We are a team of tech and innovation enthusiasts, striving for greatness, change and transformation.
Every day, we take part in hundreds of millions of auctions. Our A.I. receives billions of impressions each day and takes decisions in less than 60 milliseconds.
We make Real-Time Bidding work with a powerful proprietary technology that has been delivering global campaigns for the past ten years.
Deep Learning is a subfield of Machine Learning concerned with algorithms and inspired by the structure and function of the brain called Artificial Neural Networks. Machine Learning algorithms are normally flatter and more simple, while Deep Learning algorithms are stacked in hierarchy. They are also far more complex, abstract and efficient.
Logistic Regression model is a complex Machine Learning model that predicts the output of a binary response (e.g., click, don't click; buy, don't buy) given the input.
Our A.I. does exactly that: it predicts future clicks and/or conversions based on historical data, and obtains results.
Our auto-optimization system is powered by a Reinforcement Learning algorithm, a ML type belonging to the Deep Learning family algorithms. With rewards received in real-time into the system, our tech platform, Jenga, is able to optimize campaigns starting with the first conversion received.
With the help of Logistic Regression we can evaluate the probability of a specific event (install or post-install) will happening. With a trained set of data our system is able to filter the enormous volume of incoming requests and place a bid on the best promising ones only.
We are living the post-IDFA era so we can base our decisions on less user informations. We can of course execute a context-based targeting but we can also enrich the incoming requests with additional informations gathered from first and third party database before the bidding stage.
Our OpenRTB bidder is able to create dynamic creatives with different sizes and text variations. All creatives are rotated during the exploration stage and it keeps the best performing ones automatically in the next stages (optimization and scaling).
We can set several kinds of targeting. Our campaign specialists can fine tune your needs into our campaign settings in order to achieve your goals. Available targeting: User level, Device level, Geolocation level, Traffic source level and others.
Our anti-fraud filter protection is always enabled on both impressions and clicks layers and it works in real time. It's also updated regularly to assess the new threats that usually comes out.