As more organizations seek to uncover the true voice of the customer, the speech analytics (SA) market seems to be finally heating up, according to the findings of multiple research and consulting firms.
“Heating up” shouldn’t be confused with “overheating,” however. In Saddletree Research’s 2009 survey of National Association of Call Center members, 57 percent of respondents still had absolutely no interest in the technology. On the bright side, that figure represents a whopping 13-percentage-point drop from 2008. In another sign of the technology’s uptake, DMG Consulting’s 2009–2010 Speech Analytics Market Report found the SA space had expanded from 25 commercial implementations at the end of 2004 to 1,764 at the end of 2008. And Ovum expects SA spending to grow at a compound annual growth rate of 14 percent between 2009 and 2014.
Hardly a brand-new concept, SA is moving from a nice-to-have technology to one that has clear, tangible benefits when tied with other contact center technologies, including workforce optimization (WFO) tools. Donna Fluss, president of DMG, says vendors are now doing a better job “helping end users find the needle in the haystack and more easily…apply what they’re gleaning out of speech analytics.”
Paul Stockford, president of Saddletree Research, says there’s an inherent optimism in the contact center industry despite the recession, but that SA’s penetration is driven by a more-practical reason. “Companies are looking for every possible advantage to not only maintain business with current customers, but find new ones,” he says. “If you can mine data from current ones, you can use that to build your business or keep things moving in the right direction. That…is enough to raise the interest of people holding the budgets during these difficult financial times.”
And just who are those budgetary bosses? According to DMG, the contact center continues to be the primary buyer of SA, though more departments—including marketing, sales, operations, risk management, and collections—are also beginning to use the technology. According to Saddletree, the two verticals most interested in the technology are financial services and third-party outsourcers.
Companies interested in SA must first answer two major functionality questions, analysts say. The first is to choose one of two SA approaches: large-vocabulary conversational-speech recognition (LVCSR) versus phoneme-based SA. “There’s a huge difference,” Fluss says. “Phoneme-based analytics are so much easier than LVCSR. Fewer servers, less time, less setup.”
The second question may be the real selling point for SA, says Aphrodite Brinsmead, associate analyst for customer interaction technologies at Ovum: whether the SA solution should be standalone (players here include Aurix, CallMiner, Nexidia, and Utopy) or part of a WFO suite from vendors such as Envision Telephony, Nice Systems, and Verint Systems. “The market is becoming less about differentiating technology by LVCSR or phoneme-based approaches, but rather more about packaging the technology around core business processes,” she recently wrote. “Enterprises will begin to understand that SA is not an emerging technology but a practical business solution.”
The solution doesn’t come for free, though: Contact centers can expect to pay between $600 and $1,000 per seat (plus maintenance) for SA, according to Stockford. Given the sticker price, Fluss urges companies to allocate someone to truly own the SA deployment. “The reality is that you have to tune a speech analytics application—no matter which underlying engine you have—and tweak it so the results you get out of it make sense,” she says. “If you’re not going to be able to dedicate a resource to it, odds are you’re not going to realize the ongoing benefits.”
Shopping for Speech
Autonomy etalk (snipr.com/autonomyspeechanalytics)
Envision Telephony (snipr.com/interactioniq)
NICE Systems (snipr.com/niceinteractionanalytics)
Verint Systems (snipr.com/impact360speechanalytics)