October 2001
  

Converge With A New Generation

BY BRIAN MILLER, EMPIRIX INC.

To achieve these benefits, however, companies must be prepared to implement several major upgrade steps to combine and enhance the voice and Web "silos" that exist in most contact centers. These include:

• Computer-telephony integration (CTI) and skills-based routing to automatically "pop" all relevant customer data on the customer service representative's (CSR's) desktop and route a caller to the most appropriate agent based on various customer attributes and CSR skill sets.
   
• Shared databases between Web and voice applications so that regardless of where a transaction is initiated, both customers and agents can be viewing the same data.
   
• Browser-based agent desktops coupled with blended queuing and routing to enable live agent interaction via the Web with voice chat, collaborative browsing or instant messaging, and cross-channel applications such as Web callback.
   
• Monitoring the ongoing performance of applications and infrastructure (just as most contact centers currently monitor caller and agent performance).

Each of these steps establishes significant new quality of experience (QoE) benchmarks that didn't exist with separate silos. Failure to test against these benchmarks can result in failure to identify critical performance and functionality issues that can undermine and, in some cases, completely nullify the ROI for adopting an integrated, multimedia approach to contact center operations.

Even before they make the move to Web-voice integration, contact centers using CTI are finding plenty of coordination and performance problems to keep them worried about maintaining acceptable levels of customer service.

For example, a recent performance test on a contact center for a major online brokerage house revealed that the CTI was taking a full 12 seconds to send the correct screen pop to the correct agent -- far too long to be useful, since the call connection was completed in less than half that time. Both the caller and the agent were left waiting for the data, unless the agent asked the caller to restate the same account number he or she had just finished patiently entering on the telephone touchpad. This is very frustrating to customers and extremely expensive for the company. The ROI for such CTI systems is often based on shaving seconds off calls. Waiting for the data to pop can wash away the entire payback for the CTI investment.

Migrating to a common database for Web and voice applications creates two new QoE challenges. First, there is the added potential for Web load to affect voice performance and vice versa. Failure to test Web and voice applications under real-world, mixed operating conditions can result in sluggish performance that adversely affects both the customer's QoE and the cost of each call. Second, there is a need to verify the consistency of data across channels so customers and agents know that information entered through one channel is instantly reflected in all other channels.

Implementation of browser-based desktops coupled with blended queuing lets customers interact with agents via live chat, collaborative browsing or Web callback. To get the full benefit of live chat or collaborative browsing, customers must be seamlessly routed to the correct agent and the applications must perform under a wide range of load conditions. Web callbacks need to be timely and routed to the correct phone number, to the correct agent, with the correct data and under a variety of mixed load conditions.

Meeting these new QoE challenges is important because failure to do so can eliminate the very benefits that were used to justify the investment. Failed chat sessions can increase rather decrease call loads. Delayed screen pops can increase rather than decrease call length. Failed Web callbacks result in lost sales opportunities.

Even more important is the overall impact of poor QoE on customer retention. Recent surveys of both contact center and Web site users demonstrate that a high QoE generates more transactions, higher spending per transaction and greater customer loyalty. Conversely, those who encounter balky CTI and interactive voice response (IVR) systems -- or channels that cannot scale to meet demand -- are quick to look elsewhere.

The bottom line on customer satisfaction is neatly summarized in the March 2001 issue of the Harvard Business Review, which offers evidence to support the argument that every one percent increase in the proportion of customers who describe themselves as "satisfied" equates to a three percent increase in a company's market value.¹

Taken together, all of these factors combine to deliver three powerful messages to managers of enterprise contact centers that are anywhere along the evolution from call center to contact center:

• To succeed in the current customer communications environment, voice and Web channels must deliver the same data and consistent QoE regardless of what channel a customer uses.
   
• Your customers must be able to switch from one channel to another without worrying that their information or transaction will be lost or interrupted in the process.
   
• Since voice and Web-based applications will be sharing common data, queuing and routing, you need to know how you channels interact with one another across every device and process in your system.

In a converged, multimedia environment, these imperatives require managers to stay on top of a host of new factors. One of these is the scalability of all applications under real-world load conditions: just because a center's voice and Web self-service applications, Web chat, co-browsing and Web callback applications are all working properly in a stand-alone, lab environment does not mean that they will scale together under actual service load conditions. Even if the initial test is successful, however, the rapid rate of application change in both Web and voice applications requires fast, frequent and thorough retesting to avoid the accidental introduction of new performance and interaction problems. Because a true multimedia contact center depends on multiple systems working together, the frequent-upgrade problem is exponentially more complex.

For all these reasons, the role of testing and monitoring is becoming more important than ever. Manual testing is hopelessly inadequate to the task: in addition to being extremely labor-intensive, it simply cannot provide the consistent body of data and frequency of testing required to keep up with change in a complex environment. Yet automated testing is also inadequate unless it can accurately emulate and measure real-world, multichannel traffic loads.

Until recently, the tools to meet these challenges in a converged environment simply did not exist. Often, it has been the customers who have done the testing -- and they have paid the price in frustration and dissatisfaction.

Today, however, the comprehensive, coordinated test tools needed to ensure the performance of multimedia contact centers are arriving in the marketplace just in time to meet the growing need.

These new test solutions help control costs by tracking performance at each stage in a multimedia transaction, allowing managers to look inside what until now has been a black box. Transactions can be tracked across each component of the underlying infrastructure, enabling rapid identification of performance bottlenecks. They permit coordinated, automated scheduling and running of tests that contain both voice and data traffic, and they generate reports that present multimedia performance characteristics in a shared framework and format.

Even in traditional call centers using CTI and IVR technologies, maintaining customer satisfaction is a major challenge. In the world of converged, multimedia contact centers, the challenges and the stakes are higher than ever. In order to meet those challenges, contact center managers must do a far better job of integrating their voice and Web applications. They require automated test and monitoring solutions that ease, rather than add to, that integration process. Fortunately, those solutions have arrived, allowing multimedia contact centers to keep pace with customer expectations and fulfill the promise of efficient, cost-effective -- and seamless -- customer interaction in a converged environment.

¹ Fornell, Claes. "The Science of Satisfaction," sidebar to "High Tech the Old-fashioned Way." Harvard Business Review, March 2001, pg. 120. Fornell is director of the University of Michigan Business School's National Quality Research Center, which produces the American Customer Satisfaction Index.

Brian Miller is general manager of the Contact Center Performance Group at Empirix Inc., a provider of test and monitoring solutions for Web, voice and network applications.

[ Return To The October 2001 Table Of Contents ]