Google's Duo Video Calling With Knock Knock

22 May 2016

Google had its I/O conference this week, hosting its presentation of its latest announcements and outlook on what can best be described as a pop-star concert stage. I think the venue was a mistake but the presentation itself was markedly better than previous years. Clocking in it at two hours, the Google IO keynote is finally down to an acceptable length. Just a couple of years ago, they would run two 3 hour presentations on consecutive days.

One thing they unveiled was a FaceTime competitor called Duo. Specifically, there’s an element which struck a chord with me. When videocalling someone else, the recipient sees a live stream of the caller’s video as it rings. One side of the video call has already begun at the moment of the phone ringing. The other person can then pick up the call to start the two-way video, seamlessly transitioning into the conversation as the video of the person on the other end is already live.

It’s a fantastic streamlining of FaceTime. They also emphasised the instantaneous nature of the protocol allowing the two participants to community immediately after the call is confirmed. FaceTime’s usage model is a lot colder. One person asks to call someone else, the recipient sees the person’s name and a static image. When the call is answered, the video streams attempt to initiate a connection, which involves staring at a Connecting indicator for a few seconds, before finally succeeding to allow the two people can see and talk to each other.

The current FaceTime flow is as bad as a traditional phone call, which is basically what FaceTime is (in the same way iMessage is a 1:1 reproduction of SMS transmitted over the Internet). With Duo / Knock Knock, the call has effectively already begun as soon as the phone screen lights up on the receiving end.

Google showed how the caller could signify intent during the time waiting for the other person to respond. The user on the receiving end can pick up context from the Knock Knock video stream, such as where the person is, what they are doing or who they are with. Google showed potential with examples of people holding up movie tickets, engagement rings or simple facial expressions like happiness or sadness. (That being said, the product video — embedded above — did not do a good job of expressing the possibilities tastefully; it is too cheesy and felt too forced).

Aside from the speed and practical advantages, it’s also just damn cool to send your face to someone else. If the feature turns out to be gimmick, it encourages more people to do video calls in general, even if its just the novelty of how it works. I think it gives a meaningful benefit to picking the video option over audio, though. Even if they decline, you can imply something in those couple of seconds that would never happened otherwise. It’s almost like a transient Snapchat selfie with the opportunity to commit to a full conversation.

It’s a user experience thing that I hope Apple adopts. There are obvious knee-jerk fears of the dangers of letting people put live video onto someone else’s screen without explicit consent. I think these issues are easily mitigated by decent policy design, such as a (default) preference to only ‘enable Knock Knock for people in my contacts’. Careful attention will have to be given to the interface for callers too, especially early on, to explain what is happening — make it plain that the other person can see what you are doing right now even though you can’t see them yet. These are solvable social and technological problems and the benefits are huge, in my view.

Slight confession: I meant to write this post the same day as the event. I ended up being lazy and didn’t get to it until today. I’m glad I waited though, as it let me focus on what I was actually interested in. Almost subconsciously, my mind has concentrated on a couple of specific things.

Out of Google’s entire keynote, I can easily recall just two announcements: the Instant Apps demos and Knock Knock. Everything else is a vague blur or forgotten. Instant Apps is a technical quagmire with a lot of questions about implementation and its utility, so I’m holding off on judgements until its more set in stone … although the premise is intriguing. Duo is more concrete, complete with product videos, and made me genuinely excited. Alas, neither of these announcements have solid release dates, unfortunately. I can’t wait to check out Duo and Knock Knock sometime “later”.