Fundamentals of Contact Center Analytics

Transcription

1 Fundamentals of Contact Center Analytics A Verint Systems White Paper by Diego Lomanto, Senior Solutions Marketing Manager Bill Durr, Principal Global Solutions Consultant Verint Witness Actionable Solutions Verint Systems Inc. 330 South Service Road Melville, New York USA info@verint.com VERINT

2 Table of Contents The Struggle for Customer Retention Analytics: Innovative Solutions for Competitive Markets Speech Analytics Stage 1: Processing from Audio to Data Stage 2: Refining the Data Stage 3: Analyzing Data to Make It Actionable Data Analytics Finding Root Cause Customer Feedback Management Capturing the Voice of the Customer Beyond the Contact Center: Enterprise Benefits from Analytics-driven Workforce Optimization....8 Unauthorized use, duplication, or modification of this document in whole or in part without the written consent of Verint Systems Inc. is strictly prohibited. By providing this document, Verint Systems Inc. is not making any representations regarding the correctness or completeness of its contents and reserves the right to alter this document at any time without notice. Features listed in this document are subject to change. Not all functionality is available in all configurations. Please contact Verint for current product features and specifications. All marks referenced herein with the or TM symbol are registered trademarks or trademarks of Verint Systems Inc. or its subsidiaries. All rights reserved. All other marks are trademarks of their respective owners Verint Systems Inc. All Rights Reserved Worldwide. WASG U

3 Preface Many businesses work hard to stay connected with their customers through advertising, loyalty programs, special events, surveys, and other activities. In doing so, they often overlook the value of the customer interaction data captured in their contact centers each day. Too often, this data is used solely to track internal performance metrics, which while important seldom provide insight into customer opinions and the overall customer experience. This paper highlights the underlying technology and value of contact center analytics, including speech analytics, data analytics, and customer feedback surveys. These solutions can help centers mine customer interactions for actionable information on customer opinions, product and service issues, competitive intelligence, and more. By sharing this insight across the enterprise, contact centers can enhance the service they offer, more effectively drive revenue and contain costs, better coach and manage their staff, and deliver greater value to their parent organizations. About Verint Witness Actionable Solutions Verint Witness Actionable Solutions is the leader in analytics-driven workforce optimization. Its solutions are designed to help organizations capture customer intelligence, uncover business trends, discover the root cause of employee and customer behavior, and optimize the customer experience across contact center, branch, and back-office operations. Verint. Powering Actionable Intelligence. Verint Systems Inc. is a leading provider of actionable intelligence solutions for an optimized enterprise and a safer world. More than 10,000 organizations in over 150 countries rely on Verint solutions to perform more effectively, build competitive advantage, and enhance the security of people, facilities, and infrastructure. 1

