Smart Twilio Voice Insights and dumb service providers
Twilio announced its Voice Insights Service for WebRTC this is API exposure of call statistics plus some extra tricks. Every SIP trunk provider should offer this including incumbent providers. The truth is however they don’t get it. They don’t understand the extra value that they can provide to their customers. That exposing this information makes them more useful and more compelling to their customers. Over the top providers like Twilio will continue to grow and to steal market share if the incumbents cannot start to innovate.
The truth is even worse than this to be honest. A lot of providers don’t offer any compelling reporting tools beyond basic call records. Often the records don’t even contain quality information. This isn’t because the carriers are trying to hide the information from customers. Although this may be the case. No, it is far worse than that. A lot of carriers actually aren’t able to report on this.
A lot of carriers don’t even collect QoS data!
The service from Twilio exposes real time network stats including MOS, RTT, jitter, packet loss, audio levels device types so on. (Full List here). The tool will proactively notify when quality drops off allowing user notification and action. Additionally they are adding pre-call quality checks. (This is a cool feature which has been implemented in a few places and I’ll probably write a post on later.) Finally, you’ll be able to collect user feedback post call and Twilio will perform analysis on it. This means you can optimise delivery methods to clients based on location, time and device. Perhaps it makes more sense to use a standard PSTN dial out when the wifi connection is poor for example.
Now, there are some carriers that do offer MOS, QoS and other call information but as far as I can tell there are very few that offer anything as detailed or real time. Many might include reports delivered at the end of the month or billing cycle. This is well and truly after the horse has bolted. How can you identify a network issue or application fault with a 30 day feedback loop? You can’t! At least not effectively. There is no reason (besides organisational inertia) that SIP trunk providers couldn’t allow customers to access this information in real time. This would be a compelling service for many. Think how beneficial it would be for contact centres to combine real time voice quality information on their links with the information they already gather on their network. They could proactively route calls to alternate sites. Route data over alternate links etc.
Also, a lot of these carriers are simultaneously trying to become providers of other applications and services. Why don’t they combine this information like Twilio is doing? Capture the voice quality, tie it into location and other information that they the carrier has access to and provide this dataset to customers. The customers could then send feedback forms etc. Or better yet the carrier could do it themselves and provide the analysis as a service.
The crux of the problem however is that the incumbent carriers don’t understand the power of the information that they have at hand and they have no idea how to turn it into useful products. I worked with one that did want to offer reporting however they pretty much wanted to double the cost of the service to provide the reporting. Plus it was going to be end of the month reporting limited to pretty graphs and static tables. Twilio’s offering is a fractional increase in cost. Not a monumental one. If carriers want to be more than dumb pipes they need to start smartening themselves up.
Views expressed here are my own and do not necessarily reflect the views of my employer Oracle.