Case study
The engine that drives 23% of revenue
About 23% of revenue from suggestions, in five weeks.
- Client
- Fashion e-commerce
- Sector
- Fashion & e-commerce
- Size
- SME
- Area
- Italy
- Solution
- Advanced AI & e-commerce
- Period
- first 5 weeks after launch, 2025
~23%
of revenue from recommendations, within 5 weeks
01 · Problem
Where we started
Static suggestions, the same for everyone.
02 · What we built
The solution, piece by piece
Base level: statistical recommendations
Order statistics, every hour.
Advanced level: AI that reads the page
Reads the page the customer is on.
Full level: a model trained on the store
A model trained on the store.
Widget and dashboard
Widget on the site, daily dashboard.
How the engine grew, week by week
| Period | Phase | Share of revenue | Orders with a suggestion |
|---|---|---|---|
| Weeks 1–2 | Launch | 12% | 42% |
| Weeks 3–4 | Growth | 29% | 62% |
| Weeks 4–5 | Consolidation | 30% | 75% |
Real result. Fashion e-commerce, first 5 weeks, 2025. Base level only.
How it thinks
From an order to a suggestion
Every hour we read the orders
It starts from real baskets, not hand-written rules.
03 · Where the person steps in
The store team owns the rules
The team decides what must never appear.
Human-in-the-Loop
- Exclusions decided by the team
- Real-time stock filters
- Results reviewed together every week
04 · Results
The project's numbers
~23%
of revenue generated by the engine
base level only
Fashion e-commerce, first 5 weeks, 2025
~60%
of orders with at least one suggested product
Fashion e-commerce, first 5 weeks, 2025
48%
of revenue from the engine on the best day
Fashion e-commerce, first 5 weeks, 2025
~2.5
products suggested per order
Fashion e-commerce, first 5 weeks, 2025
Related product
Sartia · Product photos for the same catalogue, generated with AI.