Gannett, one of the largest publishers of locally based print and digital media in the United States, partnered with Adswerve to better understand audience behavior and identify opportunities to drive memberships and decrease churn. Because membership numbers remain a critical revenue component for media companies, Gannett had three primary goals:

• Increase digital memberships using propensity scoring

• Identify users at risk of unsubscribing and target with interventions to retain them

• Build a recommendation engine to determine the best content to drive action

The publisher wanted to employ BigQuery as their marketing data warehouse to power machine learning insights and interventions. They also needed to go beyond summary aggregates and use machine learning processing for more detailed analysis.

Gannett came to us with the idea of building scoring models to inform what they needed to meet their goals. Armed with this information, Gannett would be able to intercept users with relevant content or actions to increase memberships and retention.

We collaborated with the strategic development team, headed by VP of Digital Operations and Product Development Chris Biondi, to bring this project to fruition. The project team, led by Director of Corporate Consumer Data Strategy Rich Forsgren, relied on Adswerve team members to help shape and refine the strategy. We approached the initiative with five workstreams: Data Preparation and Exploratory Data Analysis, Propensity Modeling, Churn Modeling, Content Recommendations and Intervention Identification.

Now that the modeling and workflows are set up, Gannett can:

• Use propensity scoring to increase digital subscriptions

• Retain users at risk of unsubscribing

• Create compelling, targeted content that converts

The team has a solid foundation for other activities, too, including building and leveraging audiences across the Google Marketing Platform, surfacing new reporting dimensions, and building more real-time scoring to dynamically alter user experiences and develop new content personalization offerings

The project continues to evolve as the team reviews results.