May 20, 2020
How to use predictive data science to plan for utility operations post-COVID-19
COVID-19 has affected every major industry around the world. For example, utilities have stopped shutoffs for nonpayment at least through May 31, if not longer. No doubt a necessary act, but it will cause complications down the road: a significant number of utility customers will be overdue on their bills in the next 6 to 24 months.
Here’s what utilities can expect to see in that time:
- More accounts facing arrears than usual
- An increase in the cost to serve customers, stemming from increased calls to the contact center and the cost to disconnect service when the moratorium on shutoffs ends
One thing is clear: Utilities need to start planning for arrears now. And they need to put mitigation strategies in place to deal with increased costs.
We tackled this topic in an April 2020 web conference where we discussed the importance of making data-driven decisions and discussed how utilities could use predictive data science to prepare for the next 6 to 24 months. With the acquisition of TROVE Predictive Data Science earlier this year, E Source gained unmatched expertise from the TROVE team in predictive data science and artificial intelligence (AI) solutions for the utility industry. We’re thrilled to share these solutions with our E Source network.
The “return to normal” will be anything but normal
During the web conference, we welcomed Chuck Caisley, senior vice president of marketing and public affairs and chief customer officer at Evergy Inc. Caisley explained how Evergy is looking beyond the current moratoriums and seeing potential risks for its customers in the coming months. Caisley stressed that the “return to normal” will be anything but normal, and acknowledged that the historic need to rely on disconnections for cash flow runs contrary to good customer service.
Evergy is thinking creatively on ways to mitigate operations and maintenance costs. The utility is using data science and working with TROVE to analyze its residential and small commercial customers—the people who would normally call the utility contact center. Evergy hopes to achieve two things:
- To create new (potentially temporary) programs to offer to customers coming out of the moratorium.
- To be more proactive in reaching out to customers to get them on a payment plan and avoiding a wave of calls when disconnections resume.
But to be effective, Evergy needs to target the right customers: those who are most at risk of missing payments and entering arrears. This is where TROVE can help.
How TROVE uses predictive data science to identify customers at risk of missing payments
TROVE analyzes hundreds of customer-level data points around financials, demographics, behavior, etc., and combines them with utility data on customer billing history, contact center records, and payment records. Plugging these combined datasets into TROVE’s AI and machine-learning models generates a propensity to pay score for each account.