hold her attention long enough to convert her. The strategic implication was that audience size would be the binding constraint, not audience cost. Areve was not going to fail because Serbian Meta CPMs were too high. It was going to succeed or fail on whether the right woman, when she encountered the brand, would buy.
The second was about price. The instinct, in any new DTC build, is to lead with the entry-level item. The lowest friction first purchase that lets the customer try the brand. For Areve, this instinct was wrong. The customer who would pay €120 for a silk shirt was not going to be acquired by being shown the cheapest item in the catalog. She was going to be acquired by being shown the brand at its best. The right hero product was not the cheapest one. It was the one that most clearly communicated what Areve was. Across nineteen months, that meant centering the creative system around the silk pieces, Eden, the kimono, the signature shirts, and treating the entry-level items as expansion, not acquisition.
AOV grew 17% year over year, from €94 to €110, while order volume grew alongside it. The customer was not being trained to expect a discount. She was being trained to expect the brand.
04 The Approach
The work was organized as a build, not as a campaign. Three surfaces, the site, the ads, the email, were stood up in sequence and then operated as a single system. Below is what was actually done across each, in compressed form.
Surface One: The Site
Migration from WordPress to Shopify was the first material decision and the one that made every subsequent measurement possible. The previous site was editorially competent and commercially incoherent. Shopify gave the brand a working store, a checkout that performed, a measurement layer that the rest of the engagement could be built against, and a product catalog architecture that could carry a growing collection without a redesign every season.
The site was built to do the work of the atelier without the atelier being in the room. Product photography was directed to the brand's actual aesthetic standard, not to an e-commerce shorthand. Product pages were written to communicate craft without lapsing into marketing copy. The collection structure mapped to the way the customer thinks about her wardrobe, silk pieces, sets, kimonos, shirts, not to the way Shopify's defaults would organize them.
Surface Two: The Ads
Meta was set up from zero. New Business Manager, new ad account, new pixel, new event configuration, new creative pipeline. Spend started at roughly €1,400 in the first full month and scaled deliberately, never doubling without a clear data signal, never holding flat when the data justified more. The creative system centered on the brand's hero pieces, silk, atelier-craft, the elements of the brand that most clearly differentiated it from the Serbian fashion mass, and was refreshed continuously as performance data accumulated.
The architecture was simple and disciplined. A stable prospecting layer carrying the brand to new customers, a smaller and more aggressive retargeting layer converting consideration into purchase. No detail was left to default. Audiences, placements, creative refresh cadence, attribution windows. Each was set deliberately and monitored against the only number that mattered, which was blended marketing efficiency ratio against actual Shopify revenue.
Surface Three: The Email
Email was the third surface and, in a brand of Areve's price point, the one that quietly does the most work over time. Klaviyo was configured from zero. The welcome flow, the abandoned cart sequence, the post-purchase flow, the win-back flow. Each was built against the brand's actual voice, not the platform's templates, and each was tuned to do the specific job of converting a customer who was already partway down the path. Campaigns were operated on a sustainable cadence, not the saturation cadence that mass-market brands rely on. The brand's tolerance for being in the customer's inbox was lower than a discount-driven brand's would be, and the email program was built to respect that.
Three surfaces, operated as one system. Not around the constraints. Through them.
05 The Results
All figures below are taken directly from Shopify and Meta exports. Revenue is total sales, the top-line figure including taxes and shipping, as recorded in Shopify. Currency is euro at the daily NBS reference rate; all source data is in Serbian dinar.
MER = total sales as a multiple of Meta ad spend. Spend % = Meta spend as a percentage of total sales.
The Build, In Numbers
9 → 416
Weekly Orders
Week one to peak month
€94 → €110
Average Order Value
2024 to 2025
19 months
From WordPress to Peak
May 2024 launch on Shopify
From a non-functioning WordPress site in April 2024 to a Shopify channel generating €52,784 in total sales in November 2025. The first full month of the new channel produced 190 orders. The peak month produced 416. Across the nineteen months from launch to peak, the channel generated 3,865 orders and €397,311 in total sales, on €51,516 in Meta spend, for a blended MER of 7.7x sustained across the entire window.
The brand did not discount its way to these numbers. Average order value rose from €94 in 2024 to €110 in 2025. A 17% increase against a backdrop of compressed Serbian consumer spending and rising Meta ad costs. The customer who arrived through the channel paid more, on average, than the customer who arrived the year before. That is the signature of a brand that earned its growth at price, not by abandoning it.
The Peak Month
November 2025 is the clearest evidence of what the system was capable of producing. €52,784 in total sales. 416 orders. A blended MER of 7.8x against €6,779 in Meta spend. The month was supported by a coordinated November program, strong creative refresh, careful spend pacing into the high-intent window, and an email cadence built around the calendar. But the volume itself was a function of nineteen months of compounding work. The site was ready. The ads were tuned. The email program had a list and a voice. November did not happen because of November. It happened because the previous eighteen months had built a channel capable of
delivering it.
What the MER Tells Us
Marketing efficiency ratio, ad spend as a percentage of revenue, rose from 30% in 2022 to 46% in 2025. A rising MER means the brand was reinvesting an increasing share of revenue into advertising. This was a deliberate choice, made possible by the confidence the LTV model provided. When you know what a customer is worth over 4.5 orders, you can afford to spend more to acquire them. The MER rise is not a warning sign. It is the signature of a brand that understood its economics and had the conviction to act on them.
Growth Without Discounting
The 2024 to 2025 trajectory carries particular weight in a market where most fashion brands grow by training the customer to wait for the next sale. Areve did not. The brand maintained its price floor across the engagement, raised prices in early 2025, and grew anyway. The 17% AOV lift between 2024 and 2025 is not an artifact of the catalog mix. It is the result of a deliberate choice to let the right customer self-select on price, and to build a system capable of finding her at scale.
€397,311 in total sales across nineteen months. From a starting point of zero functional infrastructure. At a 7.7x blended MER, sustained without discounting, in a Serbian market that does not reward premium positioning. These were not favorable conditions. The results are not a product of favorable conditions.
06 What This Means
The Areve case is not a story about a brand that was already scaling and needed more fuel. It is a story about a brand that had everything except the channel, and the disciplined, sustained work of building that channel from zero.
The lessons do not belong to fashion or to silk. They belong to any brand owner with a product good enough to deserve a working DTC channel, who has not yet built one.
Build the surface that converts before the channel that drives traffic. The single decision that defined the engagement was the migration from WordPress to Shopify in May 2024. Without a site that could carry product, accept payment, and produce measurable behavior, no amount of Meta spend would have produced a result worth measuring. Most brands try to fix conversion with media. The right sequence is the opposite. Fix the channel. Then drive the traffic.
Do not lead with the cheapest item. The cheapest item in a premium catalog is not the right entry point for a customer who is being asked to buy into a brand. Lead with the piece that most clearly communicates what the brand is, even if it is also the most expensive, and let the customer self-select on whether the brand is for her. AOV will rise. The customer will be the right customer. The catalog will mature naturally around the hero.
Three surfaces, operated as one system. Site, ads, email. They are not three campaigns. They are three sides of the same channel, and the channel only works when each side carries its share of the work. Most brand owners over-invest in one and under-invest in the other two, and the channel runs at a fraction of its capacity as a result. The Areve case is a demonstration of what happens when each surface is built deliberately, operated continuously, and held to the same standard.