QM Picture Grows Blurry
The quality monitoring (QM) marketplace is an enigma. At first blush, a mature market appears—these products have been shipping for decades—rife with commodity products. But a closer inspection reveals a range of new capabilities (speech analytics, compliance functions, multiple communications channels) being woven into solutions that could dramatically improve customer service and stimulate sales.
“Despite being well-established, the QM market continues to spawn innovation, as well as new entrants,” notes Donna Fluss, president of DMG Consulting. In fact, more than 40 companies offer such products, including Aspect, Autonomy, Calabrio, CallCopy, Envision Telephony, Interactive intelligence, Oaisys, Nice Systems, and Verint.
Call recording and QM solutions have been shipping for years, and observers estimate that 80 percent of all contact centers have implemented at least a rudimentary system. The products have become popular because every contact center strives to improve customer satisfaction by delivering more comprehensive and consistent support, and QM has that potential.
At a basic level, the products serve as coaching aids. Supervisors review completed calls with agents, praising them for items handled well and offering tips for exchanges that could have been processed more skillfully. While the objectives are clear, reaching them is not. Through the years, how call information is captured, what benchmarks are established, and how feedback is delivered all have varied, often dramatically, from one contact center to the next. Given the array of new market forces, this disparity is expected to widen in the future.
Don’t Sit So Close to Me
Sometimes, a supervisor sits next to an agent, listens in on a series of calls, and grades the agent’s responses. However, that process may not be an accurate performance barometer because some do not act naturally when they know they are being observed. Instead, they might perform uncharacteristically well or falter under the pressure.
To address that limitation, products offer companies other ways to examine customer interactions. Monitoring might be done remotely with the supervisor surreptitiously listening in on random calls, grading them, and later providing feedback to the agent. Another option is to have the QM system, say the Envision Full Time Recording module, collect all of the calls and have the supervisor manually sift through them in search of appropriate exchanges.
However, the call evaluation process often was quite time-consuming. “A supervisor would have to sit through or sort through dozens of calls or dead airtime just to find one coachable exchange,” says Dick Bucci, a principal at Pelorus Associates. In that case, managers would spend only 10 percent of their time reviewing and coaching, with the other 90 percent spent searching for meaningful exchanges.
Speech Analytics Gains Ground
Facing ever-growing competitive pressure, corporations have been seeking ways to reduce the search time, so supervisors could devote more time to other important functions, such as targeted training or business processes. In response, vendors have tried to automate call evaluation, focusing on speech analytics. Recent technical advances have heightened interest in those efforts. Speech recognition systems have improved, so their accuracy rates are good for most callers, albeit various accents still pose problems.
Voice recognition systems structure contact center conversations and feed them into analytical applications that identify customer needs, wants, and insights. The system selects key words, such as “problem,” “trouble ticket,” or “refund.”
The underlying metrics are becoming more sophisticated. For instance, the Verint Impact 360 Speech Analytics module includes a patented Complete Semantic Index, which is designed to detect various emotions, and a self-learning Smart Category Wizard that tries to pinpoint company-specific audio interactions. Then, using programmed intelligence, such solutions automatically determine which calls to retain so a supervisor can review them later. In addition, the tools give companies the option of saving the entire call or just a portion of it, another feature designed to increase supervisors’ efficiency.
The Best—and worst—of Times
These new capabilities represent a good news/bad news scenario for customers. The positive news is that these systems make finding coachable moments easier than they were in the past; however, the new capabilities are not a panacea. First, the price for the speech recognition products is usually high, starting in the six figures and in some cases well past the million-dollar mark.
“We are interested in speech analytics, but it is a bit out of our price range at the moment,” says Ash Megalli, manager of Contact Centre Operations at ICBC, which provides car insurance to individuals and companies in British Columbia.
Another vexing challenge comes from implementing the systems. “Speech analytic systems have a lot of capabilities, but they require a lot of fine-tuning,” Pelorus Associates’ Bucci notes. Companies need to correlate key word selections to vague concepts, such as customer satisfaction. And once a company determines how to set up the system, its work is just beginning.
Companies cannot simply deploy the systems and let them run, because the analysis of the results is dynamic. Like selecting and prioritizing search system key words, speech analytics systems constantly must be fine-tuned. In many cases, the process requires full-time, in-house speech analytics experts, which would increase a contact center’s personnel costs.
Because of the various limitations, speech analytics implementations are in a nascent stage of development. “Only about 10 percent of customers are using speech analytics with QM,” Bucci notes.
Telling the Full Story?
In addition to being inefficient, the old QM approach did not accurately measure agent performance. With the historical technique, supervisors examined only a portion of an agent’s calls, so the accuracy of the sample size was questionable. Increasingly, QM systems collect all calls.
