Welcome to the Listening Post!

Customer Service Discussions as a Research Input

Tweet

 

The first of the six key trends that Bruce Temkin discussed in his Clarabridge user conference keynote was that "Customer service is moving from a cost center to a loyalty driver." This thread ran throughout the presentations.

Hal Bloom of Sage pointed out that customer service has an extreme influence, both positive and negative, on satisfaction and loyalty. "When customer service is helpful or knowledgeable, it is typically the number-one positive category influence. When customer service does not live up to expectations, it is one of the top negative category influencers."

To leverage this information, you need to: Capture, Categorize, Connect and Close the Loop.

• Capture: As Diego Lomanto from Verint said, "A conversation with a customer over the phone is one of the richest forms of data that you can acquire." Once you use speech analytics to turn those conversations into text form, you can add them to other textual records of customer service: chat logs, email correspondence, web-form submissions, social-media streams and service satisfaction surveys. Inventory all these records of customer-service interactions and set up an automated process to aggregate these listening posts into one input stream.
• Categorize: While IT is implementing Capture, begin analyzing an ad-hoc sample of the logs to develop a categorization tree. John Georgesen, Ph.D., senior director at Convergys, said, "One of our clients is a cable/satellite TV provider. The customer service group has poor satisfaction levels, and executives felt they were unfairly penalized for issues outside their control. In building code lists, we looked at whether service was breaking down at people, process or technology? It gave us a really rich set of data, so that we could categorize the source of issues and identify what was in the contact center's control."
• Connect: As Michael House, division vice president at Maritz said, "When building a categorization list, connect it with actions. It's not enough to build this construct: you want the meaning of the codes to align with the actions down the road. Connect poor service at retail outlets to the actions they can take."
• Close the Loop: Take action! Help individual customers! As John said, "If you are going down this road [of using questionnaires designed for text mining], you have to put a service recovery mechanism into your research efforts, so that you are escalating issues to a service desk or other escalation team. Look for keywords associated with negative behavior, look for later behavior."

Why go to all this effort? Chris Jones of Intuit put it well. "Never forget the power of one post, of one YouTube video that goes viral. People are much more vocal about our products and services. Every person who has an issue, regardless of how big or small, should be important. If you miss it, it might be the one person you miss who records that video. Every time we let a customer down, we take it personally."