BidMotion's ad-tech platform provides ad tracking and audience analysis to help companies make their mobile applications more profitable. Using machine learning capacities, BidMotion analyzes its customers’ audience data to display advertising in the right format and at the right time to mobile users. Optimized targeted advertising helps BidMotion customers improve their mobile applications profitability. BidMotion’s customers include service providers like Uber and game publishers such as King (Candy Crush). Based in Paris, BidMotion operates in Europe, Asia, and the United States, and continues its international development.
BidMotion stands out from the ad-tech market as a platform optimizing mobile advertising by using machine learning. Yet, machine learning, which enriches every new data analysis by benefiting from analysis previously performed, involves the storage of huge data volumes. BidMotion was looking for a highly scalable infrastructure solution with storage capabilities adapted to these big data needs. “We wanted to avoid any technical limit, while controlling storage, administration, and resource ramp-up costs,” says Juan J. Mostazo, BidMotion cofounder and CTO.
BidMotion also needed an infrastructure ensuring low latency times for mobile users in order to not hinder downloading of mobile applications. It also required a flexible infrastructure able to adapt to traffic fluctuations without increasing latency.
Why Amazon Web Services
Only a cloud infrastructure could meet BidMotion’s needs. Mostazo had used AWS before and knew the quality of its managed services. The AWS reputation was also attractive to BidMotion. According to CEO Daniel Nathan, “Information security is thus a priority for our customers, and when they find out that we run on AWS, it is reassuring for them. AWS has an uncontested reputation for data safety.”
Today, BidMotion uses about 30 Amazon Elastic Compute Cloud (Amazon EC2) instances and can have additional on-demand instances available. Amazon Elastic Load Balancing (Amazon ELB) ensures load balance to distribute application traffic while Amazon DynamoDB enables automatic and fast data and traffic distribution on servers. BidMotion uses the Amazon Redshift data warehouse for analytics and Amazon Simple Storage Service (Amazon S3) to store several tens of terabytes of raw data. Using Amazon Elastic Map Reduce (Amazon EMR) combined with a Spark framework, BidMotion handles daily processing of big data volumes necessary for optimization of its customers’ advertising. The startup also uses AWS Lambda to receive customers’ information as well as Amazon Route 53 for domain name management.
Thanks to AWS’ scalable storage and big data services, the BidMotion solution can boost its advertising audience analyses with machine learning. “So far, our machine learning solution has analyzed billions of data and analyzes more than one additional billion per month. With AWS, we are capable of storing and accessing it all,” explains Mostazo. Nathan continues, “By relying on AWS storage, our analyses by machine learning identify the best audience targeting criteria in real-time. This way, we can give our customers an advertising broadcast generating high download rates for their apps, to the point that several of them decided to grant a bigger budget to BidMotion than to Facebook.”
Additionally, BidMotion provides audience analysis results to its users. This way, they can visualize the variables and learn how to boost the download rates of their applications.
“Our AWS infrastructure provides low latency and a good mobile experience for our customers’ customers. Because we use AWS, when a mobile user clicks on one of our customers’ applications, he or she is directed to the download page in just 20 to 80 milliseconds, compared to 300 milliseconds with our competitors,” Mostazo says.
Fast service is important for BidMotion customers. “This quality mobile experience contributes directly to an increase in download rates for our customers, and thus their profitability,” Nathan says. Low latency time is maintained even during peak traffic thanks to the use of on-demand Amazon EC2 instances. “We trust the flexibility of AWS to adapt to our one-off needs, for example processing peaks of 1,000 requests per second,” Mostazo says.
Already using three AWS regions in Europe, in the United States, and in Singapore, BidMotion continues its international development. “Thanks to AWS’ global coverage, we continue to develop into new markets without heavy investment in infrastructure. Also, it is possible to launch instances in a new AWS region anywhere in the world in less than five minutes,” Nathan says.
“Our AWS infrastructure provides low latency and a good mobile experience for our customers’ customers.”
Juan J. Mostazo, Cofounder and CTO