Case Study: a leading vacation rental company enriches millions of listings with Labelbox

a leading vacation rental company automates 9 million listing tasks with Labelbox

A leading vacation rental company faced the challenge of classifying and enriching millions of its property listings using visual attributes from images. The sheer volume of unstructured data made it difficult to create signal that would improve discovery for users seeking flexible travel options. The company turned to Labelbox to standardize its machine learning infrastructure and automate this process.

Using Labelbox's managed services, which featured active learning and a shared ontology, the company automated its data pipeline. This solution allowed models to generate most of the training data, with subject matter experts providing feedback. The result was a drastic reduction in expert feedback costs and the completion of over nine million tasks. The enriched listings, particularly those outside major tourist destinations, have collectively earned more than $300 million since the start of the pandemic.