The Customer Service Survey
VocaLabs' weblog providing news and commentary on the challenges of providing good customer service.
Don't Be Fooled!
Thursday - November 03, 2005 02:26 PM in
Peter's entry about the WW2 bombers and the meaning of the term "survivor bias" is one of my favorite ways to explain how easy it is for a well intentioned survey to go very wrong; and why professional assistance in the creation and administration of any study is always a very good idea. When I first began to attend seminars and listen to speeches on how to design and deploy improved customer care services, testing of a service was seldom mentioned.
Today, that has changed, and I am most encouraged that nearly every talk has one PowerPoint slide or entry that acknowledges testing is important. Even so, there is little said about what methods work and the importance of a scientific approach to achieve believable answers. Survey bias takes many forms including the all too human tendency to ask questions that lead to the answers you'd like to get rather than objective results. Bad data is the main culprit as to why companies can provide sub par service because in asking leading questions of a skewed sample of customers they genuinely believe they are doing well. His example of agents controlling which callers get routed to an end of call survey is real. I found it recently when I was having service problems with my home ISP. Angry and frustrated, somehow I didn't get transferred to the survey I said I wanted to take and only later found that the agent got a screen message saying which callers opted to take the post call survey. And just this week I received a call from a major health insurance company who'd just realized that their agents were pre-selecting who got to take the survey and were asking for help on ways to solve this participation bias problem.
Reader, we need you to help spread the word. While most any measurement of service quality can be helpful in making things better, sometimes bad information delivers false confidence. Don't be mislead. When someone quotes a study or performance survey results, ask the hard questions before just accepting the results as gospel. Your customers will be the better for it.
Posted by Rick Rappe
Reader, we need you to help spread the word. While most any measurement of service quality can be helpful in making things better, sometimes bad information delivers false confidence. Don't be mislead. When someone quotes a study or performance survey results, ask the hard questions before just accepting the results as gospel. Your customers will be the better for it.
Posted by Rick Rappe
Posted at 02:26 PM by | | | |