Doing data-based segmentation in the real world

There are many ways to segment users - marketers typically want to define personas based on interviews with a few potential users, while UX researchers try and segment users based on intentions. By contrasts, data analysts can create segments based on behavioral data observed on all users, without trying to impute a user’s intention or persona. Of course all three approaches are complementary and needed for different purposes. The quantitative approach can help inform opportunities and offers a framework for tracking user growth, user engagement, and funnel conversion. Using VSCO’s active audience as an example, we will take a deep dive on how to apply clustering methods to identify segments and measure their evolution over time, while avoiding idiosyncratic pitfalls.