by Stephanie Overby

Big Data Analytics Gold for the Call Center

May 11, 20125 mins
CRM SystemsData ManagementIT Leadership

There's valuable data in those contact center logs that have largely been gathering dust, and companies are using new tools to mine them to boost customer satisfaction and revenue and lower costs.

There may be no corporate function that throws off more data than the corporate call center. “Every contact is counted, routed, measured and scored. Agent performance is actively measured,” says Tony Filippone, executive vice president of research for sourcing analyst firm HfS Research. “Other key process owners, like finance and accounting or claims adjudication, wish their data was as rich.”

Throughout the history of the contact center, much of the analysis of that data has been quantitative in nature—calls received, average hold time, call length, resolution rate.

“Over time, companies added more sophisticated workforce management tools including global scheduling information—to help with network call handling, scheduling, real-time adherence—but the data collected was agent performance- and efficiency-related,” says John Magliocca, principal consultant for contact center service at outsourcing and management consultancy ISG.

That’s beginning to change. Corporate call centers—and call center providers—are embracing new analytic tools to dig deeper into the big data they generate. There are a number of business factors driving the change, says Filippone. Contact center agents are being asked to handle a wider breadth of issues from more channels, including contacts from social media and online forums.

“This requires more advanced skills, better training and great real-time guidance,” Filippone says. At the same time, companies that have been disappointed with the quality of offshored call centers are now looking to emerging end-to-end CRM tools to take costs out of their operations instead.

Analyzing Unstructured Data

Some of the most interesting software being implemented today attempts to take unstructured voice recordings and analyze them for content and sentiment. “There have been efforts underway to put contact data to work to best understand the current mood of the customer and other information that can immediately mold client strategy and direction [for some time],” says Magilocca. “But until recently the systems were not robust and the information wasn’t useful.” These tools could enable the call center professional to understand what’s really bugging his next contact before he even says hello.

“Companies are applying text and sentiment analysis to this unstructured data, and looking for patterns and trends,” says Deepek Advani, vice president of predictive analytics for IBM, which has implemented its Text Analysis and Knowledge Mining (TAKMI) tool at it sown call centers to analyze call center agent records, identify customer concerns, highlight trends and patterns, and provides early warning capabilities.

“Many companies are integrating this call center data with their transactional data warehouse to reduce customer churn, and drive up-sell and cross-sell [activity],” Advani says. “Call center logs can provide invaluable insight on what customers were calling about, and can also provide insights for future product requirements.”

Most contact center operators—traditional vendors like TeleTech, Teleperfomance and Aegis, as well as newer cloud-based providers like RightNow and Kana—are implanting emerging call center analytics that bring together real-time and historical data—to isolate revenue-related calls, identify agent best practices, predict root causes of their dissatisfaction, or identify what characteristics of a contact lead to costly repeat calls, says Magilocca.

Machine learning could revolutionize the call center, as well. Everyone who’s ever called a help line has a horror story of being stuck in some seventh level of Interactive Voice Response (IVR) hell. “Being able to respond to customer requests in an automated manner, but without disagreeable and confusing IVRs, could be a huge boon for companies,” says Filippone. “Products like IBM’s Watson could one day replace traditional linear-thinking IVRs.”

ISG’s Magilocca is also keeping an eye on e-learning systems that can proactively determine a call center employee’s biggest professional hang-ups by analyzing her responses to customer inquiries and suggests training modules to improve performance.

Not all the new applications are big-ticket items. Low-cost social media listening tools are enabling contact center managers to actively search words and phrases on Twitter to identify brewing customer complaints or global issue

Change management will be a big issue in the integration of the new systems and processes. The more sophisticated analysis must actually deliver more than the more rudimentary analyses of the recent past. “Contact centers already have rich data and basic analytical capabilities. New solutions have to be proven better, otherwise the advanced analytics will be dismissed as too ‘fuzzy,’ says Filippone.

Companies will also have to figure how to effectively use the information they generate. “The analytics have to drive action, not just insight,” Filippone says. “It isn’t enough to count issues and score sentiment; new solutions have drive agent behavior change or transformations of CRM strategies.”

If there’s no hard return in short order, investment in new systems could be short lived. The customer contact industry remains hyper focused on the bottom line. “Buyers have to be confident that analytics will take out costs or identify new sales opportunities,” says Filippone.

Stephanie Overby is regular contributor to’s IT Outsourcing section.

Follow everything from on Twitter @CIOonline, on Facebook, and on Google +.