The Next Generation of Speech Analytics
1 May, 2007
By: Jeff GallinoA company’s ability to monitor, analyze and learn from customer conversations in its contact centers has long been severely hampered by the limitations of available technology. The earliest and most basic approach was simply to pull contact center personnel off the job and have them spend hours listening to a sampling of recorded calls in an effort to identify areas of concern. This took an inordinate amount of time and was only minimally effective. Speech analytics improved on manual monitoring with automated capabilities. But early speech analytics technologies did little more than identify words or phrases, which don’t reveal much about a call’s true meaning or intent. If a customer called an airline and said the word “upgrade,” for example, there was no way to tell if the customer was requesting an upgrade, complaining about not getting one, or simply asking about the company’s upgrade policy.
There has essentially never been a way to find out with any accuracy or certainty why customers are calling, what they’re talking about or how their calls are handled. This begs the question: how are companies supposed to solve problems or address areas of customer concern if they don’t even know what those problems and concerns are? The next generation of speech analytics technology is designed to overcome these limitations. This new breed of solutions makes the leap from knowing a little bit about a small sampling of calls to knowing everything about all the calls. It also provides tools to put that intelligence to work on business improvement in critical areas such as customer satisfaction, agent quality, sales performance and marketing effectiveness.
With this in mind, CallMiner has introduced the next generation of speech analytics with its CallMiner Eureka! solution. This generation differs from earlier speech analytics solutions in the amount of material analyzed, the ability to draw meaningful conclusions from the material and the ability to use the information to effect business change.
Analyzing More Complete Content
When companies can analyze only a small percentage of calls coming into contact centers, they cannot be sure that the calls captured really matter – or whether the calls with the most useful information were missed. The technology behind CallMiner Eureka! uses advanced load balancing to reallocate processing power and cost-effectively analyze all incoming calls, instead of just a few.
Another limitation of previous-generation speech analytics is their sole focus on words and phrases. If a customer ended a call by saying, “You’ve been a big help,” there’s no way to know if it was a straightforward remark – or a sarcastic one made after being put on hold eight times and still not getting a problem resolved. CallMiner Eureka! also takes into account acoustic data, call metadata and information from other enterprise systems to uncover the true meaning of calls.
Uncovering Meaningful Patterns
A hallmark of CallMiner Eureka! is its ability to analyze call content based on a broad set of call indicators – such as competitive mentions, requests to speak to a supervisor, call duration, significant silences, buy language and others – and then categorize calls based on the results.
With this technology, it becomes possible to see at a glance where buy language spiked, for example, and then to drill down into the area of sales performance to assess what happened and what’s likely to happen next. Maybe it’s time to step up production of a particular product? Previous generations of speech analytics solutions are unable to uncover these meaningful patterns.
Applying Customer Intelligence
“Drilling down” in this context means using next-generation speech analytics tools to uncover further useful information about calls. CallMiner Eureka!, for example, provides tools with which to identify variables and target areas for improvement, conduct searches to validate findings and measure and report on the success of actions taken. If a company is conducting a marketing promotion, and calls are showing an increase in product mentions and competitive mentions, but not in buy language, it’s possible to conclude that a competitor is matching the offer – and to come up with a timely, effective response.
Speech analytics has come a long way from spotting a few selected words and phrases in a small sampling of calls, with little chance of being able to draw any useful conclusions from the results. With today’s next-generation speech analytics, such as CallMiner Eureka! companies can fully discover the content, context, purpose and outcome of customer conversations – garnering unprecedented business intelligence that can yield measurable business improvements.
