How advanced analytics can help contact centers put the customer first

| Article

More and more, to put the customer first, the heads of customer service require the accurate and detailed performance information that real-time analytics can provide. The good news is that basic data and analytics tools are becoming standard practice in call centers. And while that is a solid first step, most organizations are likely not taking full advantage of the technology, meaning they are not applying advanced analytics in ways that truly put the customer first. Only 37 percent of organizations feel that they are using advanced analytics to create value; this finding reveals significant missed opportunity. Unlike earlier data and analytics solutions, which helped companies understand what is currently happening within their call centers, advanced analytics help them generate actionable insights about what will happen next, through both internal and customer-facing applications. The result is reduced costs, increased revenue, and—most important—higher customer satisfaction scores.

In this way, advanced analytics has fundamentally changed the role of contact centers from a basic service offering (and a net cost to the business) to a strategic differentiator that can make dramatic improvements in customer satisfaction and financial performance. Companies have already applied advanced analytics to reduce average handle time by up to 40 percent, increase self-service containment rates by 5 to 20 percent, cut employee costs by up to $5 million, and boost the conversion rate on service-to-sales calls by nearly 50 percent—all while improving customer satisfaction and employee engagement. While analytics is only one of a broader set of improvements, including operational changes such as coaching and process simplification, it is a powerful tool for companies to implement.

Where companies go wrong

Given these potential gains, why have more companies not taken advantage of the opportunities that advanced analytics offer? Simply put, many of them do not have the right foundation in place, due to entrenched organizational structures and processes, legacy IT systems, and other challenges. There are two root causes of slow advanced analytics adoption.

A lack of integrated data across channels: Many companies have call centers that function in silos. The centers generate plenty of data, but companies don’t have a systematic approach for aggregating that data into a single source of truth so that managers can make sense of it. Other companies buy a set of ad hoc solutions and assets to solve individual problems, rather than developing a strategic approach built around a single integrated platform. Further, the quality team, the workforce management team, and the digital team don’t talk to each other—and, in some cases, don’t even have access to the same data.

An inability to link analytical insights to actions: Other companies generate insights from analytics but don’t translate those insights into action. Or, they take some actions but fail to fully capitalize. Most organizations, for instance, run voice-of-the-customer analytics to calculate first-call resolution (FCR) and customer satisfaction metrics, but they don’t use that customer feedback to redesign processes or take other steps to make a more transformative impact. A common theme across these issues is that operations managers simply do not know what to do with analytics.

Traits of the analytics-driven contact center

There is a large and growing pool of available vendors and technologies for contact centers that are generating a lot of data but struggle to make sense of that data. It is important for contact centers to build the right foundation if they are to generate the maximum potential benefits. Those foundational elements include:

  • A clear vision and strategy: Contact-center organizations need a coherent, enterprise-wide vision for analytics. That vision must have a clear link to the overall business strategy, along with a road map for implementing specific use cases, such as improving FCR or offering more self-service options to reduce the demand on call centers.
  • An agile organization with internal analytics capabilities: Companies need to build strong in-house talent capabilities in analytics that align with the organization’s strategic goals. And, companies need agile mechanisms to capitalize on analytics-driven insights. For example, a leading credit card company has set up an interactive voice response (IVR) analytics lab that allows it to immediately assess changes in customer satisfaction and containment after every change in the IVR.
  • Platforms and data sources: Leading organizations also require a comprehensive data strategy and ecosystem that can support the broader analytics strategy. Platforms and data sources call for best-in-class data governance, data or IT architecture, and infrastructure and data security frameworks. Many top-performing contact centers have built data lakes as a single source of all data on customers, agents, product performance, surveys, and other sources.
  • An ecosystem of partners: Few companies can meet all of their data and analytics needs internally. Rather, they must determine which needs can be handled in-house and which should be outsourced to expert partners.
  • A culture of objective decision-making: Leading call-center organizations make their day-to-day decisions based on data, rather than gut instincts. Examples include analytically driven hiring, targeted coaching, performance-based bonuses, and other initiatives to improve outcomes.

Would you like to learn more about our Operations Practice?

Companies do not need to have all these elements in place before they begin applying advanced analytics. As with all fast-moving technologies, a better approach is to identify use cases with the existing data, develop them fully, pilot programs, and then iterate.

Four use cases

To see how advanced analytics can help companies proactively improve call-center performance across a range of dimensions, we identified four main use cases that employ advanced analytics.

1. Reduced average handle time
Text (including speech) comprises the highest proportion of unstructured data in most call-center operations, and therefore it offers the biggest potential impact. In fact, text is constantly being generated about a company, from social-media channels, chats with customer-service agents, surveys, feedback forms, warranty claims, and other sources. Making sense of this data requires scraping it from all available channels (including converting call-center recordings to text) and then cleaning it to remove unimportant words, punctuation, and special characters. Once the data is cleaned, companies can begin to generate meaningful insights from it (Exhibit 1).

Predicted chat intent uses different machine-learning algorithms to generate insights.

For example, a technology company used speech and text analytics to proactively reduce average handle time (AHT) by approximately 40 percent. The company analyzed a specific incident type by sifting through unstructured call description logs to find variabilities in the resolution process. Based on text data, the company mapped keywords from its logs for that incident category to better understand the implications on average handle time. Using that instance as a baseline, it created an automated self-learning solution that can use text analytics to identify potential AHT improvements in areas such as redesigning questions to better understand customer problems, optimizing processes, eliminating unnecessary steps, and standardizing agents’ resolution guides. From that foundation, the company can begin proactively identifying and mitigating other issues to improve customer service.