4 The Struggle for Customer Retention One of the ironies of today s customer service industry is that many organizations in spite of the vast amounts of data collected in their contact centers, branch offices, and Web sites each day are more removed from their customers than ever. It s not as if organizations are unaware of the problem. In fact, many businesses work hard to stay connected with their customers through advertising, loyalty programs, special events, surveys, and other activities delivered across multiple media. Research firm COLLOQUY has estimated that U.S. membership in loyalty/rewards programs reached 1.3 billion by the fourth quarter of 2006, 1 while Gartner forecasts that the customer relationship management industry will grow by 14.2 percent this year, with revenue surpassing $8.9 billion U.S. 2 These projections represent a significant corporate commitment to and investment in their customers. In spite of these efforts, customer satisfaction has eroded steadily. A recent Web-based survey conducted by Ventana Research showed that only 56 percent of the 250 respondents from around the globe measured customer satisfaction with the handling of calls, and that nearly half (49 percent) said their customers are less than satisfied with the results of their calls. 3 In the U.S., the American Customer Satisfaction Index (ACSI) 4 registered its first gain in a year in May 2008 a paltry 0.4 percent to stand at 75.2 on the ACSI s 100-point scale. 5 Although a small increase in customer satisfaction is a step in the right direction, Claes Fornell, founder of the ACSI, observed in a May 20, 2008 press release announcing the index results that falling housing prices, tight credit, and rising food and fuel costs are making it difficult for satisfied consumers to spend more even if they wanted to. Fornell points out, The smart move for companies in the economic environment is to make sure they keep the customers they have by shoring up their customer relationships. 6 Loyalty programs and CRM systems play an important role in customer retention, but they may not be enough to sustain, let alone grow, a customer base in challenging times. Organizations need more powerful solutions in their customer satisfaction arsenal solutions that can capitalize on the customer data they already have, since data appears to be the one thing that most organizations possess in abundance these days. Fortunately, such solutions exist. Contact center analytics software solutions that include speech analytics, data analytics, and customer feedback management can provide dramatic benefits for enhanced customer service, improved efficiency and performance, and greater competitive advantage. They are the next advancement in customer service. 1 CensusTalk: Quo Vadis: Sizing Up the U.S. Loyalty Marketing Industry, Rick Ferguson and Kelly Hlavinka, COLLOQUYtalk, April Market Trends: Customer Relationship Management Software, Worldwide , Sharon A. Metz, Gartner, March Customer Experience Management: Improving the Consistency and Quality of Customer Interactions, Ventana Research, June A national economic indicator of U.S. household consumer evaluations of products and service quality, the ACSI factors in scores for government agencies and about 200 companies in 43 industries. 5 American Customer Satisfaction Index, theacsi.org, May 20, ACSI: Customer Satisfaction Halts Slide; Glimmer of Hope for the Economy? ACSI, May 20,

5 Analytics: Innovative Solutions for Competitive Markets As the focal point for customer interactions, contact centers typically collect data to assess the performance of agents, help ensure organizational compliance with industry regulations, and resolve disputes. These activities are all internally focused. Contact center analytics use a different approach by examining captured interactions to detect customer perspectives on products, services, and more, as well as search for the root cause of specific performance factors. Let s take a closer look at how contact center analytics work, and how they can benefit your organization. Speech Analytics When it comes to customer interactions, the most fundamental questions asked of any contact center manager are: Why are customers calling us? Why do customers continue to call us? While most contact center managers can probably answer the first question, albeit using generalizations, they most likely would not be able to answer the second. Speech analytics can provide the necessary insight by converting conversations from unstructured audio data into data that is structured in an index that can be readily searched and analyzed. This activity takes place during three stages: Stage 1. Processing from audio to data. Stage 2. Refining the data. Stage 3. Analyzing the data to make it actionable. The technologies used during each stage affect the functionality that can be provided in later stages. This is a key consideration when evaluating speech analytics solutions. Stage 1: Processing from Audio to Data In the first stage of speech analytics processing, the conversation is converted from speech (audio) into data. There are two primary methods by which this is performed: phonetic-only processing and Large Vocabulary Continuous Speech Recognition (LVCSR). Phonetic-only processing systems analyze the sounds contained within the conversation and look for strings that represent words. A phonetic process requires training the speech recognition system to know what sounds phonemes make up a word. There are approximately 40 distinct phonemes in any language. Advantages: The key advantage of phonetic-only processing systems is that they can process calls quickly in stage one. Drawbacks: A key drawback of phonetic-only processing is that the processing engine understands only sound there is no context or meaning provided. This can lead the solution to single out words or terms inappropriately ( can sell instead of cancel, for example.) This phenomenon is known as providing a false positive. The burden of applying context and ruling out false positives falls upon the user. 3

