Comparing Contact Center Performance Management and Speech Analytics

By Rob Berry

Over the last few years, significant advances have been made in call center management solutions to improve the performance of contact centers. Among the new solutions available, contact center performance management (CCPM) and speech analytics are two solutions receiving a great deal of attention. Both offer significant benefits to call centers, their management and staff, and their clients. The challenge for many contact center managers and executives in evaluating these two technologies, though, is that they both promise similar business benefits. However, the business problems they solve, the degree and source of the benefits they deliver, and the deployment risks of each are quite different.

According to Gartner Research, a leading industry analyst firm for the contact center market, CCPM solutions are most often used by contact center management to automate the supervision and coaching of an agent and to improve the overall performance of the organization. Gartner Research explains, “These solutions integrate an organization’s established contact center technologies, CRM systems, and other data sources to provide a transparent picture of performance through role-specific dashboards and reports, and they drive actions through embedded alerting and workflow capabilities.” (Gartner Research, January 2008). The key to unlocking the true value of CCPM is deep, meaningful analytics that are actionable. Without these analytics, contact center supervisors and managers will only have a high-level view of performance, not the granular detail and insight they need to realize the full benefits of their CCPM investment.

On the other hand, speech analytics tools are primarily used by quality management (QM) analysts and, to a limited degree, contact center supervisors. Speech analytics can search for examples of calls that meet specific criteria, such as dead time, mention of competitor, or an antagonistic tone. According to Gartner Research, speech analytics is a tool that “incorporates a range of capabilities based on an analysis of a live or prerecorded voice stream. Capabilities include:

  • Searching an audio file to find a segment of interest without the need for manual indexing of the content
  • Identifying whether a specific event took place during a conversation, based on monitoring for specific phrases or keywords
  • Applying data-mining operations (such as filtering, clustering, categorization, pattern matching, and conceptual search) to audio streams
  • Using verbal and nonverbal cues to identify a speaker’s characteristics, such as emotions and truthfulness” (Gartner Research, July 2007).

An important consideration in the evaluation of whether to select CCPM with advanced analytics or speech analytics is the issue of sample versus census data. Speech analytics vendors point out that with their products, all calls can be analyzed. While true in principle, the reality is that the price of the software and hardware that is required to analyze every call recording is far too high for most organizations. Unless required by governing bodies, few organizations actually record every call. Most industries only record 10 to 30 percent of calls, and oftentimes these calls are purged regularly (thirty to forty-five days on average) to free up storage space for newly recorded calls, thus further minimizing the usefulness of speech analytics. Speech analytics is therefore limited by design in the maximum percentage of calls that can be analyzed and in the short amount of time they can be accessed.

CCPM with advanced analytics is generally considered the foundational solution for agent coaching and training, improving first contact resolution (FCR) rates which reduce unnecessary call volume, improving call center sales effectiveness, and customer and employee satisfaction. In most cases, companies will start with CCPM to address the higher return on investment areas like FCR and sales performance, and then they will deploy add-on technologies such as speech analytics to focus on areas like word spotting and script adherence. CCPM generally adds value more quickly and directly by analyzing data across all channels and affecting people and teams in the call center. CCPM solutions do not require recorded calls for their analysis. Therefore, they can provide an accurate and actionable view of contact center performance. Additionally, CCPM solutions store several years of data in their data mart – even long after the transactional details have been purged or archived – to provide access to substantial historical analysis.

Both CCPM with advanced analytics and speech analytics are next generation solutions that leverage contact center data to improve service delivery; however, there are fundamental differences in the problems these two technologies solve and the benefits that can be derived from them – although they sound similar on the surface.

CCPM is a management tool that uses multiple systems to create agent-level key performance indicators (KPIs) and enables supervisors to rate agent performance and identify opportunities for improvement based on these KPIs – typically improving targeted metrics by 5 to 20 percent. With best-in-class CCPM systems, which include automated call reasoning, an FCR module, and the ability to link to voice recordings, supervisors are able to identify agents that are struggling with particular call types or who cause repeat calls, and go directly to specific calls during training. Because it is grounded in KPIs, CCPM is not well suited for random analysis of calls, voice-dependent analysis such as dead time and emotion, or the spotting of keywords spoken in particular calls as part of a QM process.

Speech analytics is essentially an analyst search tool that allows analysts to search for examples of calls that fit a particular profile such as high emotion, mention of a competitor, and mention of a prior call; however, speech analytics solutions reject the vast majority of calls that are not an obvious match. Therefore, speech analytics is not designed to accurately tag 100 percent of calls with a call reason or to assess FCR for use in performance management – rather, it is designed to find a few good examples of each.

While speech analytics is an intriguing new technology, its early adopters have discovered that is more difficult to solve specific problems than they initially expected. Best-in-class CCPM with advanced analytics provides contact center supervisors, agents, analysts, and executives with the tools necessary to reduce the cost of unwanted call volume and improve customer satisfaction, sales effectiveness, and agent attrition.

Both CCPM systems with advanced analytics and speech analytics solutions utilize data gathered in contact centers in way that answers questions, solves business problems, and improves performance. The main difference is that each technology solves a different business problem. Speech analytics accesses only verbal communications and is best suited to aid analysts in quickly understanding the high-level nature of a call or in finding specific keywords mentioned during a call. CCPM utilizes switch, customer/CRM, agent desktop, workforce management (WFM), HR, and other data for a complete view of each interaction. CCPM solutions directly affect performance by driving targeted agent coaching and identifying the root cause of initial or repeat calls.

Given the differences between CCPM solutions and speech analytics solutions, the evaluation and selection of one of these two solutions should ultimately be made based on which business problems have the greatest potential value to solve, as well as determining the ROI hurdle for the organization.

Rob Berry is senior director of product marketing and business development of Enkata. He has more than ten years of experience in enterprise analytics, with a heavy emphasis on the call center.

[From Connection Magazine October 2008]