At Utility Analytics Institute's September 18-20, 2012, Utility Analytics Week in Arlington, Texas, two companies provided use cases to describe how they apply data analytics to solve challenges facing the customer side of their businesses.
When Greg Ravn, six sigma master black belt in process improvement at Jacksonville Electric Authority (JEA), was called into the utility's CEO office, he knew a hot-button issue had come up.
Through phone calls and media attention, JEA executives had become aware that electric usage among certain low-income housing residents was "peaking" at an astronomical eight times the median usage level during winter months.
On a per square foot basis, elderly residents on fixed incomes and other low-income customers were paying hundreds of dollars more per month during winter months for electricity than comparable-sized locations. The question was, why?
"Some of our customers are being left behind," the CEO told Ravn, who was asked to investigate the issue. After running a slew of energy consumption, property appraisal, demographic, and GIS reports on the locations in question, Ravn said, "nothing statistical jumped out at us." They were stumped as to why the exorbitant peak usage patterns existed.
JEA decided to conduct on-site appraisals of the locations. Call center agents set up 200 home appointments with customers. Technicians dispatched to the residences found each location had either an inadequate amount of insulation or none at all in attics and walls.
JEA then queried customer energy consumption data with specific metrics to identify homes with high single season winter usage peaks. A list was handed off to the utility's demand side management, which dispatched energy auditors to verify conditions at targeted sites. With confirmation of the insulation deficiency, the utility arranged funding and grants for low-income insulation upgrades.
The result, according to Ravn, was a "significant effect on peaking operations" which "played a role in the deferment of additional capital spending (power generation)" by the utility.
Low-income customers who received new insulation reduced their usage by an average of 1,000 kWh per winter month, with a total estimated savings of $4 million per winter season.
Using a two-pronged approach of legwork and analytics, JEA helped their low-income customers save money and enabled the utility to avert what threatened to be a significant public relations controversy.
CenterPoint Energy call center reengineered around data analytics
At CenterPoint Energy, Chad Andree, manager of workforce management and performance management reporting, was asked to reengineer the company's high-volume call center.
Handling between six and 12 million calls per year from the utility's five million metered customers, the call center sent to the business division a maze of manually generated, difficult-to-analyze performance reports, creating a headache for managers. Data resided in a variety of islanded databases, information formats and terminology.
"Management couldn't trust the accuracy and consistency of the information being generated by call center reports," according to Andree.
In his effort to gain the utility's executive and business partner buy-in for his overhaul strategy, Andree said he had to become an "ambassador for analytics" within the organization. Management consistently wanted to know "how are you going to measure it?" Andree stated he had to take raw call center data and "turn it into a story that made sense to management."
Key to the success of the data integration and automation project was creating standardized key performance indicators and metric names so that everyone in the organization was for the first time speaking the same language.
Measures included the number of calls, type of call, and average call handling time. Previously this data was manually generated. With automation, the data was now accurate, timely and accessible. This information is vital to forecasting call load, staff scheduling and generating performance management reports.
Using graphical models of call center metrics, Andree presented a case to managers that conveyed a clear message: "Here is what we need and here's what it will get you." Andree stated he became a "consultant to his company's business and executive management teams."
The more managers trusted the data, the more enthusiastic their buy-in to proposed changes. Once performance targets were set top-down by the executive team, gaining buy-in from call center agents and managers naturally followed.
Eliza Castillo, business analyst with CenterPoint Energy, added that the call center analytics-based reengineering project enabled the utility to save hundreds of thousands of dollars in overtime staffing costs and other benefits.
Michael G. Albrecht is a student at Austin Community College in the Technical Communications program.