For the Love of Data, What is Predictive Modeling?

The ArtsBoston Audience Lab has taken on a vital challenge: how can the greater Boston arts and culture sector be more inclusive and engage audiences we don’t currently see among the mid to large arts organizations in the region? The City of Boston alone is majority people of color, so why is it so difficult for organizations to reach more audiences of color? To help us figure out this challenge, we looked to other sectors which have a strong track record of reaching audiences of all kinds, and mobilizing these audiences to take action. One of the best examples we found came from the civic engagement world, and particularly,their use of predictive modeling for political campaigns and policy change.

In the Pilot phase of the Audience Lab (Audience Diversification Pilot Project), we worked with Clarity Campaign Labs, a DC-based firm which helps organizations use their data for effective analytics, research, and strategy. Using the current data from existing audiences of color attending performances and events of the participating organizations, Clarity Campaign Labs created a predictive model that would allow us to find prospective audiences of color with a high likelihood of arts attendance.

WHAT IS A MODEL?

A model is a tool used to assign each individual a probability that they will either

• take an action (sign a petition, vote, attend a show, etc.)
• be receptive to a campaign or marketing program
• support an issue or candidate

In order to find these audiences in greater Boston, we partnered with the Mass Voter Table, which uses the Voter Activation Network (VAN) to help organizations find Massachusetts residents who may be interested in their particular cause. We were able to use the predictive model created by Clarity Campaign Labs to score all Massachusetts voter records in VAN, in order to pull contact information of our target audience. From there, participating organizations will be able to run their marketing experiments to their respective new audience. Prospecting often yields a low return on investment (ROI) which means arts organizations are mostly focused on their existing audiences which are not particularly diverse. By building a predictive model and finding patrons who have a high interest in the arts, the Audience Lab is able to overcome this challenge with these new tools, strategies, and partners.