May 1, 2013 • 82 Views • Start a discussion •
Attending CS Week in rainy Tampa, Florida (more on what I learned here next week) inspired me to pull out a customer analytics story about credit and collections from our Utility Analytics Institute project profile archives.
If you have a great story about the analytics work you've done at your utility -- or know a client who has done some great work -- we'd love to hear from you. Interested in accessing these stories? Just let us know, and we'll be happy to get you more information about the database, or you can go to member.utilityanalytics.com.
Also, if you can't get enough of stories like the one below, I'd encourage you to participate in our customer marketing and segmentation working group with the Utility Analytics Institute. This group will give you the opportunity to hear stories from other utilities about their customer analytics efforts, as well as engage in real-time with the group through phone calls and online collaboration. If you'd like more information about this group, please don't hesitate to contact me.
Alrighty then, enough of my digressions, let's get on to the good stuff.
A large investor-owned utility has developed a predictive model for one of its operating companies. This model is used to prioritize residential customers for collections efforts post-disconnect for nonpayment. Customers are scored based upon their expected loss which combines their likelihood of paying and their outstanding debt. Customers new to the post-disconnect collections process are scored once per week and all customers within that group are scored monthly.
The utility has limited collections resources and does not use a collections agency. Therefore, this is the last opportunity to recover this debt. These predictive model-based scores enable the company to better focus collections resources and spend more time on those customers with the highest expected loss.
The tools and technology necessary to support this initiative were already in place and no external resources were involved. The internal resources required to develop this model totaled approximately five to six person-months.
There were two primary drivers behind this project:
- Maximize revenue recovery from post-charge-off collection efforts
- Improve operational efficiency
A statistical program, running on individual personal computers, was used to build the predictive model. An image of the utility's customer information system (CIS) containing customer and billing system data is secured via a query, though it is not a full replica of the production data. This avoids potential impacts related to running analytics against the production CIS.
Data from the CIS and a separate charge-off database were used to build the model and to produce customer collection prioritization scores. New customers to the process are scored weekly, and the entire population of post write-off accounts is scored monthly.
The data sources involved are fairly straight forward and include customer account information. (e.g. customer name, address, contact information, age of debt, payment history, customer attributes, an internal credit score based on a point system, external credit scores obtained when the customer first sought to initiate service) All of this data originates from the CIS or the database containing information about charge-offs.
The predictive model scores each customer and that information is made available to the manager of collections who in turn shares the file with those actively engaged in collecting post charge-off debt, and the files are retained for future use.
The business process associated with collecting post charge-off debt did not change significantly. However, the prioritization process has been automated and is now based on predictive model-based scores that indicate the customer's propensity to pay versus a more dated model.
This project facilitated interactions and collaboration between the marketing services team, CIS experts in the IT organization, and subject matter experts in the collections group.
Measuring the value of the new model with precision can be problematic because it requires that all other things in the process be held constant, which is never the case. The utility is confident that there will be an improvement in total collections of several percentage points which will be many times the cost of developing the model.
One of the utility's key objectives is to provide high reliability at affordable prices with the goal of having very satisfied customers. Improving efficiency in the collections area is one way to reduce their cost of doing business, benefits that will ultimately be passed on to their customers through lower prices.
The utility hasn't adopted any strategic analytics-centric metrics or goals, but analytics-driven information is a fundamental input to helping them improve their understanding of their business and their customers.
The title of the technology lead for this project was senior data analyst. The senior data analyst fulfilled the roles of technology lead and business-side project manager. The title of the executive sponsor for this initiative was vice president of customer service. External providers were not involved in this project.
Several analysts with expertise in statistical modeling, data extraction and transformation, data sources and subject knowledge were all a part of this effort.
This initiative did not create any significant change management needs. In general, individual project teams care for change management (e.g. training, communications, process change documentation) as there is not a standard change management framework per se.
So that's a taste from the customer analytics side of our utility analytics project database. If you'd like to learn more about the database and how to get involved, please don't hesitate to reach out to me at email@example.com. If you're a member of the Institute, you may already be able to access the full database at member.utilityanalytics.com.
Thanks for reading!
H. Christine Richards is the director of knowledge services for the Utility Analytics Institute, a division of Energy Central.