Social networking, analytics are boon to financial industry

Financial services firms are employing knowledge management solutions to break down silos within their organizations, provide better analytics of customer feedback and improve profitability forecasting.

Social networking enhances internal KM

TD Bank Group, headquartered in Toronto, started employing social media for some of its internal communications in 2008. The financial institution distributed basic news and announcements, and enabled employees to comment on different items.

That limited rollout produced unanticipated results. Comments from internal social media users helped TD Bank identify opportunities more quickly and resolve issues more efficiently, according to Wendy Arnott, TD VP of social media and digital communications. With traditional group and committee meetings, concerns would typically take longer to rise from the level of someone noticing an issue, to the stage at which someone with authority could take action.

To expand its social media capabilities further, TD Bank Group undertook an extensive vendor selection process at the beginning of 2010. Bank officials wanted the application to: handle the social media communications the company already had, as well as scale for growth; provide a user-friendly interface; manage the technical issues of operating within a financial services environment; and work both in the United States and Canada.

More than a million connections

TD Bank looked at several different solutions, choosing IBM Connections, which met its needs better than any other applications, Arnott says. After choosing the solution at the end of 2010, the bank and vendor worked together to ensure that all aspects of the application were thoroughly tested for security, reliability and functionality. TD Bank launched the solution enterprisewide in November 2011.

Within the first six months, more than 85,000 users were participating in IBM Connections-driven social networking activity, making more than 1 million connections and forming 3,000 different communities, as well as countless blogs and wikis.

By forming separate, smaller communities, TD employees and executives have tightly focused groups, enabling solutions and discussions to be on point, according to Arnott. Queries and suggestions are more likely to lead to solutions than in a larger group, and there is less unrelated information through which to sift.

TD Bank Group is evaluating the newest version of IBM's application that offers additional social networking features.

Enhanced customer feedback analytics

Fidelity Investments, a large, diversified financial services company headquartered in Boston, was receiving hundreds of thousands of pieces of good customer feedback annually from surveys, social media and contact center feedback. The content included invaluable suggestions for improvements and enhancements, as well as customer sentiment information (e.g., agree, strongly agree, disagree, strongly disagree) about particular products and services. But the sheer volume of the structured and unstructured data was too cumbersome to conduct any meaningful analysis, says Parrish Arturi, Fidelity senior VP, customer experience.

"We wanted to be able to harness the voice of the customer and the voice of the associate," Arturi explains. So in summer 2010, Fidelity, based on information from its technology research arm, looked at a handful of different analytics providers to see which could best meet its needs. After narrowing the search to two firms, pilot tests were conducted using thousands of pieces of actual customer feedback to produce analytics reports.

From the pilot, the company chose the sentiment analytics platform from Clarabridge, based on a combination of the vendor's people, process and technology, Arturi says. Fidelity initially made the decision in the second quarter of 2011, and then spent a few months outlining priorities for rolling out the application, which Fidelity started using at the beginning of 2012.

"It has helped us to prioritize improvements in different categories and different services offerings," Arturi says. Additionally, the analysts who had been manually categorizing all the unstructured data shifted from data entry/data categorization duties to true analysis work.

Arturi expects to expand the use of the Clarabridge application to Fidelity's business-to-business unit, and then to enhance the application with Clarabridge's collaboration capabilities for better teamwork within the company.

Improved profitability forecasting

PREMIER Bankcard of Sioux Falls, S.D., like other credit card issuers, saw reduced revenue potential with the enactment of the CARD Act in 2009. The new law limited some of the fees that credit card issues had placed on cardholders.

PREMIER had relied on high fees to protect itself from the increased risk it took compared to typical card issuers. While most card issuers build their business around consumers with good credit who were likely to pay their credit card bills, PREMIER caters to people with poor credit histories. The potential for default is greater and credit lines are lower, so potential earnings from card usage are lower.

PREMIER had built its business around charging high upfront fees for card usage. The issuer had to be much more exacting about whom it extended credit to, the size of the credit line and other factors to continue to be profitable, explains Rex Pruitt, manager of MIS, profitability and risk requirements.

"Our model is different; we provide credit to individuals who have been through hard times," Pruitt says. "We work with people who have damaged credit scores and who are trying to re-establish their credit."

More time to analyze

The issuer had been using a predictive modeling tool, SAS Enterprise Miner, since 2003. But the new law required that PREMIER be more exacting and look longer term at predicting a customer's expected profitability. Pruitt explains, "We needed to be able to automate the process, not just enter macro codes."

So in summer 2011, the issuer started looking for a solution that would provide more detailed forecasting capability. SAS again had what PREMIER officials considered to be the best solution, SAS Forecast Server. The application provides forecasts from a GUI, using variables supplied to the system in the modeling process.

PREMIER selected the SAS solution in October, started installation in November and was on track to go live with the application in mid-June of 2012, according to Pruitt. When the application is fully operational, Pruitt expects PREMIER analysts to be able to spend more time with actual analysis rather than simply writing and adjusting code to develop forecasts.

"We wanted our analysts, who were spending 80 percent of their time writing macro code and only 20 percent of their time in actual analysis to be able to change that around to only 20 percent of their time writing code and 80 percent in actual analysis," Pruitt says.