Kinoo

End of Call Quality Survey

Set the Stage

Kinoo is an interactive co-play video chat platform with a mission to create the best possible future for children, their families, and the world through connection, play, and learning. Kinoo launched in Winter 2021 with the Kinoo Connect app focusing on children 3-7 years old, grandparents, and other older loved ones. Along with the app, we also launched the Kinoo Wand introducing another layer of magic and AR within the app, along with encouraging physical movement and engagement during a call.

Problem to Solve

Designing an end of call quality survey for our users in an effort to gain insight on calling inconsistencies. While our backend team built a system to flag any abnormal activity, we were lacking in understanding what exactly our users were experiencing in these situations. In an effort to understand more, I built a survey based on backend client events tracked during users calls. This resulted in more information which helped inform my design decisions.

My Role

I was the designer and owner of this project collaborating with our lead client engineer and Unity engineer. We had 1 month to complete the project and due to that I have limited documentation.

Design Considerations

Working together with our client engineer, we categorized different user behaviors upon a the end of a call. We noted all unusual behavior or an unsuccessful end to a call and grouped users in 1 of 4 categories based on client events that were recorded:

Nothing unusual: user initiated and ended a call with no seen issues, normal patterns of game play
Unexpected behavior: quick back to back calls, backgrounding the app, force closing the app, call crashing, or users putting the call on hold
Service failures: internet failure, loss of service, how many times app tries to reconnect, handshake fail, loss of Twilio, or a loss of Photon
Call quality issues: ping rates for servers, Twilio video quality, number of disconnects and reconnects, volume info (did they enter the call with no volume), activity troubles or sync issues

I built variability into the system reducing the chance of the user ignoring the questions and motivating them to give us feedback. The design requirements I put into place are:

Only show same questions 3 times
If a user is still having trouble after two weeks time, then bring the questions back
Start with high-level, but offer more specific questions as we learn more overall and more about this user and their calling environment