6 Another key drawback is that end-user searching for calls (stage-three processing) is very slow. This seemingly minor distinction has major usability implications, since business users may become frustrated by the time involved in locating useful calls. Moreover, because phonetic-only systems save and search on only predefined keywords in stage three, only a fraction of content is used for analysis. These systems also limit the kind of analytics that can be performed in later stages of processing. LVCSR systems process calls with phonetics, then convert the results into full text by leveraging a 60,000+ word dictionary to establish a meaning for each word. Pure LVCSR solutions usually need their generic dictionary expanded with industry- and company-specific words. Unlike phonetic-only systems, LVCSR systems understand the actual meaning of words. Using the dictionary, the software recognizer scans the conversations and creates a searchable index. Advantages: The key advantage of LVCSR processing is that it permits root-cause analysis in stage two. It also allows users to search for calls much more quickly in stage three. LVCSR systems achieve higher levels of accuracy and present fewer false positives than phonetic recognition systems. Drawbacks: The primary drawback to LVCSR processing is that stage-one processing (converting audio into searchable data) takes longer than with phonetic-only systems. Stage 2: Refining the Data In the second stage of the speech analytics process, data provided from stage one is refined to make it usable. What makes one speech analytics system more accurate than another is not only its ability to convert interactions from unstructured audio to structured data, but also the type of data mining (also known as speech mining) it conducts on this data. Stage two is where the primary differences between different vendor systems exist. The only information processing that phonetic-only systems conduct is to create a list of the user-predefined words detected in the processed conversations. LVCSR-based systems can conduct a deeper analysis. However, some LVCSR systems simply count words and use a frequency of terms as a means of determining the key reasons for customer calls. This method can overlook comparable terms or phrases that might indicate potential issues or problems. More value is delivered from a solution that uses phonetics and LVCSR, while also providing an additional, specialized layer of speech mining to count frequency, detect where words are spoken, and note their relation to other words contained within the conversation. Verint Witness Actionable Solutions Impact 360 Speech Analytics solution works in this manner. Its speech-mining application creates a Complete Semantic Index of all processed calls to provide meaning and context, helping business users quickly detect trends and patterns within customer interactions. The Complete Semantic Index is the amalgamation of several proprietary speech-mining algorithms that have been researched and developed to categorize calls and provide a root-cause analysis for each category. The Complete Semantic Index compares the subset of calls that you are interested in (cancellations, for example) against the content of the rest of your calls. It then searches for a relevancy score and correlation to other terms. For example, the term cancel might be highly correlated with competitor names, or the terms that represent specific product features or process issues. The Complete Semantic Index can discard terms that are common in both the subset and superset of calls, such as words like hello. The addition of these technologies to Impact 360 Speech Analytics powers features such as TellMeWhy, which provides automated root-cause analysis. 4

7 Stage 3: Analyzing Data to Make It Actionable Once conversations are converted to structured data and an analysis/refinement method applied, the third stage of the speech analytics process is initiated. This stage involves analyzing the data and requires an interface for users to extract actionable information. The simplest interface provides keyword spotting. Nearly all speech analytics solutions offer this functionality, which is a simple search for words or phrases that are spoken in customer interactions. Retaining the entire content of the call enables users to search for almost any term or phrase in millions of calls and quickly receive a prioritized list of results. Categorization provides a more sophisticated level of analysis by using call content to classify large volumes of calls into defined categories, such as customer complaints, billing issues, product feedback, or repeat calls. Categorization can provide quantitative information (such as an increase in customer complaints about late shipments), as well as qualitative information (such as a set of calls in which customers state that they are switching vendors because of a better pricing offer). Categorization requires a category-building wizard to guide users through the process of building categories by suggesting terms and automating processes. LVCSR solutions can provide categorization; phonetic-only systems cannot. An effective category-building wizard, in addition to the Complete Semantic Index, can reduce the time it takes to build and train categories from weeks to hours. The most sophisticated speech analytics solutions, such as Verint Witness Actionable Solutions Impact 360 Speech Analytics, can provide root-cause analysis that goes beyond simple search and categorization by leveraging advanced data-mining technologies to show why customers are calling. These solutions can analyze a set of categorized calls (such as calls related to billing issues) and automatically surface underlying issues, regardless of whether the user has defined these issues or root causes in advance. Dividing recorded conversations into root-cause categories makes it easier and less time-consuming for the person reviewing the calls to take action on the issues identified. Since contact centers are dynamic in nature, root-cause enabled speech analytics can continuously categorize and mine calls when they take place, surfacing new root causes. Few other speech analytics solutions provide the ability to conduct true-root cause analytics to tell you why. The layering of these three analysis capabilities delivers the true essence of speech analytics. An effective speech analytics solution provides the ability to search unlimited terms and phrase combinations, categorize calls based exactly on how customers express themselves, and mine specific categories and call subsets to automatically surface the root causes that drive these calls and impact overall performance. 5

