Connecting data providers, brands and agencies, Audiencerate uses trustworthy targeting data to help marketers run more efficient campaigns. The company undertook the difficult task of starting a data platform from scratch and rapidly grew into a major operation.
Audiencerate CTO Enrico Vecchio explains how opting for AWS helped the company facilitate this growth: “When we started out, we sketched on the wall things we didn’t want to do, including maintaining infrastructure, machines, autoscaling, or databases. Then we found the best AWS managed services to deal with each of these.”
“So AWS CloudFormation and AWS Lambda deal with infrastructure, Lambda and Amazon Kinesis look after machines, AWS Elastic Beanstalk solves our autoscaling issues and Amazon Simple Storage Service (Amazon S3) and Amazon DynamoDB deal with the database.”
Opting for an infrastructure on AWS enables the relatively small Audiencerate technical team to quickly respond to around 250 million requests daily and analyse roughly 50 terabytes of data each month. AWS means the company doesn’t have to fret about security or scalability and can instead focus entirely on the client.
AWS services also allow the company to remain compliant with different data regulations in different territories. Data is encrypted throughout its journey, whether that’s in Amazon S3 Buckets, Amazon Kinesis or DynamoDB.
Utilising templates with AWS CodePipeline and AWS CloudFormation to build platforms also means they can quickly be reused should they need to be moved to a different region. AWS enables the company to react instantly to any changes in data regulation.
In spite of its rapid growth, AWS microservices have also allowed Audiencerate to keep its costs under control. Since it started, the company has scaled up from approximately 100 requests per second to 10,000 per second without incurring extra charges.
Each microservice that Audiencerate adds or removes from its platform has an individual billing target. This feature means the company can easily monitor the cost of each component it is using.