Data-Driven Audience Creation

Overview

I led a team of designers in collaboration with the audience strategy team at a large media enterprise. Our goal was to craft a prototype for a tool that utilized the company's extensive data resources to expedite audience creation, facilitate ad syndication, and offer accessible audience insights and scale these practices to the rest of the company. By working closely with the client and incorporating as much user research as possible on a tight timeline, we were able to develop a prototype for a tool that allowed the client team to do something new in a way that was familiar.

Process

• Through discussions with the client we identified the goal of the project as enabling a wider array of roles to use advanced data methods when creating audiences.

• We determined that we would accomplish this goal by developing a prototype of a no-code tool that would demonstrate how data-driven practices could be adopted by other teams.

• The tool would allow the team to use machine learning to identify audiences for ad campaigns and explain the character of these audiences to their clients without needing to code.

• Through user interviews and contextual inquiries we gained a baseline understanding of the team’s workflows and tools before developing our initial prototype.

• Starting with a user journey sketch, we iteratively developed a fully interactive prototype that aligned with the vision of the a client through biweekly design sprints and reviews.

• Meanwhile, we maintained a continuous dialogue with the data team to ensure the feasibility of our prototype and a smooth transition into development

• Additionally, we also worked with a team in IBM Research to incorporate explainability tools into the the prototype.