Cadreon is the advertising tech unit of IPG Mediabrands, the media innovation division of Interpublic Group. Cadreon offers a specialized digital performance platform for media brands that integrates technology, data, and inventory to target audiences in real time. The organization’s solutions are used by many of the world’s top brands to manage and increase the effectiveness of advertising campaigns.
In early 2016, Cadreon sought to develop a new audience-targeting platform that would integrate data from multiple data sources, with the goal of helping media companies maximize the effectiveness of digital advertising campaigns. “We wanted to build a solution that could manage very large datasets—up to 20 billion rows and containing close to a petabyte of data—while also supporting an application to enable real-time data analysis,” says Tushar Patel, senior vice president of engineering for Cadreon.
Cadreon, however, was concerned that using its existing on-premises environment to support the platform would be too challenging. “We only have a few people here to manage an IT environment, and it was already taking up too much of our time,” Patel says. “We want to focus on creating and running new solutions, not managing the backend of it all.”
The company also needed to ensure its new platform could scale to support a potentially fast-growing volume of data, such as demographic data and television-viewing trend data. “We projected a lot of growth for the solution, and we had to have the ability to quickly scale up to support that without investing a lot in hardware and personnel to manage it,” says Patel.
Why Amazon Web Services
As Cadreon considered its technology options, it focused its attention on the cloud. “We saw that the cloud would give us the flexibility to procure new compute resources very quickly and bring them down as needed,” says Bhupen Patel, vice president of data engineering at Cadreon. The company’s search led it to Amazon Web Services (AWS). “AWS is the leader in the cloud space, and we also felt the services from AWS were much more comprehensive than what other providers were offering,” Bhupen Patel says.
Cadreon created its new audience-insights platform on AWS, using the Amazon EMR managed Hadoop framework to process large amounts of data across scalable Amazon Elastic Compute Cloud (Amazon EC2) instances. The company also runs the Apache Spark framework on Amazon EMR and the Druid open-source analytics data store on Amazon EC2 for indexing data. The platform contains a total of 240 TB of data from 11 different advertising sources.
Cadreon runs its audience-insights database on Amazon Relational Database Service (Amazon RDS), and collects television-viewing trends, demographics, and a range of other data in an Amazon Redshift data warehouse. IPG Mediabrands analysts use the solution to conduct advertising-campaign analysis for media customers throughout the world.
Cadreon was able to easily get its audience-insights platform up and running on the AWS Cloud. “We built and launched this solution very quickly on AWS, and we were able to immediately spin up several hundred instances,” says Tushar Patel. “It would have taken us three months to order the hardware and get it all set up in a data center. Using AWS, we built a sophisticated big-data platform in the cloud that frees us from having to manage the backend environment. We can support 400 instances with only a few people, which we could never do in a traditional environment.”
Leveraging the audience-insights solution, IPG Mediabrands analysts are able to quickly process data for their customers. “Our solution can process big-data queries from thousands of sources in a few seconds, so media companies can better define ad campaigns and audiences. With that data, these companies can target their audience to best achieve the goals of an ad campaign,” says Tushar Patel.
Additionally, Cadreon has the elasticity to scale the platform up or down when necessary. “We get our third-party data on a monthly basis, so we really only need to access and process that a few days each month,” Tushar Patel says. “Using AWS, we can scale the platform up quickly and bring it back down when that time period is over. That saves us both time and money.”
As the platform grows, Cadreon will be able to easily meet demand. “We want to make sure that if this solution takes off in a big way, we can support that growth without compromising the user experience,” says Tushar Patel. “We can definitely do that using AWS. And the fact that we can do it without making a large upfront investment in hardware is a huge advantage for us as a company.”
Cadreon is planning to explore additional AWS services—including Amazon Kinesis—as it expands the platform. “We are always interested in seeing how we can integrate additional AWS services into our platform,” says Tushar Patel. “We want to always keep our focus on building and enhancing our products, and AWS helps us do that more efficiently and cost effectively.”
“Using AWS, we built a sophisticated big-data platform in the cloud that frees us from having to manage the backend environment. We can support 400 instances with only a few people, which we could never do in a traditional environment.”
Tushar Patel, Senior Vice President of Engineering