Rethinking processes and solutions to grow audiences and build products at New York Times

By Scott M. Scher, Columbia University

The New York Times has a three-fold approach to growing audiences and building products: people, ideas, and things. They developed a centralized organizational structure that acts as a center for excellence. There are three main data teams at the New York Times; Data science- focuses on developing and deploying predictive algorithmic modeling for newsroom and business challenges, Data analytics- partners with finance, newsroom, and business teams to specify goals and metrics, design and run experiments, and conduct analytics, and Data engineering- intersects data migration, warehouse, and software engineering to store business data, power analytics products, and fuel data driven strategic decision making. These teams all sit together and attend the same meetings. These teams work in open communication to solve problems and share knowledge.

In the opening slide of Aram’s presentation the audience is shown the Hierarchy of Data Needs. The pyramid from bottom to top reads: stable infrastructure. Fast, easy SQL access. Data governance and common definitions. Bi tooling and dashboards. Experiments. Machine learning.

Chekijian emphasizes that without access, accuracy, and reliability, it is impossible to effectively deliver on the higher order needs and attempting to solve for all needs simultaneously may result in a state close to paralysis. When it comes to working with data it is essential to “identify the first order problems that must be solved, then move up the pyramid”.

The data and insights function of the New York Times is built by data engineering, data science, and data analytics disciplines. At the New York Times, the analytic data environment takes unified disparate data sources to enable an holistic view of content, behavior, and revenue. Using real-time event trackers of web, app, and messaging data they have been able to create advanced dashboard platforms, develop unified data environments, tools for accessing data for decision support, and delivering insights through analytics. Using dashboard reporting enables business and newsroom operators with timely and actionable information. Aram and his team are using AI to enable experimentation through their ABRA platform.

This platform creates a low latency solution in order to maintain user experience standards, front-end and back-end integrations that can be adapted to any environment. This gives the New York Times the ability to explore preferences of audience segments and provides them with behavioral and revenue metrics without any additional implementation. All of this enhanced data will allow them to launch a cross platform home screen with a new, persistent programming strategy that will maintain and increase engagement with their product. This will result in the ability to deploy product changes to a subset of customers and track the impact these changes have on behavioral and revenue metrics. As well as provide feedback for the newsroom on the impact of editorial decisions.

AI is enabling New York Times to enhance metadata for deeper clarity, efficiency in internal recommendations, and understanding retention.  Using AI to run statistical modeling they are able to predict trends in key business revenue and audience metrics, and estimate the effect marketing, product, news, and exogenous factors have on their business.