Rambler Group begins to use neural networks for smart targeting

2021: Start using the neural network for smart targeting

Rambler Group March 12, 2021 announced the beginning of the use of the neural network for smart targeting.

The developers conducted numerous studies on the use of algorithms based on neural networks and reached a different, previously unattainable, bar of quality. The addition of neural networks increased advertising capacity by more than 200 million shows per month, available for targeting accurate socio-demographic data.

When solving any ML problem, the moment comes when quality goes to the plateau and it is more and more difficult to achieve tangible improvements. We refracted this trend by radically changing our approach. Using neural networks based on the behavior of users on the Internet, we built embedings, that is, for each visitor to our media resources, we created a vector of numbers that characterizes it. Adding this vector to existing models allowed us to significantly improve the quality of socio-demographic segments, said Artyom Vybornov, head of the data processing platform AdTech Rambler Group

Rambler Group is one of the few Russian digital companies that uses its own technologies to place and target advertising.

For the most accurate personalization, the Rambler Group uses its own DMP platform, which collects and processes a huge array of unpersonified data (more than 10 TB per day). The system analyzes and structures disparate information of more than 500 million profiles and forms it into 10 thousand audience segments, updated in real time.

Socio-demographic segments are the key and most popular category among our advertisers. Increasing the quality and capacity for them gives an increase in coverage in almost all advertising campaigns. Thanks to the work of our AdTech specialists and research carried out over the past years, the Rambler Group already accurately identifies the socio-demographic segments of users from the entire array of unpersonified data. Therefore, improving performance by even a few percent is a difficult technological task. We strive for high efficiency and twist every tool to the ideal. As a result, advertisers receive more accurate audience targeting and increasingly choose the Rambler Group as a priority partner for their campaigns, noted Yulia Konovalova, CPO AdTech Rambler Group

Technology based on machine learning and artificial intelligence helps not only in determining the socio-demographic data of users, but also in optimizing performance advertising, which is used to increase the effectiveness of the campaign. Real-time artificial intelligence estimates a large number of factors, including user behavior, platform, advertising place, date, time, weather and others, and advises the most optimal solutions for showing a particular creative.