2. Reduced call volume
Advanced analytics allows companies to conduct full, end-to-end analytics on millions of customer data points, looking at text and call flow volumes to proactively identify potential improvements. Based on that output, the company can design a solution that offers an improved customer journey—prioritizing individual measures based on impact, required investment, and feasibility—and then use an agile approach to roll out minimum viable products to be tested and iterated via two-week sprint cycles. IVR rapid simulations can further accelerate testing and refinements, and an interactive dashboard can measure impact by call types and customer types.

For example, a financial services firm was experiencing a high number of repeat calls. For every 100 customer issues, the company received more than 160 calls. It used analytics to look at three specific factors: customers, agents, and processes. Among customers, the firm identified people who would call frequently for minor things, such as status updates on a resolved issue, or those who would call repeatedly if they were not happy with the initial outcome. If some customers do not receive a credit from one agent, for example, they might call back and try their luck with a different agent.

Similarly, the company ran analytics on call-center agents to segment those who had a low resolution rate or were spending too much time on minor issues, compared with top-performing agents. Finally, the company took an analytical look at processes to identify gaps or systemic issues (such as a replacement card that does not get delivered on time). Armed with this information, the firm identified an opportunity to reduce repeat calls by 15 percent.

3. Proactively enhanced network resilience
Organizations increasingly underestimate the challenge of network resilience when they think about workforce management—and the impact of outages on customer satisfaction. The typical company faces up to five major outages each year, and 25 to 30 site or queue disruptions. For a typical financial-services firm, a two-day outage can require up to a week to return to business as usual, with an increase in abandonment rates of 10 to 30 percent. Most workforce-management teams in contact centers don’t proactively model the outcomes of the outages on service levels.

Customer first: Personalizing the customer care journey

Customer First: Personalizing the Customer-Care Journey

By using workforce-management advanced analytics, however, companies can run simulations that predict the impact on service levels and estimated recovery time for different types of outages. Companies can also identify and prioritize a set of specific actions to reduce the recovery time, customized depending on which sites or queues are affected. As a result, customer-care leaders can understand the implications of events such as a call center completely shutting down, partial staff availability, a downed server, or other disruptions.

For example, a simulation could determine that a three-hour failure of one site in an organization’s contact-center network could cause the average speed of answer to increase from ten seconds to 350 seconds, with a recovery time of 17 hours. By modeling different interventions—such as adding capacity, rerouting customer calls, or making an IVR announcement, among others—and different start times and durations of each intervention, a company could reduce the recovery time from 17 hours to 8.5 hours (Exhibit 2).

Failing to respond immediately during emergency scenarios might lead to much higher wait times for customers and a longer recovery period.

4. Improved service-to-sales conversion
To truly improve performance, advanced analytics tools should not only increase efficiency and reduce costs but also proactively unlock new revenue. A virtual sales coach can accomplish that goal by assessing factors about a customer—not only existing data such as demographic and behavioral profiles and purchase history but also real-time data from a current service call—to predict the next product the customer is most likely to buy. It can then pull up a script to give the sales agent specific language designed to improve conversion rates for that customer (Exhibit 3).

A virtual sales coach can alert call-center agents of the next product that a customer is likely to buy and provide scripts to aid sales.

For example, companies using advanced analytics can analyze the text of previous successful sales calls for the same customer (regarding a different product). And they can proactively tailor specific aspects of a pitch based on the customer’s behavioral profile—for example, if a customer is identified as potentially regretting a purchase he or she made, the script can highlight a hassle-free return or cancellation policy.

A telco company used this approach to boost the conversion rate for sales initiated through a service call by 46 percent. The company identified a range of input variables and ran analytics to determine which variables had the biggest effect on a customer’s willingness to buy, broken down by specific product and service offerings in the company’s portfolio. The telco then developed specific sales scripts for each product.

How to get started

Building the right foundation is crucial if call-center organizations are to generate the biggest benefit from advanced analytics, but companies should not wait until all of those elements are in place. On the contrary, analytics is a rapidly changing field, and organizations must start applying advanced analytics tools and techniques right away and learn through experience.

To begin, companies can identify the potential value pools from an analytics initiative and prioritize them based on measures such as the payoff relative to required effort, data availability, customer demand, and competitors’ moves, among other things. Prioritizing will help a company focus on a specific use case—for example, improving FCR by 20 percent at a particular call-center site—and map the data requirements it will need, such as agent notes, voice-of-the-customer information, routing data, and automatic call distributor information. Most organizations will not have perfect data, but that should not be an excuse for a lack of action. Rather, organizations should begin by working with whatever they have and refine their data over time.

With the goal of using advanced analytics to improve FCR performance, a company can begin to generate hypotheses—for example, calls may be routed to the wrong queues—and then analyze the data to prove or disprove each hypothesis. Next, it can build an analytics model and test it with users, gauging results and refining the model based on user feedback (in close collaboration with IT and the line organization).

Finally—and most important—a company can scale up successful pilot tests across the entire call-center organization to maximize their potential impact. This rollout requires working with managers at other sites, applying lessons learned, and—when possible—automating analytics use cases to improve efficiency.

Companies often talk about identifying customer pain points, and call centers are a clear opportunity, yet most customers view them with dread. The best organizations recognize this as a chance to proactively differentiate themselves from the competition—but they cannot get there without advanced analytics. There is a real difference between first-generation data and analytics already in place at many companies and the advanced analytics techniques and methodologies now available. By implementing these new tools, companies can more accurately predict what’s coming—allowing them to literally control their own future.

Explore a career with us