TCIM Services Inc. runs outsourced call centers for Fortune 500 companies from facilities spread around the globe in a 24/7 environment. The Wilmington, Del.-based company’s operations are further complicated by the need to support multiple customer interaction points without downtime – and to make business intelligence (BI) data readily available to its clients.
“We have to continuously manage our internal business while also managing and providing information to our customers so that they know how their programs are being serviced and supported,” said Scott von Kleeck, TCIM’s chief information officer. He added that each customer has a different set of key statistics and performance measurements and that how often they want to see the information varies.
Many clients get daily reports, for example. But recently, some upped the ante and started asking for data on an hourly basis, or even more frequently. “Now it’s often the last half-hour that is of interest to customers,” von Kleeck said.
To meet those emerging requirements, TCIM has moved toward real-time BI capabilities – a shift that required the company’s data warehousing and BI infrastructure to “grow up,” in von Kleeck’s words. He said new analytics technology that TCIM implemented has enabled the company to easily integrate both cumulative and real-time data from a variety of data sources, including call center switches, predictive dialers, payroll systems and internal databases.
“In contrast with traditional BI approaches, we can add data sources without having to redesign the entire architecture,” von Kleeck said.
Self-service approach aids creation of real-time BI capabilities
It also helped that five years ago the company switched from an IT-driven data analysis process to a self-service BI infrastructure, which was designed to provide answers to analytical queries more quickly via a Web-based interface. Since then, the amount of data that TCIM needs to manage has continued to grow, while BI reporting intervals have gotten smaller. But the combination of the new data integration technology and the self-service infrastructure – which now is being extended to mobile devices – has provided a robust framework for enabling the real-time BI capabilities, von Kleeck said.
Claudia Imhoff, president of consulting firm Intelligent Solutions Inc. in Boulder, Colo., said that when organizations are considering real-time business intelligence projects, it’s critical to recognize the implications. With both real-time BI and its operational BI cousin, analytical processes are embedded into operational workflows and systems. That means a real-time BI system can't go down – it requires hot backups and full disaster-recovery capabilities, things that typically aren’t found in a traditional BI environment, Imhoff said.
“Scalability is mandatory,” she added, because real-time BI processes require large volumes of data – often much greater than what companies are used to handling in their BI systems. Similarly, the performance of real-time BI applications must mimic the performance found in the operational environment itself – that is, sub-second response rates in most cases.
Imhoff said that standard operating procedures likely will need to be rewritten and personnel will have to be retrained to use a real-time BI system. She suggested that flagging mechanisms or alarms be built into a system to detect if employees are bypassing the BI steps and returning to "the old way of doing business." On the other hand, she noted that real-time BI tools must be easier to use than conventional ones are because the audience for them tends to be broader and less technical than the user base that many BI managers are used to dealing with.
Some of those considerations were on the mind of Mitchell Swindell, chief technology officer at Fort Worth, Texas-based Novo 1, when he was evaluating real-time BI software. Novo 1, another company that manages contact center and customer service operations for corporate clients, started a real-time BI project last year, in part because one customer asked for a dashboard application that could track key performance indicators for its customer service program.
Swindell said he wanted to be able to pull together data from a variety of source systems so he could create customized dashboards showing up-to-date information on service performance and workforce statistics, such as worker productivity and absenteeism, “all the way down to the agent” level. But, he added, “I didn’t want to do a lot of heavy lifting.” The technology that Swindell eventually selected made it easy to integrate information feeds and display them within the dashboards, he said.
In addition, Novo 1 purchased the real-time software under a pricing model that was based not on the number of users but on an initial license fee with incremental charges for deploying the software on additional systems. As a result, a product "that was originally installed for 80 users could be rolled out to 1,500 with no additional cost,” Swindell said.
Real-time BI aims to reduce pain at the pump
U.S. Xpress Enterprises Inc., a trucking company based in Chattanooga, Tenn., began using a real-time BI system last summer to collect data from on-board systems in trucks and make the information available to executives and other users. For example, fleet managers trying to reduce idle costs – the price of fuel used while trucks sit running in parking lots – get color-coded alerts in a dashboard when drivers exceed normal idling thresholds.
The data, which is updated every 13 minutes, was there for the taking, said Timothy Leonard, CTO and vice president of information management at U.S. Xpress. But creating the real-time BI capabilities posed some challenges. Speaking at The Data Warehousing Institute’s BI Executive Summit 2011 in Las Vegas, Leonard noted that the truck systems capture more than 900 different data elements, which needed to be integrated and organized in an effective way.
U.S. Xpress also had to establish new processes to drive behavioral changes among its drivers, including incentives to encourage good performance on idling and “some negative consequences” for heavy idlers, Leonard said. “It may sound all harsh initially, but what that did was set up accountability,” he added. “Really, the key is the driver, not the truck.”
From the end-user perspective, the real-time BI technology lets the company’s fleet managers get information without having to request a report or wait for a centralized data warehouse to be updated. And now, the fleet managers would like to see the data “even faster,” Leonard said. “If we could get it to them in five minutes, they want that.”
Alan R. Earls is a Boston-area freelance writer who focuses on business and technology.