Transforming search with Machine Learning

Podimo, a Danish leader in the podcast and audiobook industry, sought to revolutionize their search functionality to better serve their diverse user base. With a focus on rewarding creators and providing a seamless listening experience, Podimo aimed to enhance the accuracy and relevance of the matches according to user preferences.

Addressing the challenges of search constraints

Podimo's existing search system faced challenges in meeting user expectations, often delivering irrelevant results due to a limited understanding of queries and content semantics.

To address this, STX Next proposed the incorporation of advanced methodologies such as Learning to Rank (LTR), semantic search, and conversational search to elevate the experience. Additionally, we decided to implement Machine Learning models that would allow analysis of user interactions, understanding their intent, and facilitating a conversational search interface.

Combining Machine Learning with semantic search models

Taking the customer's needs into consideration, our dedicated team proposed the implementation of new features, including:

A Machine Learning model:

A dynamic ML model trained on user interactions to adjust podcast and audiobook rankings, aligning search results with user preferences.

A semantic search model:

A sophisticated language model designed to understand user queries' context and semantics, enabling content discovery beyond lexical matching.

Conversation search:

Introduction of a conversational search assistant to bridge user interactions with the search engine, gathering feedback for further enhancements.

Search results combination:

An integration of lexical and semantic search outcomes through an ML model to re-rank results for accuracy and relevance.

podimo mobile graphics case study stxnext

Our collaboration with STX Next has been exceptionally professional and fruitful. They have consistently delivered high-quality ML models, improved our search engine, and successfully prototyped applications with LLMs, demonstrating their reliability and expertise.

Benjamin Biering

Global Director of Artificial Intelligence

Podimo

Curious about the impact of Machine Learning on your business?

Chat with our experts and discover what STX Next can do for you.

Get started

Better search results for improved user experience

The incorporation of these features led to significant benefits for Podimo and its users:

Increased relevance:

Average position of clicked shows improved by 27%. Users now receive highly relevant search results tailored to their explicit queries and content preferences.

Enhanced discovery:

Conversion rate (searches resulting in a selected show) increased by 3.5%. Semantic search capabilities broaden content discovery, enabling users to find relevant content matching their interests and intent.

Innovative interactions:

Meaningful listens (plays exceeding 5 minutes) rose by 1.3%. The introduction of a conversation search assistant showcases Podimo's forward-thinking approach to user interaction, setting them apart in a competitive market and paving the way for future enhancements.