DESIGNER & RESEARCHER, JAN 2021 - DEC 2024
Over the course of three years, NPR migrated 200+ radio station websites to a modernized third-party Content Management System (CMS) as part of a project named Grove. The vendor provided the basic design template for home and story pages, but we added our own enhancements post-migration.
NPR radio stations are unique. Some play music almost exclusively while some serve as the only local news in their area. Each station had its own high priority feature requests. My product manager asked for help prioritizing work that would benefit the most stations first. I approached the research for this three ways.
I first conducted a review of secondary data. It was the easiest to access for immediate input while I sought permission for direct user outreach. Every two weeks, station staff attended an open office hours session with member reps to answer questions and troubleshoot issues. I audited the questions raised in those meetings, as well as four months of ticket requests.
Customization was a big theme, so I explored visual improvements such as color schemes, font alternatives, and module enhancements (above.)
Pros: Readily accessible data, quick turnaround
Cons: Self-selected user base, environment primed for obvious bugs
Continuous discovery
I then pivoted to continuous discovery and invited my team observe weekly interviews of station staff. After 10+ interviews, it was clear what users really loved about Grove was how it made their work faster. This, of course, was not something that was evident from the secondary data of bug reports. Here was an opportunity to lean into efficiency for product happiness.
I plotted the steps of a common workflow in Grove and identified 13 high-impact wins to save time. One change alone saved 6.5 hours of pointless clicking annually per person.
Pros: Direct and recurring access to real humans, encourages team investment
Cons: Requires lead time to operationalize, can lead to slow-forming insights
A learning framework
Our ticketing system was filled with people wanting custom features, but our interviews included people that were quite happy with Grove. I issued a survey to better understand user groups. 55% of stations (more than 100) participated.
The results helped us craft trainings and support documentation in the short term. Long term, it gave us a starting point for curating permissions settings based on realistic personas.
Pros: Data at scale, fast reach
Cons: Numbers can lie, tells you what but not why
Ultimately, mixed methods give the clearest picture, so all of the above combined formed a powerhouse of knowledge about our users.