8 Data Analytics Data analytics tends to be poorly understood, largely because it is often confused with data analysis a typical activity in many contact centers, where center personnel use spreadsheets to track and chart a wide array of key performance indicators (KPIs) that can vary widely from one center to another, such as: Queues, agent, and team productivity from the automated call distributor (ACD) Agent time utilization and activity management from the workforce management system Individual agent and team quality scores from the quality monitoring team Agent time and attendance Absenteeism Skill development scores CRM system data While the sheer quantity of data about what is happening in the contact center can be overwhelming, it typically provides very little insight into why the center is experiencing these outcomes. Lacking this insight into root cause, organizations may have difficulty knowing which actions to take to prevent specific problems from happening again. Finding Root Cause Unlike data analysis, a data analytics solution doesn t just track and report data it provides new information from it. For example, Impact 360 Data Analytics leverages Verint s patented data mining technology to process the extensive volume of data generated by the contact center and surface specific call scenarios that may be helping or hurting overall performance. It can be compared to having an army of analysts run every possible report with every possible filter and parameter to uncover findings with the greatest impact. Going far beyond presenting KPIs, the solution provides data that can help centers: Uncover hidden service and quality issues in the contact center. Reduce costs by pinpointing actions that harm productivity. Increase revenue by uncovering conditions favorable to up-selling and cross-selling. Increase first-call resolution by detecting key reasons why customer issues are not resolved. Identify star performers and underachieving agents. Establish a return on investment for initiatives to improve performance. Impact 360 Data Analytics uses all of the data associated with a call to help isolate the root cause of contact center performance. When a metric is selected for further investigation, Impact 360 Data Analytics presents it as a prioritized list showing specific call scenarios that could positively or negatively impact performance. These call scenarios, or findings, can point to the root cause of an underlying business problem. By automatically detecting the key call scenarios most likely to affect performance, the solution can eliminate the need for users to know which questions to ask. As a result, instead of simply measuring a KPI such as average talk time, Impact 360 Data Analytics can show what drives average talk time. It could be a type of customer at a certain time of the day or month, an agent group, or even a specific agent that is abnormally inflating the overall score. Impact 360 Data Analytics can bring that previously hidden knowledge to the surface so that it can be acted upon. Estimation tools help model the impact of corrective action to establish a priority for action. 6

