Automated Research Repository

At Hotmart — a leading creator economy platform serving 30M+ users — user satisfaction data was scattered across multiple disconnected sources: CSAT surveys, NPS scores, and support tickets. There was no consolidated view, which made evidence-based backlog prioritization nearly impossible.

Product teams were making decisions based on gut feeling or whoever spoke loudest in the room. The challenge: build a system that centralizes, categorizes, and visualizes all this feedback automatically — and actually gets used.

UX Researcher responsible for conceiving, structuring, and implementing the entire repository from scratch. I worked cross-functionally with Product Managers, data analysts, and engineers to build something that lived at the intersection of research, data, and product strategy.

I started by mapping every existing data source — surveys, NPS tools, support ticket systems — and understanding how (or if) teams were using them.

Then I defined a taxonomy for automatic categorization: tags, themes, and severity levels that could be applied without manual effort. This was the hardest part — the categories had to be specific enough to be useful but broad enough to scale.

I built interactive dashboards in Looker Studio, choosing it for accessibility and integration with the existing tech stack. The key constraint: PMs had to be able to use it without training.

I tested dashboard usability directly with PMs — observing how they navigated, what confused them, what they ignored. Several iterations came from watching people try to find answers and fail.