"By adding Amazon Athena, AWS Glue, and Amazon Aurora to our AWS services stack while also adopting the latest Amazon EC2 instances alongside serverless technologies, we have reduced our infrastructure costs by 50 percent."
 
Carl Leskinen Chief Technology Officer
  • About Burt

    Launched in 2009, Burt serves the online advertising market, including marketers, publishers, and advertisers. Since 2012, the startup’s data intelligence and analytics platform has been focused on media companies in Scandinavia and New York.

  • Benefits of AWS

    • Cuts the time to onboard customers by more than half
    • Reduces IT costs by 50 percent by leveraging new AWS services
    • Increases agility to support more complex query modeling
    • Establishes new business unit for big data services
  • AWS Services Used

    • Amazon Redshift
    • Amazon Athena
    • AWS Glue
    • AWS Enterprise Support
    • Amazon Aurora

     

Imagine you’re a media company running tens of thousands of online ad campaigns. Now imagine trying to produce reports on the performance of each one when the relevant data comes from hundreds of sources. Even the most efficient employee would struggle—which is why media owners worldwide are working with Burt. Using Burt’s data intelligence and analytics platform, large media owners can improve their advertising revenues to the order of millions of dollars by using data intelligence to streamline operations and reduce advertiser churn.

In 2017, Burt looked to increase its data intelligence and analytics capabilities to run the increasingly complex data models that customers requested. Theo Hultberg, chief architect at Burt, says, “We wanted to get ahead of the market for data intelligence and analytics. To do this, we also needed to reduce the time we spent modeling the data queries that give customers insight into the performance of their campaigns.”

Burt had been running its IT in the Amazon Web Services (AWS) Cloud since 2009. The company began by using core services such as Amazon Elastic Compute Cloud (Amazon EC2) and Amazon Simple Storage Service (Amazon S3). In 2013, it added Amazon Redshift to its AWS services stack. Hultberg explains, “Amazon Redshift helped us dig deeply into complex data sets to answer specific queries from customers. Amazon Redshift always delivers good performance.”

Now, to reduce time spent modeling queries, Burt looked at Amazon Athena, a serverless interactive query service. Hultberg comments, “Amazon Athena delivered performance comparable with Amazon Redshift, but with slightly more flexibility around query modeling. The inclusion of Amazon Athena allowed us to change and iterate on complex data models quickly, reducing the onboarding time of new customers by more than a half.”

At the same time as Burt assessed Amazon Athena, it also looked at AWS Glue, a fully managed extract, transform, and load (ETL) that can be used for both Amazon Athena and Amazon Redshift. Indeed, the AWS Glue Data Catalog can serve as a single data catalog across both Amazon Athena and Amazon Redshift. Because it can take time to build internal expertise around new AWS services, Burt turned to AWS Enterprise Support to accelerate its ability to leverage Amazon Athena. Carl Leskinen, chief technology officer of Burt, says, “Having AWS Enterprise Support and a technical account manager helped us seamlessly adopt Amazon Athena and AWS Glue.”

AWS Enterprise Support ensured that Burt gained access to the product developers behind the services. “Things are so much easier when you can talk directly to the creators of these services,” says Hultberg. “Our technical account manager set up biweekly calls with the AWS product teams, so we could talk through any challenges. It was also clear that AWS valued our feedback on the services because they were so new in the market.”

Hultberg continues, “Working with AWS Enterprise Support encourages innovation and increases agility. Having such a high level of AWS expertise ready and waiting means you can pursue development with greater confidence. And it doesn’t matter if you’re not an enterprise-sized AWS customer—the service is within reach even if you don’t have a huge budget.”

Having tested AWS Glue and Amazon Athena successfully, Burt transformed its data analytics and intelligence platform. The company built a data lake where Amazon Athena and AWS Glue sit alongside Amazon Redshift and Amazon Aurora for PostgreSQL. The ability in Amazon Redshift to query the same tables as Amazon Athena through Amazon Redshift Spectrum helped speed up critical reports. Plus, Amazon Aurora for PostgreSQL gave Burt great performance when reporting on smaller data sets, where very low rates of latency were critical.

Depending on the customer use case, Burt engages Amazon Athena, Amazon Redshift, or Amazon Aurora to run queries against data in the data lake. “Our customers seamlessly get served by Amazon Athena, Amazon Redshift, or Amazon Aurora when asking a query without ever being aware of the underlying technology,” says Leskinen. Regardless of the complexity of the data modeling, we can choose an AWS technology to meet the needs of our customers”. He continues, “By adding Amazon Athena, AWS Glue, and Amazon Aurora to our AWS services stack while also adopting the latest Amazon EC2 instances alongside serverless technologies, we have reduced our infrastructure costs by 50 percent.

Boosted by the agility it has gained using the latest AWS technologies, Burt launched a new business unit focusing on big data services. The team has developed a platform that consolidates customer and advertiser data into a single view for in- depth analysis. It handles billions of messages a day and is based on extremely large data volumes from one data source that Burt has created.

Underpinning the platform are the same services—Amazon Redshift, Amazon Athena, and AWS Glue—on which the campaign-focused platform runs. Leskinen says, “We formalized the new business unit six months ago, based on a demand for big data services across the advertising industry and the flexibility of our AWS services. So far, the results have been excellent.”

Learn more about Enterprise Support