Dow Jones Drives Business Value by Transforming Data Analytics on AWS

Dow Jones & Company Inc. (Dow Jones)—a global news provider that delivers content across digital, audio, video, and print formats—manages over a century of data. To position itself for the future and create an environment to gain greater insights, initiate innovative projects, and unlock the power of its data, Dow Jones selected Amazon Web Services (AWS).

On AWS, Dow Jones centralized its data and developed an internal analytics environment constructed on Amazon SageMaker, a machine learning (ML) service that developers can use to build, train, and deploy high-quality ML models for virtually any use case. Using this solution, Dow Jones transformed its analytics processes, making it simpler for its teams to drive business value.

Founded in 1882 and headquartered in New York City, Dow Jones oversees multiple consumer publications, including the Wall Street Journal, Barron’s, and MarketWatch as well as leading business-to-business products including Factiva, DJ Newswires, and DJ Risk and Compliance. To advance its operations, the company set out to centralize its data and modernize the analytics infrastructure. “With a company that’s nearly 140 years old, we have accumulated a treasure trove of data across several systems,” says David Chivers, senior vice president of emerging growth platforms, data, and product strategy. “Without a simplified way to reach across these silos, it was difficult to surface meaningful insights in a timely way.”

Dow Jones wanted to enhance the ability to provide its marketing, sales, product development, and finance teams with useful and actionable data. It also hoped to create an environment for various teams to create their own ML algorithms for specific use cases, fostering projects that would drive business growth. To harness the power of its data, Dow Jones replaced its on-premises data warehouse with a data lake that uses Amazon Simple Storage Service (Amazon S3), an object storage service that offers industry-leading scalability, data availability, security, and performance. By late 2018, the company had cleansed and uploaded 3 PB of data into Amazon S3.

Dow Jones wanted to activate its centralized data store by launching an internal analytics service for its data science teams. In 2020, it began to build this environment on AWS using Amazon SageMaker. “AWS really leaned into our analytics project,” says Chivers. “The AWS team detailed how to use Amazon SageMaker and hosted training days for our team. AWS won us over by fully understanding our needs and with its account management, strength of services, and speed of innovation.”

Using Amazon SageMaker and Amazon S3, Dow Jones rebuilt and modernized its ML model deployment pipeline and facilitated stronger access to data for multiple teams. The company powers its analytics service using Amazon SageMaker Studio, a web-based visual interface where users can perform all ML development steps. After uploading project datasets into Amazon S3, data scientists can clean, prepare, test, and develop their ML models using a variety of built-in open-source tools and AWS services. Once the ML models are ready, users can provision infrastructure using Amazon SageMaker for further model training. Dow Jones stores ML models on its Amazon S3 data lake using Amazon SageMaker Pipelines, a purpose-built, easy-to-use continuous integration and continuous delivery (CI/CD) service for ML. Teams can then access Amazon S3 to pull ML models for specific use cases, such as developing personalized marketing campaigns or predicting subscriber cancellation rates.

To process large amounts of data, Dow Jones employs Amazon SageMaker Data Wrangler, a service that seamlessly integrates data preparation workflows and Amazon SageMaker Pipelines to automate ML model deployment and management. The company is automatically running and monitoring over 20 ML models that were previously managed manually by data scientists. This automation decreases the overhead of training, running, testing, and monitoring ML models, reducing the workload of data scientists by 50 percent on average. As a result, employees can spend more time developing innovative, value-generating solutions.

After building its analytics solution, Dow Jones began testing various customer use cases. “Our goal is to be a truly data-driven organization in service of our customers, and using Amazon SageMaker, we are activating value from our data in service to that goal,” says Chivers. In one instance, the company built an ML model to determine customers’ preferred time of day for email communications, which increased response rates by 50–100 percent. Dow Jones also developed trust-based ML models to predict churn rates, or the rates at which readers cancel their subscriptions. These models help Dow Jones create better experiences to best serve customers and grow deeper relationships. 

Dow Jones uses its transformed ML analytics solution to continually streamline its operations. For example, the company deployed ML response models to find areas in which it can optimize its advertising costs. It also created algorithms to track webpage metrics like read times, dwell times, and completion rates, unlocking actionable insights into reader behavior. Using this data, Dow Jones can strategize ways to continually increase the quality of its products.

Dow Jones transformed its data analytics capabilities using Amazon SageMaker and Amazon S3, saving time, enhancing customer experience, and driving business value. Dow Jones will further use its analytics service to support even more sophisticated projects as it continues to its growth trajectory. For example, it plans to use natural language processing to help review and structure data from over 33,000 periodicals and articles. 

With the successful launch of its analytics solution on AWS, Dow Jones will continue to support its applications and long-term operational health by using AWS Enterprise Support, which provides companies with concierge-like service to help them achieve their desired outcomes and find success in the cloud. “On AWS, we have unlocked a rich dataset that we could have never amassed outside of the cloud,” says Chivers. “Thanks to the collaboration with AWS Enterprise Support, we are well positioned to apply increasingly cutting-edge solutions to enhance our industry-leading experiences for our customers.”