Opportunities and challenges in personalization of online hotel search
With the rise of digitalization in hotel industry the hotels became available for online booking, causing a significant shift from traditional offline booking behavior to online methods. As a hotel metasearch engine, trivago aims to aggregate hotel prices from all online booking sites for assisting visitors to find their ideal deal. This generates massive amount of information flow about the market and the user behavior. In order to improve the satisfaction of the visitors, their preference should exploited individually, which requires a real-time personalization service in our website.
In this talk, the basics of trivago business model will be introduced. I will discuss how personalization can boost the key performance indicators in short- and long term, and what are the main differences between a hotel search engine and a booking site. I will walk through the main information sources about the users that are useful to understand their preferences. Then some use cases were expounded where personalization can give additional value to the service. I will also point out some specific challenges that we can face with during the solution of these problems. Finally, the limitations of personalization will be discussed.
Data Science & Analytics Lead, trivago
David holds MSc in Computer Science and PhD in Business and Management. In the last 8 years, he worked on recommender systems and personalization techniques in several domains (traveling, IPTV, streaming media, video sharing, jobs, news and coupons) as a data scientist. He gained significant industrial experience in designing and optimizing data-driven large-scale systems to improve various business performance indicators. Currently, he is leading the data science and analytics team for user modeling in trivago, focusing on understanding user preferences and increasing user experience by personalization. Besides his product development projects, he takes care of academic relationships and research activities in data science.