For anyone buying a smartphone today, the burden for supporting that device typically falls with the network operator, who is responsible for making sure that the phone works well. In the days of feature phones, this meant dealing primarily with call quality and battery issues. Today, technical support means resolving sluggish performance, application problems, battery drain, configuration issues and bill shock due to unexpected data usage.
As smartphone sales have grown, the impact to customer care has been dramatic. The length of time to resolve each technical support call has, on average, risen by ten minutes and the number of escalations from Tier 1 to Tier 2 call agents has doubled.
Source: Carrier IQ customer survey, July 2011
This has had a direct impact on support costs and profit margins – a fact highlighted in numerous quarterly earnings reports by major operators.
What can be done?
At the center of the cost growth lies one essential problem – the amount of time, effort and expertise required to understand each consumer’s problem and then resolve.
Much like a doctor, the care agent has to gather a complex list of symptoms from the patient (the consumer). Unlike a doctor, the care agent has very few diagnostic tools to verify the symptoms and to make a diagnosis.
What is missing is the expert information necessary to allow the care agent to assist the consumer quickly. Not surprisingly, the best place to gather details on device performance and consumer experience is directly from the mobile device. Gathering performance metrics from the mobile device and then analyzing this information provides expert knowledge to care agents, which can be used to identify and help resolve the consumer’s technical problem.
Loading diagnostic agent software on mobile devices prior to shipment, usually at the request of the network operator, allows for this insight. Once in the hands of a consumer, the device works as normal and, with user consent, periodically submits metrics on the performance of the device. Areas covered include call quality, data throughput, device stability, battery life, configuration, and application performance.
Once gathered, the data is analyzed to understand what, if any issues are being experienced and to provide details of this experience to care agents. When a consumer calls in for technical support, the care agent is able to use this diagnostic information to identify the consumer’s problem quickly, then resolve or escalate the problem based on the information provided.
Call Triage
When a call first reaches the Tier 1 Technical Assistance Center, the first task is call triage – understanding what type of problems the customer is experiencing and identifying if the call can be solved at Tier 1 or if Tier 2 escalation is required and to which group or department. An experience summary can allow the care agent to identify which areas of device performance appear to be causing issues and assist the consumer quickly and appropriately.
Understanding normal vs. abnormal performance
In every network and with every device, there is an accepted (if not acceptable) performance and behavior of networks and devices. For example, how often do calls drop in your town or how long does the battery in an Android phone last? Understanding “normal” is important in the context of support calls because the care agent needs to understand the consumer’s experience relative to other users.
Insight gathered directly from devices helps aggregate the performance of the network by defined area (block, city, region, country) and devices type (make, model, firmware version). This average is then matched against the caller’s details which can then be used to determine the caller’s experience and define what is or is not a “bad” experience.
Making the knowledge actionable
This information, coupled with machine inferencing of the data allows triage to move forward by categorizing the customer’s experience into network performance, battery life, application issues, system crashes and so on. This allows the Tier 1 agent to address the issue directly – “It appears you have 70 applications loaded on the phone and 10 of them are running and draining your battery” – or pass the call on to tier 2 or tier 3 – “I am sorry I can see you are having issues making calls from your office, let me have one of our network engineers call you back.”
When the call does need to be escalated, having the knowledge and insight to track what happened and indeed why it happened requires providing the data behind the machine inferencing to allow human inferencing to take over. Contacting application providers, analyzing crash logs, and examining layer 3 messages between the mobile device and tower all from a set of data delivered from the device and never previously available to answer questions. This data goes a long way to help improve the performance of the network and devices and therefore impact customer satisfaction and churn rates.
Unlike other solutions, which allow care agents to “log-in” to a mobile device to view settings and try to understand an issue in real-time, accessing information from the previous ten days activity enables a complete picture of the user experience to be understood. In this way, problems that occur infrequently or can’t be repeated during a support call can be captured and understood.
Savings
By providing immediate knowledge of a consumer’s issue and insight on the likely cause to Tier 1 agents, call times are radically shortened and the number of calls escalated to Tier 2 or Tier 3 is reduced. If calls are escalated, the details of the consumer’s problem and further data – such as the signaling between the radio tower and the device are captured for further analysis.