X3M Drives Average Revenue per User with Real-Time Waterfall Optimization

X3M emerges as a fresh and dynamic AdTech enterprise with strong industry support. Their primary goal is to establish a publisher-centric approach in an industry predominantly focused on advertisers. Originally receiving support from Etermax, X3M has since evolved into an autonomous entity committed to transparency and impartiality. Their comprehensive range of services encompasses a fully managed AdOps service, a mediator, an exchange, and various other offerings.

Challenge

X3M’s mediator encompasses a key aspect of their services: optimizing waterfalls. This presents a unique challenge as publishers possess less data compared to advertisers and must navigate multiple advertising platforms with varying behaviors, all while contending with the ever-shifting landscape of buying mechanisms.

Essentially, it becomes a strategic game where one party aims for higher CPMs while the other strives for impressions at lower CPMs . Consequently, X3M must continuously fine-tune their logic to gain a competitive advantage, considering the intricate understanding of each network’s behavior and developing a scalable solution that accommodates diverse demographics and game scenarios.

While X3M had been refining this process for a considerable time, they wanted to consider a fresh approach leveraging AI and Machine Learning to achieve the next significant breakthrough. With this in mind, X3M determined that Mutt Data was the ideal partner for tackling this ambitious endeavor.

Solution

Our team of expert #DataNerds worked with the X3M team to successfully develop an innovative real-time waterfall optimization algorithm driven by state-of-the-art AI systems. By utilizing predictive models for fill rate and latency, the task force created real-time waterfalls that optimize revenue potential for each advertising opportunity, considering the available placements.

To validate our logic, an offline system simulator for any waterfall policy was developed, enabling the team to evaluate the behavior of the new implementation and further iterations. This tool proved instrumental in assessing the impact of upcoming changes and became an integral part of the daily toolkit for data scientists, AI specialists, and machine learning engineers working on waterfalls. Through these simulations, X3M efficiently iterated on new machine learning models, business rules, heuristic adjustments, and other design decisions.

Throughout our AI journey, we elevated the AB testing process by automating analyses, conducting significance testing, and establishing key performance indicators (KPIs). To achieve this, we harnessed the power of Vowpal Wabbit, Airflow, MLFlow, Databricks, and good old-fashioned maths.

25% Decrease in CPCs

“Mutt Data has been a trusted strategic consultant from the very beginning of this project, working with us to define a clear roadmap and providing end-to-end capabilities. Their expertise in AdTech and AI is unparalleled, and their flexible engagement model has made them a perfect ally for our success.”

Ignacio Ortiz Freuler
CEO at X3M

Impact

Online tests showed an increase of more than 25% in the amount of clicks, without affecting the platform’s overall organic GMV in a significant manner.