Communications
Quality-Assurance Application Redesign
Role: UX Lead
Artifacts: Wireframes, Data Viz Style Guide, Agile Stories, Reconfiguration of Product Roadmap.
The Challenge
Synopsis: Our oldest customer-care Quality-Assurance (QA) application was built as a bundled add-on, but never thought of as a primary product. As customer demand for a robust QA application drove up the application's profile, the deficits were more apparent, and it became clear that we needed to rip up the old system and start again. UX and Product designed a first-in-the-industry quality-assurance application.
However, many clients had been using the client as the primary workflow application for customer service quality assurance. Consequently, as usage of the client became more widespread, so too did the complaints about its shortcomings, including one large client threatening to cancel their entire network phone contract the next time it was up for renewal.
With a lot riding on making an application that enterprise clients wanted to use, we were tasked with redesigning and building a completely new experience for quality-assurance monitoring.
The Work
Data viz card concepts
(Click to enlarge)
Research Methods
Compiled and taxonimized data from customer feedback logs.
Subject matter expert interviews, stakeholder interviews, contextual interviews with QA associates.
Weekly user testing rounds.
Design Work
Work with PO and enterprise clients to identify requirements and opportunities to increase user satisfaction and customer retention.
Co-design sessions with enterprise stakeholders.
Feature prioritization exercises and workflow optimization with end-users
Data visualization ideation sessions with data designers.
Because of the time-sensitive business demands of the project, a panel of end-users and subject-matter experts was put in place for rapid, regular feedback. The only way to get back critical feedback was to design quickly, present lightweight artifacts, and solicit frequent critiques from subject-matter experts and end-users.
With the old application, the major complaint was that users spent as much time organizing their workflow as they did executing it. In the mind of the end users there was an established pattern for consuming media from years using applications like Spotify and iTunes, but what needed to be explored was how to translate that idea to the domain of a call-review application.
Also there was a demand from our enterprise client to integrate data visualization into the application. Some of the organization's more industrious power-users had taken to outputting .csv files from the old application and converting them to visualizations in Excel. As an end-to-end product, we needed to do better.
The Result
From our co-design design exercises, we discovered an implemented the following:
A new player layout: Most end-users said they wanted something that looked like iTunes, but in usability tests, the “Spotify” play controls (controls placed at the bottom of the screen) won the day.
Playlist management: End-users preferred playlists to be placed in a hidden drawer panel, and to have agency in accepting new playlists: They preferred to drag them manually to the main queue, rather than have a playlist automatically placed in there for them.
Data Visualization: The ability to drill-down, as well as pivot data were both enthusiastically used, but management end-users still needed the ability to export call data as an Excel spreadsheet for further review purposes.
Deep linking to specific recordings: The idea that a single recording can be a link to a file instead of an email attachment was met with an incredible amount of skepticism from end-users. To quote one “How do I know that the system won’t just erase the file?” We quickly realized that there was still not a lot of understanding of how cloud storage could enhance their workflow.
Multiple Workflow modes: When building playlists, managers had adapted to different ways of building playlists. This meant that we had to take a belt-and-suspenders approach to giving them the ability to build a playlist. In the end about 75% could transfer their old habits to the new interface, while the remaining 25% had to be taught how to build a playlist with the new options. After 2-3 attempts all were able to adapt.
Completed interface displaying playback UI, complete playlist, and corresponding options.
(Click to enlarge)
The pilot was successful with the enterprise customer, and within 30 days it was decided that the pilot could be expanded to all their rooftops. They also reported an average 60% reduction in the time it took for call-reviewers to finish a playlist!