20% increased PVs by recommendations for the Largest Media Company in Beauty and Cosmetics in Japan.
The challenge for this media company was to improve the quality of their recommendations to its significant number of users and content.
Due to its 300 million USD annual sales were driven largely by ad revenue from its advertisers, slight improvements in session time and PVs had a significant impact on their revenue. They have tried in-house development for its recommendation engine, worked with different vendors for improvements but did not see much success.
CNN - Convolutional Neural Network
LSTM - Long Short-Term Memory
Our technology made it possible to make accurate decisions all by itself and achieve 20% increase in PVs.
To put it simply, by combining the three technologies above, we were successful in creating a constantly updating recommendation engine based on content information on each page and how, when, how much users interacted with them. The AI automatically restructures itself based on user input, so there is no need to use a team to decide what to recommend based on customer trends.