9 Impact 360 Data Analytics presents findings in simple tables, visual scorecards, and straightforward text, rather than in complicated spreadsheets or Web-based reports. Scorecards are easily configurable so that managers can focus on the subset of data most relevant to them, and various groups within the organization can be provided with access to the metrics, goals, and data-mining findings appropriate for their roles. Customer Feedback Management For years, contact centers have used a number of techniques, including mailed questionnaires, interactive voice response (IVR) surveys, and live telephone conversations (either immediately after the transaction or later, via a call-back) to solicit customer feedback about the company, its agents, and its products and services. Apart from what to measure and how to go about asking the questions in the right way, customer feedback processes present an additional problem: a cost/time trade-off. Mail surveys are fairly inexpensive but require considerable time to obtain results. Telephone surveys using live interviewers are expensive and require time to obtain statistically valid numbers. There is also the problem of reduced response rates due to the time lapse between when the transaction took place and when the survey is performed. The search for a middle ground led to the use of IVR technology. IVRs are well-suited for branching questionand-answer survey construction. One approach is to transfer the caller to an IVR script upon conclusion of the call. Another approach is to ask the caller initially if he or she will agree to take a short customer satisfaction survey after the call. The up-front agreement helps ensure that no matter how the transaction turns out, the customer s perceptions are captured. The single biggest problem with all of these approaches is the participation rate. People are reluctant to take surveys when interacting with contact centers. Mailed surveys frequently produce a response rate of less than five percent. Call-back surveys have a better response rate than mail, but the small sample size afforded in this approach seldom produces results that are applicable down to the individual agent level. Even post-call, IVR-based surveys typically have participation rates that are very low and bi-modal only customers whose experience was really good or really bad tend to make the effort. The low participation rate generally is attributed to survey length. With few customers agreeing to take a survey at all, pollsters are primed to ask many questions of those who do. But in doing this, they tend to annoy the survey takers and actually discourage the very participation they are seeking to promote. Capturing the Voice of the Customer Impact 360 Customer Feedback is a browser-based, scalable solution that uses short, dynamic surveys to capture information from customers across the IVR, Web, and in real time. Flexible and highly reliable, the solution is installed on site (behind company firewalls using internal security policies) and interacts directly with existing telephony and company networks to capture customer comments. Surveys are presented based on user-defined business rules after the completion of the call or /web contact. Because these surveys are context-based and timely, they can engage customers and deliver response rates far greater than those obtained through traditional, one-size-fits-all questionnaires. Impact 360 Customer Feedback organizes survey questions into a library to allow them to be easily reused in other surveys across multiple channels. Templates enable users to create surveys easily using question types from the library or to develop questions of their own. 7

10 Unlike traditional IVR surveys, Impact 360 Customer Feedback can use survey responses to automatically trigger alerts notifying the appropriate people in the organization of particular circumstances requiring immediate attention, such as potential customer defections. Alerts can also trigger workflows a predefined series of activities that must be performed or acted upon within a prescribed period of time. Links to actual call recordings can enable contact center managers to drill down to specific calls for further analysis. Having gathered critical information, Impact 360 Customer Feedback presents it in reports based on the specific people, products, and processes that each customer experiences. It knows who the customer is, who the agent is, the transaction type the customer has undertaken even the revenue opportunity involved. Reports are easy to create and can be ed directly from the Customer Feedback interface, so they can be shared throughout the enterprise. Reports include: Survey Results Reports Provide overviews and detailed views of response and completion rates, question details, individual and aggregated responses, access to recorded calls, and more. Performance Analysis Reports Summarize effectiveness metrics for agents, supervisors, and managers, helping you identify strong and weak performers. Quality Reports Correlate and plot internal (company-driven) and external (customer-driven) metrics, illustrating discrepancies between the two to depict areas of strength or weakness. By gathering data that might otherwise be overlooked or lost, Impact 360 Customer Feedback offers insight that can enable contact centers to respond quickly to customer comments. As a result, organizations can do more than simply measure customer satisfaction they can influence it. Beyond the Contact Center: Enterprise Benefits from Analytics-driven Workforce Optimization Contact center analytics offer organizations the opportunity to gain deep insight into the concerns and sentiments of their customers a key advantage in customer retention and enterprise performance initiatives. But the benefits and return on investment of contact center analytics can be increased further still. When deployed as part of a unified workforce optimization solution suite, contact center analytics can capitalize on the other functionality within the broader suite. The result can be powerful functionality that offers unprecedented insight into contact center operations and can extend into other areas of the enterprise, including branch, remote, and back-office operations. Verint Witness Actionable Solutions Impact 360 Workforce Optimization solution is an example of a unified suite. In addition to Impact 360 Speech Analytics, Impact 360 Data Analytics, and Impact 360 Customer Feedback, the Impact 360 Workforce Optimization suite includes solutions for quality monitoring and recording, workforce management, performance management, and elearning. Separately, these individual solutions have value in the contact center. But when deployed together under a unified framework, they offer organizations analytics-driven workforce optimization a new way to improve their entire customer service delivery network. Armed with analytics-driven workforce optimization, enterprises can gain a formidable solution to help them improve performance, competitive advantage, and the customer experience while controlling costs even in challenging economic times. 8