Use Speech Analytics to Reduce Calls That Frustrate Customers and Hurt Productivity


Share on LinkedIn

The global economic downturn is likely to cause business leaders to cut budgets in all areas, including contact centers. Does the customer experience need to suffer? Not necessarily—if cuts are made with a scalpel instead of a machete.

Analyzing the audio from recorded customer calls can reveal opportunities to fix the root cause of problems, which can save money and improve the customer experience at the same time. In other words, it’s in your best interest to listen to the “voice of the customer” by listening to the actual voice of the customer. No survey required!

Untapped Resource

If you’ve ever called a customer service department for help, you’ve probably heard a message like this one: “Your call may be recorded for quality assurance.” Perhaps you’ve wondered, as I have, what happens to those call recordings.

At Blue Cross of Northeastern Pennsylvania, Bob McDonald wanted to learn why customers were calling. Sure, agents can note in their records that customers were calling about a benefit, claim or other issue, but as McDonald, director of customer service, said, those notes didn’t “give me actionable items to work on.” So in 2007, he turned to a speech analytics solution from Nexidia to “mine” all of the audio recordings and get to the root cause of calls.

Take, for example, the extremely high call volumes Blue Cross experienced in the fall of that year. With the help of Nexidia’s tools, McDonald could validate that, in one case it was the result of something anticipated—a recent system change. The data gave him “ammunition that the problem really needed fixing.” In another case, McDonald discovered, much to his surprise, that customers were circumventing processes to get faster service. Customers are clever that way! Armed with this insight, McDonald changed the call flow and improved agent training.

From Emerging to Surging Market

Speech recognition and analysis technology has been around for quite some time, with applications in government, security and call centers. But contact center guru Donna Fluss of DMG Consulting figures that 2004 was the year it “burst into the commercial world,” with a grand total of 25 implementations. By 2007, that figure had grown to more than 1,200 implementations.

Despite the down economy, Fluss forecasts in her 2008 Speech Analytics Market Report that the market will expand by 70 percent in 2008 and 50 percent in 2009. A significant driver of this growth is “ROI that is absolutely real and compelling,” resulting in a payback in less than a year if done properly.

It’s not just about saving money. Fluss contends that speech analytics can also improve the customer experience—a key factor in customer loyalty and retention.

Despite the down economy, Fluss forecasts in her 2008 Speech Analytics Market Report that the market will expand by 70 percent in 2008 and 50 percent in 2009.

Contact center analyst Keith Dawson of Frost & Sullivan shares this optimistic outlook. However, he sees speech analytics as part of Business Intelligence, rather than a separate market space. Contact centers have traditionally been concerned with improving agent performance, including staying “on message” and offering the correct up-sell or cross-sell offers. Speech analytics can help in these areas, of course, but Dawson notes that most call centers and marketing departments have not “parsed the customer side” of calls. Now technology makes it possible to understand how customers are responding to the agent, including their emotional state.

Vendor Landscape

Whether or not speech analytics is a standalone market is an interesting debate, but the real issue is whether vendors have solutions that can deliver business benefits. Fluss counts some 22 vendors now and expects more to come. As with most any other segment of the IT industry, you’ll have to decide whether a “full suite” or “best of breed” solution is right for you.

Multifunction vendors, such as contact center stalwarts Autonomy, Envision, NICE Systems and Verint, include speech analytics in conjunction with other solutions including workforce management and quality monitoring, says Dawson.

NICE Systems has been investing in a range of speech technologies—including phonetic search, speech-to-text transcription and emotion detection, says Barak Eilam, president of the Interaction Analytics business unit. He believes organizations should look take a unified approach across interaction channels: call center, web, IVR, email and chat.

Verint sees speech analytics as a part of “workforce optimization,” according to Daniel Ziv, the company’s vice president of Customer Interaction Analytics. Generally companies only review a subset of recorded calls to ensure agents are doing their job correctly. With rapid advances in computer processing power and speech recognition technology, Ziv says companies can now mine all their calls (or at least a much larger subset) to identify customers at risk for defection and find the root-case for calls.

Specialist vendors include CallMiner, Nexidia and UTOPY…and many more. Marketing Vice President Jeff Schlueter says Nexidia’s claim to fame is a patented phonetics technology that can rapidly index and search audio. Customers can use the solution on a standalone basis or in conjunction with systems from such companies as NICE Systems and Verint. As you can see from the screen shot here, one of the applications is to understand call drivers.

Colette Yee, UTOPY’s Director of Strategic Marketing, touts the company’s SpeechMiner solution as best-of-breed. What makes it better, she says, is “phrase-based” recognition technology which is more accurate than phoneme- or word-based schemes. UTOPY can help improve agent performance, but can also be used for other enterprise applications like churn prediction and assessing product popularity.

Smart Practices

Industry experts and vendors agree that the challenging economy could accelerate the growth and adoption of speech analytics in contact centers. The main driver is the need for operational efficiency, but increasingly business leaders recognize that the customer experience is critical, too. Even when consumers are spending less, service quality will factor into their buying decisions.

To have the best chance for success and getting a fast payback, consider these four tips:

  • Given that budgets will almost certainly come under pressure, plan to use the insight gained to make wise choices about where to cut costs, where to fix problems and which customer segments need different treatments.
  • Because this is still a relatively new market, make vendor decisions carefully. As I mentioned earlier, there are the usual trade-offs between full suite and specialty vendors. Likewise, the technology “under the hood” does matter, so invest the time to understand which methodology fits your requirements best.
  • Speech analytics experts and vendors give much the same advice that applies to CRM projects: Take small steps toward your grand vision. Pick an application to learn from, prove out the ROI and then expand.
  • Most importantly, Fluss counsels you to ensure your speech analytics project is staffed appropriately. Done right, you can get a return in three to nine months, but don’t assume it’s an install-and-forget kind of application.

Budgets cuts may be a painful reality in the coming year, but you don’t have to share that pain with your customers. Use insights gained from speech analytics to improve the customer experience while operating your call center more efficiently.

Then, when the turnaround comes, your customers will have stuck around. Better yet, they’ll be ready to spend more money with you and not your competitors.

Resource: 2008 Speech Analytics Market Report


Please use comments to add value to the discussion. Maximum one link to an educational blog post or article. We will NOT PUBLISH brief comments like "good post," comments that mainly promote links, or comments with links to companies, products, or services.

Please enter your comment!
Please enter your name here