Search

Search

Consumer Discretionary

Consumer Discretionary

Databricks Enables Real-Time Personalized Hotel Recommendations at Scale for Global Travel Platform

Hotels.com, a global online travel booking platform, accelerated machine learning deployment and real-time personalized recommendations by migrating to Databricks Data Intelligence Platform.

Value Results Summary

Data processing volume increased by 20x without impacting performance

Data processing volume increased by 20x without impacting performance

Data processing volume increased by 20x without impacting performance

Highly accurate image classification and optimization for personalized property search results

Highly accurate image classification and optimization for personalized property search results

Highly accurate image classification and optimization for personalized property search results

Higher customer conversion rates through improved sales efficiency and relevant image display

Higher customer conversion rates through improved sales efficiency and relevant image display

Higher customer conversion rates through improved sales efficiency and relevant image display

Hotels.com, a leading online travel booking platform with over 325,000 hotels listed across 41 languages and 70 million app downloads, faced critical challenges managing the volume and complexity of its data operations. The company processed millions of hotel photos daily from properties and customers, requiring rapid analysis to eliminate duplicates, classify images (kitchen, pool, gym), and organize them for ranking. Additionally, delivering personalized hotel recommendations to customers in real time demanded massive compute power and advanced analytics capabilities. The existing on-premise Hadoop cluster using SQL and SAS proved insufficient, taking 2 hours to process data pipelines on just 10% of the company's dataset, creating a severe bottleneck for data science and engineering goals.

Hotels.com implemented Databricks Data Intelligence Platform to transform its data infrastructure and enable real-time machine learning operations. The solution provided advanced cluster management to scale data volume without adding infrastructure complexity, an interactive workspace to foster collaboration across data science, engineering, and business teams, and Databricks Runtime to accelerate streaming data processing at enterprise scale. By migrating to run 100% of its workflows on Databricks, the company gained the agility and flexibility needed to rapidly deploy deep learning models into production and anticipate customer behavior patterns in real time.

The platform delivered measurable business impact across data processing, customer experience, and revenue. Hotels.com increased data processing volume by 20x without performance degradation, enabling near-instant analysis of millions of images and customer interactions. The deep learning models now deliver highly accurate and contextually relevant image classification and display during property searches, optimizing how customers view hotel listings. By pairing the right hotel with the right images based on individual search behavior, Hotels.com achieved higher customer conversion rates and improved overall sales efficiency, directly strengthening its competitive position in the global travel booking market.

Similar stories

Keep exploring