Case Study

Scaling Personalization Across 12 European Markets: A Decade of E-Commerce Growth with Miinto

Segment

Apparel, Fashion

The Results

Martin & Servera validated Retail Media as a scalable revenue stream using Algonomy’s Recommend™ and DSW, without investing in a costly point solution.

From September 2025 through April 2026, the program delivered strong outcomes across all key metrics.

Placements closer to the point of purchase, such as the search results page, consistently delivered the highest engagement and conversion rates.

0 %

Average Purchase Rate

0 %

Revenue attributed to Discover across core Europe

0 X

Incremental revenue from Recommend™ in Sweden

0 +

Markets across Europe

The Client

Miinto is one of Europe’s largest fashion marketplaces, connecting shoppers with boutiques and brands across 12 markets, from the Nordics to Southern Europe. At that scale, personalization isn’t a feature decision. It’s an infrastructure one.

For almost a decade, Miinto has run Algonomy’s personalization suite across its full market footprint, using Recommend™, Discover, and Engage to cover every meaningful moment in the shopper journey, from product recommendations to discovery and homepage content.

Twelve markets. Three products. One consistent personalization layer.

The Challenge

Different languages, different shopper behaviors, different catalog structures. Most personalization tools break down under that complexity. The usual fix is to customize market by market, which creates inconsistency and makes it hard to improve anything at scale.

Miinto’s approach, built on Algonomy’s personalization suite, takes a different path.

Rather than managing each market separately, Miinto runs a single, unified personalization stack across all twelve. A unified stack across all markets means the same logic, data infrastructure, and optimization levers, regardless of geography. What gets refined in one market strengthens the others. What works in Stockholm gets tested in Warsaw.

The Solution

1 Discover: Turning Exploration Into Revenue

For shoppers who arrive without a specific destination, Discover surfaces personalized product feeds that turn browsing into buying.

Personalized Accessories PLP based on the customer’s affinity towards sunglasses.

Revenue attribution from Discover, the share of total sales traceable to a discovery interaction, reached 33.65% in Belgium and 31.37% in Norway in a single year in 2025.

Across the mid-tier European markets, including Poland, the Netherlands, France, Italy, and Germany, attribution consistently sat between 22% and 27% that same year.

Even in earlier-stage deployments in the UK, Discover accounted for 17.46% of revenue.

The consistency across markets at different stages of maturity is what the almost-decade of optimization delivers. Discovery isn’t a feature that works for Miinto’s best market. It works for all of them.

2 Recommend™: Closing the Loop at the Conversion Moment

Where Discover captures exploratory shoppers, Recommend™ works at the product and cart levels, surfacing relevant suggestions when a shopper is closest to making a purchase.

Relevant product recommendations at PDP, powered by Recommend™.

Revenue attribution from Recommend™ reached 5.76% in Sweden and 5.39% in Norway in 2025, with most European markets contributing in the 3–4% range that same year.

These numbers reflect what a well-tuned recommendation engine should deliver at this stage of the journey: incremental revenue from shoppers already in purchase mode, adding to baskets that are already forming.

Relevant product recommendations at the Add-to-cart page, powered by Recommend™.

Almost a Decade of Compounding

Ongoing optimization is built into how Miinto and Algonomy operate together.

Strategy refinements on recommendation placements, multivariate testing on discovery configurations, and continuous tuning across markets. The partnership is structured around improvement, not maintenance.

Almost a decade of that work shows in the 2025 results. And that’s what the next chapter is built on.

Content

The Results

The Client

The Challenge

The Solution

Almost a Decade of Compounding

Get in touch

“We just wanted personalization that actually worked for our customers across every market, without the complexity. Almost a decade with Algonomy has given us the infrastructure, the data, and the partnership to make it real. Today, we have a program that scales across twelve markets, delivers consistent results, and keeps getting stronger.”

Paloma Hellfeier-Truong

Director of Product & CX

Miinto

Your Retail Media program starts here.

Book a Demo

Case Study

Leading American Fast Fashion Retailer Improves Assortment

Get the PDF

Segment

Fashion and Apparel

Challenge

Reducing inventory spending and preventing loss of sales due to sub-optimal assortment mix

Product Used

The Results

0 %

range planning efforts

0 %

overstock and understock

0 %

variance between range and financial plan (from 12%)

0 %

full price sell-through

0 %

end-of-season leftovers

The Overview

The client is one of the leading fast-fashion retailers with 700+ stores across the US. The client offers apparel, accessories, and footwear targeting young men and women looking for affordable and trendy fashion. The client faced challenges with its existing assortment planning processes:

Assortment range and size planning required 840,000 man-hours every month for 300+ product groups across stores.
The large deviation between range plan and financial plan resulted in overstock of 1.6 million units.
Ad-hoc merchandise size planning led to frequent out-of-stock instances and customer dissatisfaction.

The client was looking for an easy-to-use, intelligence-driven, and proven algorithmic solution to optimize assortment and achieve merchandising financial goals.

1 Algorithmic Decisioning Platform for Assortment Planning

ADA Global Assortment Edge (AE) was the perfect fit for the client’s requirement. AE is designed to expedite and optimize the process of building an optimized assortment. With its smart 1-click automation, AE helped the client automate the time-consuming process of planning and the advanced AI algorithms recommended a demand-driven assortment breadth, depth, and size pack – minimizing markdowns, controlling inventory spendings, and delighting customers.

Our ML-based ensemble of algorithms provided highly accurate, granular, and attribute-based sales forecasts. Additionally, it performed store clustering based on key dimensions such as product class, attributes, consumer segment, etc.

2 The ROI of Algorithmic Decisioning

The client made the transition from an ad-hoc manual assortment planning to a demand-driven assortment planning, powered by intelligent features of ADA Global Assortment Edge:

With the help of smart automation, the range planning for each store was reduced to a click from the earlier 4 hours, resulting in savings of over 840,000 man-hours per month.
Algorithmic range plan recommendations helped achieve the ideal assortment breadth and depth at store level, reducing the variance between financial and range plan from 12% to 3%.
With size pack recommendations by store, cluster, and region, the client optimized the size curve and introduced multiple size packs suited to meet the demand, reduce overstock, and understock by 4%.

With AE’s smart automated workflows and ML-based algorithmic assortment optimization, the client was able to improve the productivity of its assortment planners by 700%, achieve 3% higher full-price sell-thru, and reduce end-of-the-season leftovers by 2%.