A cutting-edge model of user behaviour that predicts the likelihood of ad clicks and conversions
Build a behaviour model that for a given user and product predicts the likelihood that the user will click an ad of this product, and/or will purchase it.
RTB House is one of the first retargeting companies in the world to have developed and implemented a proprietary technology based entirely on DL solutions for the purchase of RTB ads. In 2018, the company was ranked eighth among technology companies in the Financial Times 1000:
Europe’s Fastest-Growing Companies 2018, won the Silver Stevie® Award as the most innovative company of the year, received the AIconics Award in the Best Application of AI for Sales & Marketing category and won the Big Innovation Award granted by the Business Intelligence Group.
Our algorithm made it possible to accurately predict user actions, which is a key component of RTB House technology.
In other words, for a given user and product, we need to predict the likelihood that the user will click on an ad of this product. Our job was to test a variety of models including neural networks, gradient boosting, random forests, factorisation machines and various types of regression. We used many techniques, such as k-NN-related functions obtained with LSH (locality-sensitive hashing).
Our LSH regression is available open source in the Practical Approximation Algorithms Library (PAAL). The solution was created by verifying various hypotheses and testing many different ideas. The joint project contributed to the rapid development of RTB House and its global success in the online advertising market!
Revenue increase of 14.2 percent with the same percentage of gross margin and reduction of required resources.