The worst WebRTC demo yet.
I tweeted that the demo AT&T used to launch their new (beta) WebRTC service at CES was the worst I’d seen.
The demo consisted of:
- User browses to flight booking website from a laptop
- User enters flight details
- User can’t get website to book the flight he wants
- User sees ‘talk to us’ button and clicks it
- User is put into a video call with a customer service agent
- Agent asks what the problem is
- User describes the flight he wants and explains that the web site won’t book it.
- User realizes that they need to leave to collect child and presses a button to transfer the call (in audio only) to their cell phone.
(end of demo)
Presumably the user completes the transaction on the phone, whilst driving.
There is so much wrong with this that I hardly know where to start.
The demo seems calculated to make webRTC look useless.
Why would you create a video call ? At best seeing the agent is a distraction, at worst the agent seeing you is a privacy violation. It certainly does nothing to aid the user’s progress to booking a flight. Amazon’s ‘mayday’ has this right, where the agent can see your screen and you can see the agent’s face.
Why does the agent force the user to repeat the details that have already been entered into the web page? This seems calculated to waste everybody’s time and brings no possible benefit to either the user or the agent. All that wonderful context, clickstream data, meal preferences, frequent flier status etc is thrown away.
Why is the agent not able to present options to the user on their mega-pixel laptop display? The demo stays faithful to the 100 year old call model where once you are on the phone you can’t do anything else.
The WebRTC element added nothing to the resolution of the issue, a fact the user demonstrates by transferring the call to mobile voice and seemingly losing no functionality.
The thing that infuriates me is the AT&T did this same demo 2 years ago. They have learnt nothing in the interim. I’m especially irritated by the fact that as The Verge pointed out, the call quality was not up to that of Skype. Why not? Because AT&T have elected to use a 20 year old codec rather than the modern and superb Opus. One can only assume this was because they didn’t want the shoddy mobile call quality to be heard back to back with Opus.
So how does such a shockingly bad user experience get to be center stage at the Consumer Electronics Show?
You have to look at it from the perspective of an telecom equipment maker – The last thing they need is independent web developers seeing that they can integrate real-time voice and video into their user journey without needing to use phone numbers or big iron PSTN networks, or indeed without charging by the minute. They have afterall just sold the MNOs huge heaps of IMS hardware centered around these concepts. So the demo is calculated to keep those elements front and center rather than actually try and meet the user’s needs efficiently.
From that angle the demo seems like a good one, IMS is leveraged, the phone number is the central concept and some minutes are burnt. As a bonus the latest web fad is made to look stupid.
Could I have done better ? Judge for yourselves and take a look at my Fragment demo….