Another challenge stems from human nature. Supervisors forge positive and negative bonds with employees, so their evaluations are rarely objective. Some companies have established QM professionals, whose job is to listen, identify calls of interest, and coach employees. Again, such a move would raise personnel costs, which could bring resistance. Some companies justify the hiring of more employees by the benefits gained in freeing up supervisors’ time and improving agent interactions.
Regarding the latter, businesses are trying to gain more insight into how agents handle calls, via another recent phenomenon: the inclusion of keystroke monitoring in QM systems. Rather than just look at what the agent says to a customer during a call, software like the Oaisys Tracer monitors what screens are open, what data are presented to the agent, what information the agent inputs, and how the employee moves from one screen to the next. With such capabilities, companies can present their agents with helpful information without cluttering their computer screens.
Compliance Issues Arise
The screens are becoming cluttered because agents are being asked to do more now than they have in the past. For instance, the number of government and industry regulations has grown, making compliance a contact center concern. New laws and industry regulations touch on privacy, outbound calling, and mandatory disclosures.
In many contact centers, agents collect sensitive personal information, such as credit card data or Social Security numbers, and rules have sprouted that outline how that data is handled. Many QM systems now offer businesses ways to encrypt the information. In some cases, for instance with Calabrio Compliance Recording and Quality Management, the agent turns encryption functions on and off. Some systems now automatically encrypt certain fields, secure transmission of recordings as they travel across the network and sit in storage, and deliver audit-trail reports.
While helpful, such steps may fall short of some compliance requirements. For example, to be Payment Card Industry (PCI)–compliant, companies must ensure that sensitive information is not captured or stored at all. As a result, more solutions include automatic pause/resume functionality, in which call-recording systems would shut off whenever certain words are spoken.
Focus Remains on Voice Calls
Yet another area of emphasis with QM systems is their multimodal capabilities. Even though electronic customer exchanges emerged more than a decade ago, most QM systems concern themselves only with voice calls.
“Many companies want to respond to their chat and email exchanges in a more consistent manner,” says Kathy Krucek, product manager for quality management at Aspect. QM systems have been enhanced with reporting templates and best practices so that companies can manage these exchanges in a consistent manner. For instance, with Aspect’s Unified IP, online customers establish a time and a number, so they are contacted through a Web callback function.
The focus on electronic communications tools has sparked interest in real-time QM coaching, says Matthew Storm, director of innovation and solutions for Nice. The Nice Enterprise solution lets organizations understand customer intent. Using embedded presence and speech analytics, supervisors can identify high-risk calls as they arrive and help agents find the appropriate resources to address concerns immediately.
In addition, companies use the multimodal features to more accurately measure customer satisfaction. Traditionally, contact centers have relied on key performance indicators, such as average handle time or call abandon rates, to determine quality. Companies assumed that if their contact centers scored well in those areas, customers were happy and their needs were being met, but that may not always hold true.
New Feedback Avenues
“Businesses are using new tools to gather more feedback from customers,” says Brynn Palmer, director of the customer experience at Verint. For instance, ICBC uses email surveys to validate such impressions. “We want to make sure that the customer has the view of the experience that we do,” Megalli explains.
Along those lines, the contact center, beyond fielding customer inquiries, is assisting departments like marketing and sales. While answering a call, an agent can upsell other products, such as premium movie channels to cable TV customers. In addition, they can take advantage of new social media technologies to enhance their brand. With CallCopy’s cc: Survey, companies can determine which customers will actively promote products and services on their social networks.
As companies use their contact center to generate additional revenue, they need greater visibility and connectivity into other applications, such as supply chain and Enterprise Resource Planning solutions. One challenge is that this data typically resides elsewhere in the organization. Consequently, companies need to scoop data from these sources and set it aside in a central data mart, where it can be extracted and manipulated.
Data Integration Challenges
A couple of factors complicate this process. First, companies house two types of data: structured (information in a database) and unstructured (free-form information, like emails). Businesses need tools to integrate the various data sources. Since few standard interfaces are available, the customer usually has to tweak the tools to get them to work together.
Moreover, the company must figure out how to analyze and correlate information. “Companies have lots of data; the challenge is trying to make sense of it,” Verint’s Palmer notes. As with speech analytics, companies need to assign personnel to the task so that it is done well.
As companies try to expand in these areas, they face organizational issues. “One big question is: Who owns the data?” DMG Consulting’s Fluss says. “In many cases, marketing wants to take charge of these interactions even though they take place in the contact center.” As a result, infighting can occur as departments try to protect their turf.
In sum, the QM market is evolving beyond its historical function of providing agents with feedback on problem calls. It is becoming a powerful tool that companies can use not only to improve the customer experience but also to enhance the enterprise’s bottom line. Hurdles remain for businesses to clear to make the transition, but those willing and able could reap great rewards.
Paul Korzeniowski is a freelance writer who specializes in technology issues. He can be reached at email@example.com.