By Larry Skowronek
One of the most interesting trends in the analytics space today involves real-time analytics, and rightly so. The ability to analyze calls while they are in progress, and then provide the information the frontline staff can use to take immediate action, is a capability from which every contact center can benefit. The insights gleaned from a call while the customer is still on the line improves the agents and supervisors’ ability to remain compliant with regulations and policies, increase the effectiveness of their sales efforts and improve the overall customer experience.
As the migration of voice systems from legacy TDM telephony to VoIP technology nears completion, the costs for deploying real-time audio analytics have fallen dramatically, bringing the application well within the reach of the broad population of contact centers of all sizes. Now that real-time analysis is a real possibility for companies in any industry, the question remains: How exactly can a company use real-time interaction analysis to achieve the best possible business results as it handles its customer interactions? Following are four key areas:
1. Identifying at-risk compliance situations
One of the most straightforward use cases for real-time voice analysis, applicable (to varying degrees) to all industries, is maintaining compliance, both with external regulations and internal policies and processes. By successfully managing the former, a business can sometimes avoid large fines from regulators, even as it protects its reputation in the marketplace. Just as important, tracking internal compliance allows a company to work more efficiently and become far more capable of achieving its business objectives. How does real-time interaction analytics help companies manage both aspects of compliance? Real-time analytics as an external compliance aid is especially important in industries that are heavily regulated, such as healthcare, financial services and debt recovery. Be it adhering to HIPAA, the Dodd Frank Act or ensuring that agents are contacting the right party and delivering the mini-Miranda and recording disclosures, every regulated business has room for improvement. Where the goal is to maintain compliance with disclosure requirements, the real-time analytics system analyzes the call in progress, listening for the agents to state the required disclosures. If a disclosure is not said in the required time frame, such as the first 90 seconds of the call, the system sends an alert to the agent, reminding them to make the disclosure. If they still don’t deliver the disclosure in, say, the next 30 seconds, the system sends a reminder alert to the agent and an escalation alert to the supervisor. With this information, the agents have the best chance of saying the right thing at the right time. In much the same way, reviewing calls as they are happening can give a company the ability to monitor how its agents are performing in real time to ensure that they are following company guidelines and providing the most effective and profitable service possible. For example, a company could provide alerts to agents when it detects that the agent is building a quote for a customer to remind them to disclose an activation fee (or warn them if it detects that they still have not mentioned the fee after a sale). This disclosure very often will prevent the angry customer call back when they get their first bill. These little nudges have a powerful affect on agent performance. Through real-time analytics, it is entirely possible for a company to monitor and manage potential compliance issues faster and more accurately than ever before.
2. Increase customer retention
For many businesses, increasing customer loyalty is their most important objective. Of course, any program designed to drive improvement in customer retention will have many components. One important retention strategy includes identifying areas of customer dissatisfaction by understanding customer sentiment, then taking pre-emptive action to improve it. Real-time analytics should employ both an acoustic and linguistic approach to the calculation of customer sentiment, with the ability to recognize pitch, tone, cross talk, laughter and, most importantly, the actual words and phrases spoken. Based on these markers, the realtime system flags calls exhibiting negative sentiment and sends appropriate alerts to the agent desktop or supervisor console. These alerts offer instructions designed to save the call, such as guidance on resolving the specific issue or a promotional offer the agent should make. The goal is to pinpoint an issue that can lead to customer defection at the moment it occurs and then address the issue before it’s too late. The science of predictive modeling is another approach to increasing customer retention. First, an analyst uses post-call analysis to identify interactions of customers known to have churned. The modeling process compares the topics discussed on the calls for those customers that eventually left with those discussed by the customers that stayed, finding the topics that lead to, or predict, that the customer will eventually churn. These results are then combined with additional data points such as customer spend and tenure data. Based on this combined set of training data, the analytic process builds a predictive model that prioritizes each at-risk customer. When applied to a call in real time, this same churn prediction model is used to flag interactions with high propensity to churn scores and send alerts to agents and supervisors to take action right away. With this information, companies can implement real-time retention strategies such as special offers or discounts that can significantly minimize customer churn.
3. Increase operations and sales effectiveness
Similarly, the assistance provided by real-time analytics enables agents to handle calls with the confidence they will be far more effective. After post-call analysis identifies the behaviors and statements that are most effective at selling a product, overcoming objections, etc., the real-time system identifies the appropriate times to send alerts to the agents and supervisors, prompting them with specific up-sell, crosssell and next-best offer queues. When used in such a fashion, real-time analytics has practical applications beyond understanding, designing and measuring sales tactics, including driving their improvement. For debt recovery agencies, the use case is similar. The agency that used post-call analytics to understand the best practices for maximizing collections can then use that information to create realtime alerts that provide an agent with the best route to successful collection. As an agent handles a call, she receives alerts guiding her through those best practices to ensure that the proper strategies are being deployed to maximize the call’s effectiveness. For example, the system might detect phrases such as “can’t pay,” “broke” or “not working.” This could trigger an alert to the agent containing guidance on how to position a payment plan to which the customer can agree.
4. Keep real time in context with post-call analytics
Burdening agents with alerts or screen pops for every interesting situation that might occur defeats the purpose of any real-time monitoring program. Agents that receive more than just a few reminders will eventually ignore these alerts, suffering from “alert fatigue.” For real-time agent alerting to be successful, the business must alert agents only to the most important issues and actions, enabling them to handle their calls most effectively. To do this, management must know that the action requested by the real-time alert addresses an issue the business is facing on a large scale (more than just an individual agent basis) and that the requested action will have a meaningful impact. Post-call analytics is the key to the successful prioritization of real-time alerts and requests for agent action. For instance, to optimize customer experience an analyst would examine the full range of call history, understanding which factors indicate negative customer sentiment and elevated stress and their correlation with low net promoter scores. Alerts would be sent to the agent offering techniques to prevent a potentially adverse outcome, including how to be empathetic and reminders to not talk over the customer. Once the business knows how to change agent behavior for a negative customer situation, management can train everyone on the new ways to handle this type of scenario and configure the real-time alerts to support the agents with useful reminder information as they handle those tough customer calls.
As the volume of customer interactions across channels continues to grow, it is vitally important that companies not only take advantage of real-time analytics, but that they use the collected information to enact valuable changes. The key to getting the most from real time, as is true with any sort of analytics, is to take effective action on the findings. With each new insight discovered through post-call analysis, it is important to turn that information into the best practices that are then used during live calls and chats. It’s with that effort that a company can discover just how valuable a tool real-time analytics can be.
Larry Skowronek is senior vice president of Product Management for Nexidia, where he is responsible for the research and development of the overall market strategy, product strategy and development roadmap for Nexidia’s speech analytics solutions. Skowronek has nearly two decades of experience working with contact center tools and management. He has led product management, quality assurance and consulting organizations for contact center management software vendors across the industry.