The Customer Service Survey
VocaLabs' weblog providing news and commentary on the challenges of providing good customer service.
What's a Test?
Tuesday - January 10, 2006 04:34 PM in
It happens a lot. We mention that VocaLabs is in the business of testing customer care and if the other person is in the industry the response is often: "Oh yah, just like such and such company." Well no. In general terms, customer service applications need to be tested for technical performance, in particular load or stress testing to make sure the design will continue to function when loaded up with traffic. Often this is done in a somewhat risky manner by the use of a pilot phase/controlled introduction in which the design is phased in. Other entities, notably CT Labs (now a part of Empirix) have figured out how to mitigate the risk of using real customers as test call guinea pigs by simulating traffic into the functioning (but not yet activated) application. Then there is technical performance monitoring on an ongoing basis, and a leader in this sphere is IQ Services who monitor the system for throughput including such items as making sure all the phone lines are in service.
These are important functions, but as we like to point out, a system can be performing just fine from a technical perspective, but still not serving the caller well. This ease of use, or if you prefer, caller satisfaction test is the Vocalabs niche.
By our broad definition of usability, anything that keeps the caller from having a positive experience and easily completing the reason for their call is where we focus.
So while we might uncover a technical issue, it is only because such a glitch impacted the caller's ability to conduct their business. Most often what we find are design issues such as scripting and/or commands that might have seemed logical to the system architect, but which confuse the caller. Sometimes this is the use of insider jargon, sometimes an off-putting persona in an IVR, or perhaps an inability of a speech recognition application to understand a caller's verbal response.
In the end, all these tests are critical to fielding and keeping a good system operating at peak service levels, and the best test regimens incorporate all three tests, both during design/install AND on an ongoing basis.
Rick Rappe'
By our broad definition of usability, anything that keeps the caller from having a positive experience and easily completing the reason for their call is where we focus.
So while we might uncover a technical issue, it is only because such a glitch impacted the caller's ability to conduct their business. Most often what we find are design issues such as scripting and/or commands that might have seemed logical to the system architect, but which confuse the caller. Sometimes this is the use of insider jargon, sometimes an off-putting persona in an IVR, or perhaps an inability of a speech recognition application to understand a caller's verbal response.
In the end, all these tests are critical to fielding and keeping a good system operating at peak service levels, and the best test regimens incorporate all three tests, both during design/install AND on an ongoing basis.
Rick Rappe'
Posted at 04:34 PM by | | | |