Deepening Audience Understanding for Nationwide Public Broadcasting
PBS is the most prominent provider of public educational television programming in the U.S., servicing over 350 member stations. With over 800 million annual pageviews to PBS.org, PBS had working assumptions about visitor behavior — but needed more to help them take action. They looked to us to help deepen their understanding of their online audience segments.
We leveraged big data in the Google Cloud Platform to mine insights across the PBS ecosystem, using these insights to to identify new segments, develop in-depth personas per segment, and inform content development. We empowered the network with the data they needed to fine-tune their site so that they could better meet their audiences’ needs.
PBS.org serves as an online content hub, supporting the television experience with local schedules, new program information, clips and full episodes, and more. To inform site design and content structure, PBS needed to uncover specific visitor behaviors beyond metrics available through Google Analytics 360.
But the sheer size of the PBS dataset was overwhelming, with over 330 million sessions, 800 million pageviews, and 17.5 million episode plays per year. In addition, average page views and other descriptive statistics only summarized the big picture; PBS needed a solution that tested their hypotheses and provided a deeper knowledge of key segments and behavior.
Using Google Cloud Platform to draw on data science techniques, we helped PBS identify user patterns that would have been impossible to detect from averages alone. This detailed process involved exporting Google Analytics 360 data into BigQuery to streamline the dataset, making it manageable for analysis. Then, Google Cloud Platform was used to connect user trends and ultimately yield seven distinct user groupings based on similarities in digital behavior. Machine learning validated and expanded upon the contextual-based theories PBS had formed about its audiences.