Understand What Customers
Are Saying Regardless of Language

By Cliff LaCoursiere

June 2008

Shifting demographics are changing the mix of spoken languages in the U.S.  The 2000 U.S. Census Bureau survey found that 18 percent of the total population, or approximately 47.0 million people, spoke languages other than English.  These figures were up from 14 percent, or 31.8 million, in 1990 and 11 percent, or 23.1 million, in 1980.  These trends are expected to continue and will compel contact centers to offer service in multiple languages.  This adds an additional layer of complexity to a business' understanding of their customers' needs, and it can challenge their ability to achieve key performance metrics.  Sophisticated speech analytics technology that supports multiple languages can positively influence not only contact center operations, but also business processes as a whole.

Background: The Benefits of Speech Analytics: Speech analytics can be employed in contact centers to identify key topics that point to customer dissatisfaction, such as naming a competitor or asking to speak with a supervisor.  The more advanced speech analytics solutions capture silence, tempo, and stress in a call along with topic identification algorithms.  These applications enable businesses to understand what customers are saying as well as the intent of what they mean, providing a complete picture of all calls.  This capability enables businesses to identify customer dissatisfaction trends before they become issues that could dramatically affect the company's bottom line.  In addition, these solutions produce reports that allow businesses to respond quickly to create and implement programs that will satisfy and retain customers before they decide to take their business elsewhere.

The following is an example of how leveraging speech analytics technology could have helped a leading U.S. communications services provider.  This company reported losing $720 million in 2005 due to customers canceling their services.  While most of those customers did not cite customer satisfaction issues as the main reason, a sizable amount did attribute their cancellation to competitive providers that offered better pricing and better products.  If this company had used speech analytics to uncover why customers were leaving, then some percentage of those customers might have been saved, and company-wide tactics could have been implemented to save future customers as well. 

Using the above communications services provider example, implementing speech analytics to modestly improve customer satisfaction by one quarter of one percent would save $5 million in revenue from dropping off the bottom line.  Increasing customer satisfaction and reducing cancellations by a meager 1 percent boosts that figure to $20 million. 

Taking It a Step Further: Speech Analytics Technologies in a Multilingual Contact Center Environment: As you can see, speech analytics technologies have the ability to enable contact centers and the businesses they support to reduce costs, increase revenues, and improve customer satisfaction.  Adding multiple languages into the mix makes achieving customer satisfaction goals a daunting task unless the speech analytics technology can address the issue of multiple languages. 

This can be accomplished with speech analytics technology that uses the appropriate language and acoustic models to drive customer interactions into the database for analysis.  However, while some speech analytics vendors support multiple languages, they do so in a monolithic fashion and only support one language recognition model per application.  Switching language applications is usually accomplished through the contact center's IVR.  More advanced speech analytics solutions, on the other hand, are using a combination of IVR and CTI data to trigger the correct language application. 

To benefit outsourced contact center providers that need to support multiple languages and accents, oftentimes via multiple contact centers, in particular the speech analytics solution implemented should be able to tag the origin of the contact center audio and execute the appropriate language model for converting those recorded calls into data that can be mined.  The ability to switch between languages and accents is critical for businesses that support a diverse customer base with multiple languages.  This gives the outsourced contact center a flexible tool to deliver value-added services to their clients.  This capability is particularly of interest to outsourced contact centers since the value-added services have good profit margins and can differentiate the outsourcer from those that don't offer sophisticated call analytics. 

Future implementations of speech analytics will possess the capabilities of handling different languages spoken on a call automatically by having the application recognize the language and each word to trigger the appropriate function. 

Undeniable Results: It's clear that speech analytics can positively affect contact center - and overall business - operations.  By providing the ability to analyze every aspect of every customer interaction, regardless of the language being spoken, sophisticated speech analytics technology that supports multiple languages enables call centers to better understand all of their callers' needs, wants, and concerns.  With this knowledge, companies can then act accordingly to enhance customer satisfaction, reduced churn, and improve customer loyalty.  This, in turn, can help companies reduce their overall costs and increase revenue.

Cliff LaCoursiere is a founder and senior vice president of business development at CallMiner, provider of quality management, analytics, and call center solutions.  CallMiner is a speech analytics solution that extracts customer intelligence from recorded conversations.