Case Study
A leading subscription streaming platform engaged our team to perform a sophisticated segmentation analysis of its subscriber base. The objective was to categorize subscribers based on their consumption behaviors to identify key engagement cohorts. Through the application of advanced techniques like Principal Component Analysis (PCA) and K-Means clustering, we uncovered critical insights that enabled our client to tailor its marketing strategies more effectively.
A leading subscription streaming platform engaged our team to perform a sophisticated segmentation analysis of its subscriber base. The objective was to categorize subscribers based on their consumption behaviors to identify key engagement cohorts. Through the application of advanced techniques like Principal Component Analysis (PCA) and K-Means clustering, we uncovered critical insights that enabled our client to tailor its marketing strategies more effectively.
Our client offers a subscription-based streaming service featuring live sports broadcasts, game replays, and a variety of region-specific content. With a diverse range of offerings, the platform needed a more sophisticated method to understand and engage its subscribers. They wanted to move beyond traditional segmentation methods and focus more on the nuances of user behavior, enabling them to identify their most engaged users and target them with tailored marketing initiatives.
The primary challenge they faced was the need to segment its subscribers based on complex consumption behaviors rather than relying on a single metric. A simplistic approach could lead to inaccurate segmentation, missing out on valuable insights about highly active users who may exhibit different types of engagement. For example, distinguishing between a user who streams for long periods in a single session versus one who streams in shorter, frequent sessions was essential. This required a more nuanced segmentation strategy that could capture these variations and provide actionable insights for targeted marketing efforts.
To address this challenge, our team implemented a data-driven, multi-faceted approach to segmentation, leveraging advanced analytics techniques:
We began by analyzing 19 distinct engagement metrics, including average monthly minutes streamed, average session length, and total programs viewed. These metrics provided a detailed picture of user behavior but were highly correlated, making direct analysis complex.
To reduce complexity while retaining the richness of the data, we applied PCA. This technique allowed us to condense the original 19 metrics into four principal components, each representing a unique dimension of user engagement. These components, derived as linear combinations of the original metrics, varied in influence, with some metrics playing a more significant role on certain components. These components included “Primary Engagement,” “Secondary Engagement,” “Tenure and Frequency,” and “Recency,” named according to the metrics that most strongly influenced each one, effectively capturing the different aspects of user interaction with the platform.
Using the derived components from PCA, we performed K-Means clustering to segment the subscribers into distinct groups based on their engagement patterns. This method enabled us to classify users accurately, identifying high-value “super subscribers” and other key segments with great precision.
The final step involved visualizing the segmentation results, allowing Bally Sports+ to see the distribution of subscribers across different engagement levels. These insights were crucial for the marketing and UX teams to develop more personalized and targeted strategies.
To ensure the segmentation model remained current and actionable, we integrated the entire process into Google Cloud. This setup facilitates continuous updates and refinements as new data is collected, ensuring that the insights remain aligned with ongoing marketing and UX efforts by providing the most current information.
The advanced segmentation strategy we implemented yielded significant and actionable results. By leveraging Principal Component Analysis (PCA) and K-Means clustering, Bally Sports+ gained a comprehensive understanding of their subscribers’ diverse engagement behaviors. This newfound insight allowed them to craft highly targeted marketing campaigns, resulting in improved user engagement and retention. Furthermore, the ability to identify and focus on “super subscribers” enabled them to optimize the user experience for their most valuable customers, driving satisfaction and loyalty. These strategic enhancements contributed to increased subscription growth and a more efficient allocation of marketing resources, ultimately boosting the platform’s profitability and positioning the global leader for sustained success in the competitive streaming industry.
