Suggest, Complement, Inspire: Story of Two Tower Recommendations at Allegro.com

TL;DR

This paper presents a flexible, scalable content-based recommendation system at Allegro.com that efficiently handles multiple recommendation tasks and reduces maintenance costs, validated by extensive A/B testing.

Contribution

The paper introduces a unified Two Tower recommendation architecture adaptable to various tasks with minimal modifications, improving scalability and maintenance.

Findings

Significant engagement and profit improvements confirmed by two years of A/B testing.

The same model architecture effectively serves similarity, complementary, and inspirational recommendations.

Reduced maintenance overhead through a unified, adaptable system.

Abstract

Building large-scale e-commerce recommendation systems requires addressing three key technical challenges: (1) designing a universal recommendation architecture across dozens of placements, (2) decreasing excessive maintenance costs, and (3) managing a highly dynamic product catalogue. This paper presents a unified content-based recommendation system deployed at Allegro.com, the largest e-commerce platform of European origin. The system is built on a prevalent Two Tower retrieval framework, representing products using textual and structured attributes, which enables efficient retrieval via Approximate Nearest Neighbour search. We demonstrate how the same model architecture can be adapted to serve three distinct recommendation tasks: similarity search, complementary product suggestions, and inspirational content discovery, by modifying only a handful of components in either the model or the serving logic. Extensive A/B testing over two years confirms significant gains in engagement and profit-based metrics across desktop and mobile app channels. Our results show that a flexible, scalable architecture can serve diverse user intents with minimal maintenance overhead.