by Mark L. Black and Jeremy Oosthuizen, Origin Inc.
Utility executives continue to wrestle through legions of data for answers to the question: How do I redirect my organization to increase revenue, reduce costs and grow margins?
Some have embarked on a marathon run—others in a mad dash—to implement reporting systems that might, or in many cases might not, produce tangible results.
A few have ventured into the world of business intelligence, the proposed nirvana of corporate information and reporting.
Business intelligence proponents say it can help improve the bottom line by finding lost revenue, improving collections and increasing efficiency.
But can it?
As utilities transform to keep up with new technologies, smart grid pressures and advancing interests in energy alternatives, appetites for more meaningful and accurate information are on the rise.
These pressures and challenges prompt questions such as: Can utilities become predictive? How can data be used to greater effect and not just become more data? Does a business intelligence approach deliver on its promise?
Most utilities have reams of transactional data reports: lists 20 columns wide on stacks of paper up to your knees identifying customers who have not paid their bills, showing how long the payments are outstanding, and a cacophony of other details.
These reports are necessary and useful, but they are limited in their application.
As technology advances, the proliferation of reporting tools is inevitable, too, providing new ways to access the same information.
Adding to the confusion is that no standard reporting tool fully satisfies every need for every utility. And not every utility has use for all the reports that are generated.
The situation is exacerbated by the exponential growth in the volume of data utilities are required to track. A Tier 1 utility’s data volume would increase from about 300 terabytes to almost 700 terabytes as soon as advanced metering infrastructure (AMI) is deployed, according to the Electric Power Research Institute (see Figure 1).
Unless all of this data is analyzed properly, it will lead to more data overload and less useful information.
Knowing what each customer consumes and pays or does not pay is one thing; knowing the trends and using them in a productive, more lucrative way is another.
Historically, utilities think in arrears, generating transaction reports that tell an audience what happened.
Utilities are better equipped to relay, “This is where we are because we have the data to prove it.” What if, instead, new behavior were driven by data analysis and trend prediction?
A data warehouse, coupled with a business intelligence system, combine to produce reports that focus on the action that must occur rather than transactional data that shows what has occurred.
When data is placed in a properly designed warehouse, the data changes to a format designed for analytical processing. Data elements are sorted naturally according to what logically must be grouped together, averaged or arranged to identify trends.
Viewpoints or dimensions from which that data will be viewed, sliced, diced and analyzed are conformed across the groups of data.
This arrangement makes it much easier for an analyst to know where to find data and how to use it.
Data warehouses have a striking advantage compared with transactional systems because they can sift through immense amounts of data to identify patterns.
A data warehouse produces a strategic piece of information in minutes or seconds; a similar query would bring a transactional database to its knees.
For example, consider if an analyst tried to compare billed consumption, revenue and arrears month over month for the past year for hundreds of thousands or even millions of accounts.
This kind of query would be difficult for a transactional processing database (i.e., an operational system) but is made easy by data warehousing.
Consider another example with two utility customers: Joe and Jean.
Joe pays his bill on time every month, but this past month he has not responded.
Jean frequently is late with her payment, missed a payment twice in the past 12 months, and has not paid this month.
Both look the same in a monthly transactional data report. Joe might be out of town. Jean also might be out of town, but other factors likely are at play.
Jean has had four addresses in three years. Joe has lived at his address 16 years. Transactional reports don’t show that.
The approach to each might be the same, but it also might be different depending on the data available.
The ongoing relationship between utilities and customers can be affected significantly based on the approach taken.
Usually the only communication between customers and utilities occurs in sending bills and receiving payments unless there is a problem.
Sometimes customer service representatives can help customers during their first phone calls, but sometimes the process take weeks or months and involves many departments.
Analysis such as that which business intelligence provides can be used to shorten cases that take longer than the first phone call to solve.
Retailers have been using this kind of business intelligence analysis for years, and the applications are endless.
Retailer A knows, for example, that 70 percent of its stores in one area of the country for the past week have had increased sales in flu and cold medicines, tea bags, chicken soup and tissues.
Retailer A also knows that chicken soup and tea bags sales have been at an all-time low in the adjacent area, where temperatures have averaged more than 100 degrees for 22 of the past 30 days.
Prevailing weather, however, might be going in that direction, and, thanks to knowledge of what sold in the first area, the flu is headed that way, too.
Inventory shipments to regions likely will be adjusted, and positioning of certain products in individual stores might change.
That is intelligent use of business information.
Designing a data warehouse that responds effectively to queries from a business intelligence system is not a straightforward task.
It takes special skills and a complete shift in design thinking; it requires an architect who can step away from the normal design of a transactional database.
For example, data warehouses contain redundant data—a difficult concept for a transactional data analyst, but an essential ingredient for successful data warehousing.
Many vendors are creating sophisticated products for the utilities market, specifically for their vertical products, that can work with this challenge.
All products, however, are not created equal. Compatibility with existing systems is essential, as is continuing vendor support.
Let’s return to Joe and Jean.
Customer care and billing units at utilities already know which customers have outstanding balances, how long they have been outstanding, and which collection processes have been implemented to try to collect the revenue.
But an integrated business intelligence system implemented with proper data warehousing knowledge can look at that data across any set of viewpoints: type of customer, commodity, collection process, geography, all of the above, and many more.
It then can be used to determine the effectiveness of each process across any of those attributes.
At this point, it is all still just information. It becomes intelligence when the new knowledge is used to funnel groups of nonpaying customers into the most effective processes for their peers, resulting in increased collections.
It can be used to tweak individual steps in processes, further shortening the time to get the debt paid. When this newfound intelligence is used to uncover new questions to ask, it becomes wisdom (see Figure 2).
Business intelligence is a relatively new concept in the utilities industry; collecting data is not.
The value is in integrating the two within existing systems to yield maximum benefit for utility companies, collecting revenue that is not being collected or is being collected inefficiently and providing better customer relations and response.
Utilities likely will have multiple avenues to follow in applying business intelligence principles, some to much greater effect than others.
Finding the right partner for that learning path is essential to achieving success.
Experience from other industries and early indications in the utility industry are that utilities can use business intelligence to cut through the transactional data noise to get to the knowledge nirvana.
Mark L. Black is president and CEO of Origin Inc., a consultant in the implementation and support of revenue management and billing solutions for utility and energy industries, taxing authorities and financial services. Reach him at firstname.lastname@example.org.
Jeremy Oosthuizen is a principal at Origin Inc. with expertise in implementing customer care and billing and business intelligence for utility products. Reach him at email@example.com.