• February 29, 2008
  • By Jeff Gallino, cofounder and CTO, CallMiner

How to Find Out What Customers Really Think, Want, and Hate

In a time of cutthroat competition and razor-thin profit margins, keeping customers and cultivating their loyalty is critical to any company's success. But how do you find out what you need to know about your customers to make sure you're doing everything possible to keep your relationships with them strong? In a word: Listen. It's not as simple as it sounds. For years, companies have been using methods such as customer surveys, exit interviews, and call monitoring to get customers to tell them whether they're satisfied, whether they're thinking of leaving, or, if they have left, what drove them away and what could bring them back. The bad news is, those methods don't work particularly well in and of themselves. The good news is, there is something else that can help: advanced speech analytics. The Trouble With Traditional Approaches Surveys, interviews, and call monitoring share a weakness that limits their value for determining what's on customers' minds. They're all, by nature, one step behind. For example, if a customer satisfaction survey says a customer was happy with a service experience last week, that doesn't mean that customer isn't fuming and thinking about leaving this week. Or if a survey shows that a customer had a bad service experience last week, it may be too late to do anything about it now. Exit interviews have similar limitations. What good is it to find out why a customer left after the customer is already gone? And how eager is an unhappy former customer going to be to answer a long series of questions from a company he or she may be in a big hurry to forget? If the customer participates at all, the answers are likely to be perfunctory, inexact, and affected by the customer's state of mind at the time of the interview -- something an interviewer may not even pick up on. Both surveys and exit interviews also have the limitation of only representing the self-selected customers who choose to participate. Other, more-valuable intelligence may rest with customers who elected not to respond. The monitoring of customer calls to contact centers, on the other hand, presents the opposite problem. With thousands of hours of content available to monitor, this approach may provide too much, not too little, information. Companies end up listening to a small sampling of calls to try to find information of interest -- which circles back to the problem of not enough information. The odds that they will end up listening to the calls that contain the most-valuable information are slim when they're listening to only a limited number of the calls recorded. It all comes down to a case of too little information that's likely delivered too late to do much good. That's what makes speech analytics so attractive. The Power of Advanced Speech Analytics Speech analytics was developed as a way of automating the process of manually listening for relevant information in calls to contact centers. The earliest versions -- some of which are still in use -- simply searched for pre-selected words and phrases in calls. Today, advanced speech analytics can do much more, going beyond words and phrases not just to reveal what customers are saying, but also to assess how they're feeling, and to determine whether they're having problems that could threaten a company's relationship with them. The most effective advanced speech analytics solution will: Not just analyze what customers are saying, but also assess their state of mind by analyzing non-speech characteristics such as how fast or how loudly they're saying it -- even taking into account the meaningful silences in their conversations with contact center agents. Construct patterns of meaning based on indicators such as mentioning competitors, reacting negatively, exhibiting high levels of stress, or overtly stating intentions ("I'm taking my business elsewhere!"). Identify patterns and trends across calls over time, such as whether a particular indicator of customer dissatisfaction (like asking to speak to a supervisor) is happening repeatedly and if it's related to a particular issue or agent. Have the technology to quickly and cost-effectively analyze the content from more calls -- 100% of calls, ideally -- to ensure that valuable customer intelligence isn't left behind with calls that went unscreened. The information provided by advanced speech analytics makes it possible for a company to understand trends in customer satisfaction and dissatisfaction as they develop. This puts the company in a position to take action to address these trends before they negatively affect the customer relationship. Since customer churn has a costly impact on companies, preventing churn and finding more-effective ways to win customers' loyalty and solidify relationships with them is critical. When surveys, interviews, and call monitoring aren't enough, adding advanced speech analytics to the mix can make a difference. About the Author
As chairman and chief technical officer of speech analytics solution provider CallMiner Inc., company co-founder Jeff Gallino oversees current product development and future product direction.
CRM Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues
Buyer's Guide Companies Mentioned