How do we put an end to human error in our live video broadcasts?
Why Did I Prioritize This?
Our team had recently implemented the interactive broadcast solution, Tokbox, in our Major League Baseball app for mobile and web. The show is called Edward Jones Chatting Cage, and it presents live, participatory interviews with players and important figures from baseball.
Unfortunately, the implementation required administrators to add and remove metadata to tags on media posts as Chatting Cage changed state before, during, and after every presentation. Errors were being introduced, causing the show to fail.
Since this was a high visibility feature, we needed to immediately address the issue.
Problems To Solve
I met with engineering to better understand the nature of the problems we'd been seeing, then met with the men and women who were charged with managing our broadcasts. Detailed ID numbers had to be entered at just the right time and it was all too easy to forget to make meta-data changes at the appropriate moment.
Rather than changing meta-data on the fly, we needed a tool in our Brand Administration Platform to automate and simplify the process.
Chatting Cage Program
I documented our admins’ user flow, noting areas of potential error. Quickly sketching out ideas on paper was the best way to iterate through possible solutions. The goal was to make setting up a video broadcast as simple as possible, while constraining the process to reduce all possibilities of error.
After the initial discussion with engineering I’d understood this to be a multi-step process, however, while working through it on paper, I realized it could become a single step process by contextually disabling certain fields to insure that information was entered in the appropriate order. After confirming this was a workable solution with the appropriate stakeholders, I added error states and notifications to prevent introduction of new forms of user error.
Next I went over the proposed solution with the men and women who would actually use it, getting a thumbs up to move forward.
Once we were all in agreement, I fleshed out the design in Sketch, wrote the functional requirements and JIRA tickets, uploaded the Sketch pages to Zeplin, and then prioritized this project for the tool team’s upcoming sprint.
The team was able to deliver their update to the Brand Portal right away. After implementation, our error rate dropped to virtually zero.