Analytics are used to leverage contact center data in order to resolve problems and improve business processes. Specifically, speech analytics refer to software that can go through millions of hours of recorded text (like e-mails and web chats) and audio to analyze and gather information.
Speech Analytics to Improve Quality Assurance
It’s impossible to evaluate every single agent interaction and it can even be difficult to evaluate a significant sample of conversations. However, since the main goal of analysis is to improve overall performance, taking random samples of customer-agent interaction would be a waste of time. Instead, purposeful targeting allows for only specific, valuable interactions to be assigned to evaluators.
Creating Buckets to Segment Conversations
Creating pre-defined categories, sometimes called buckets, allows contact centers to choose the telltale phrases that are most likely to identify a high-value conversation. One bucket may be for repeat calls and another may be for customer complaints. A bucket may contain just one word, like “ridiculous,” or a series of words and phrases. According to KPMG, in the insurance field, when a customer uses the word “ridiculous” on a customer service call, they are 80% more likely to switch insurers within 90 days, regardless of whether or not the customer service agent is sympathetic. Additional types of buckets may include:
• Repeat calls with the customer indicating they’ve called before. Possible phrases include “called earlier,” “spoke to someone yesterday,” “had to call back” and “called twice already.”
• Repeat calls with the customer indicating their issue is unresolved. Possible phrases include “keeps happening,” “hasn’t resolved it,” “having the same issue” and “never heard back.”
• Calls with complaining or dissatisfied customers. Possible phrases include “frustrating,” “annoying” and “you people.”
• Calls with agents who are not being helpful to the customer. Possible phrases include “that’s our policy” and “can’t give that information out.”
• Calls with happy customers who show their appreciation. Possible phrases include “you’ve been so helpful” and “thank you so much.”
Speech Analytics vs. Post-Call Surveys
Contact centers that have invested in advanced speech analytics software find that the process of identifying and measuring repeat calls is now quicker and more accurate than before. Post-calls surveys have long been a method of analyzing customer-agent conversations, but they’re known to have low participation rates. With speech analytics tools, contact centers can now be proactive when it comes to exploring problems and commonalties across conversations.