Using the power of BT smart devices and big data to fix problems before customers notice
A software team from BT told the Big Data Belfast conference audience about how big data is improving the quality of its customer experience using BT network and TV products.
Gareth McKittrick is head of data analytics and reporting for TV and broadband at BT. He reminded the 400 delegates in the IIC Belfast venue that telecoms and TV are very fast-moving technological areas with rapid convergence. The days of customers having a single device in their home are long gone: Devices are now smart, connected, and interdependent. Delivering a consistently good service is important for customer retention in a busy multi-player industry.
Customer helpdesks receive a lot of calls and online interactions. While 80% of calls from customers tend to be related to issues within the home rather than the access network, the diagnostic tools available to customer agents have typically been skewed towards understanding the state of the access network rather than the home.
Back in 2012 the SMART Diagnostics Programme began. Telemetry data from 7 million BT home hubs (broadband routers) and BT TV set-top boxes is collected using the TR-181 standard every two hours. A quality of experience application on devices generates test traffic in order to detect jitter and other customer experience issues.
Recent advances mean that diagnostic data is no longer siloed within specific product sectors but can now be joined up to build better insight into how customers are being affected by any problems they experience.
The analysis is now pushed through to engineers, consumer mobile apps and advisor assistance systems as well as being made visible in operational command centres. As the information and techniques become further embedded in BT’s processes, they hope to be able to pre-emptively resolve some issues before customers see that there is an issue.
“What is normal?” asked Séainín McCoy, the lead engineer responsible for the big data design and architecture within the TV and Broadband team. “We knew what bad was; we knew what good should be. But ‘normal’ can be quite different in a low activity home when compared with a high usage household.”
She outlined how BT’s patented Brandenburg algorithm identifies patterns in the device data to detect whether the fault is in the network or whether a home visit from an engineer was required to fix a router fault. An algorithmic calculation of line length along customers’ products and usage is used to cluster comparable users together into peer groups. Each customer is given a WiNscore (What is Normal score) based on the number of standard deviations they are away from their peer group’s ‘normal’ performance.
The WiNscore metric can be viewed at an individual level, street level, or exchange level. Knowing the health of a street or the health of other smart products within a home allows intelligent decisions to be made about the dispatch of engineers when a whole street begins to report a problem.
As well as its uses within service assurance – “taking us away from just reacting to what a customer says [about a fault] but allowing BT to get out in front” – the data can also be used to make product recommendations to customers based on their historical pattern of use and performance.
“It’s putting the customer in the middle of what we do” says Mckittrick. “We shouldn’t always rely on the customer coming to us, but when they do we should maximise the conversation we have with them.” People want their problems dealt with quickly and don’t want to stick on the phone while agents ask them questions they should already know the answer to.
Another validation of BT’s work is that customer service agents feel that the WiNscore information is helping them with their jobs: “they’re happier to come to work with better information”. As McKittrick explained, “it’s not an easy being an agent dealing with people who have a problem and want a fix quickly. They say this is making their life easier and it’s also helping the engineers who are out in houses and have access to the same information.”
Smart devices in your home can help utilities deliver better services to you and your neighbours using the power of big data.
