Six months of data
Established recurring patterns and the six clusters to investigate.
Globo · Behavioral segmentation · Mixed methods
From clusters to product questions.
How we connected quantitative navigation patterns with the routines, motivations and contexts behind them without mistaking correlation for identity.
6 monthsof product data
28interviews
6 × 7 daysdiary study
The decision
The clusters described what happened. Product needed to understand why.
Six navigation patterns had emerged from homepage behavior. The opportunity was to make them useful for personalization, navigation and content discovery without turning them into simplistic personas.
Research design
28 interviews
Explored context, motivation and the meaning participants gave to their behavior.
Seven-day diaries
Revealed how routines and content needs changed over time.
Analytical stance
A cluster is a pattern in data. It is not a person.
I worked with Product, Design and Data Science to connect pattern, context, motivation and recurrence while resisting the temptation to treat behavioral correlation as fixed identity.
What changed
The output was a set of testable product questions.
Which patterns are stable enough to support personalization?
When does context matter more than historical behavior?
How should navigation adapt without narrowing discovery?
What evidence would distinguish a useful signal from noise?
The documented impact is directional: the research translated statistical clusters into hypotheses and decision criteria. No downstream product metric is available, so none is claimed.