I partnered with the Product Manager to lead problem definition research with clients and internal stakeholders to understand where the product was falling short. Through interviews and affinity mapping, I identified patterns around usability gaps and missing data that were driving CSM dependency. Research revealed users needed hundreds of additional data points to understand the audiences they were targeting. The challenge was making that volume of data easy to navigate and discover. Our backend didn't support fuzzy search, and building it would have significantly delayed launch, so users would rely entirely on a newly designed taxonomy. I worked closely with CSMs who understood mental models across different industries to develop a structure that matched how users actually thought about audience data. After validating through usability testing and refining labels based on feedback, I designed several UX improvements. Audience selection was separated from analysis to keep the critical decision at the top while maximizing screen space for data. Navigation became sticky so users always knew where they were. Users could now sort by the biggest differences between audiences to surface key insights rather than interpreting raw percentages. And bulk data export let users pull everything they needed in one action instead of chart by chart.