Expect the next five years to be some of the most exciting in the history of the contact center market -- thanks in part to new advances in speech analytics technology.
In a recent report, West Orange, N.J.-based DMG Consulting predicts the quality monitoring/recording liability market will grow 12.5% from $817 million in 2004 to $919 million in 2005. The report contends that the coming year will see significant adoption of speech analytics.
"As you structure unstructured conversations and then mine the information you can take advantage, in real time, of a customer reaching out to you," said Donna Fluss, principal at DMG. "The name of the game is to structure 95% to 98% of what flows through a contact center."
Speech analytics offers three approaches. Keyword detection identifies when a customer uses certain words -- such as "refund" or the name of a competitor -- or complains of a broken product. Emotion detection technology can detect when customers begin to raise their voice and then sends an alert to the contact center manager. Talk analysis technology identifies problems such as lengthy hold times, silence or when an agent talks over the customer.
"The two things we hate the most are being put on hold, and being talked over when we're trying to get something addressed," Fluss said.
Yet the real benefit of speech analytics technology is not just what it brings to monitoring agent performance, but how it brings the value of the contact center to the wider organization. Take, for example, a defective toy distributed at Christmastime. A parent might call into the company to file a complaint. It's a situation that could be playing out at several of the company's disparate call centers. With speech analytics, alerts and reports are generated in real time to flag such circumstances and address the problem before it worsens.
"Now, a supervisor might see a pattern, but it's still very different versus a report that reflects the fact that 10% of the calls were about a broken item," Fluss said. "Today, at best, managers figure it out and call whichever operating group is responsible."
However, if the information comes in a report that is then distributed to sales, marketing or the executive staff, they can craft a standard response. The next customer calling in to complain can then be told a replacement part is on the way, along with a coupon.
Speech analytics, Fluss said, can begin to capitalize on potential cross-sell opportunities, turning the contact center into "full-fledged corporate citizens" as a revenue generator and not just a cost center.
Speech analytics is only now beginning to catch on with large organizations, but it will grow rapidly as companies start to see ROI in six to nine months, Fluss said. By the end of 2004, Fluss counted approximately 35 speech analytics applications in production. However, the vast majority of requests for proposals to large software vendors from their customers include references to the technology.
"It's going to grow a couple hundred percent," Fluss said. "It's not just a value proposition for the contact center; it's a value proposition for the greater enterprise. This is phenomenally rich information because it's coming from the true source -- the customer."
Nice Systems, in Israel, began in the compliance environment, offering applications for organizations that had to record all calls, before moving into the quality monitoring market. In July it released its emotion detection technology and now has 20 implementations, said Eyal Danon, director of sales and marketing.
"Most call center managers listen to about 1% of their calls," he said. "You never know the really important calls to your business. Nice Perform records all the calls coming into the call center and then uses emotion detection, word spotting and call flow analysis."
Businesses are implementing the technology to determine how they are perceived by their customers and also for use in agent training, both what to do and what not to do, Danon said.