More than a simple optimization exercise, rate design requires a complex balancing act to cover the costs to reliably provide service, charge a fair and transparent amount to all customers, and not overburdening any specific customer.
When analytics are applied to Advanced Metering Infrastructure (AMI) data, the resulting intelligence serves as an essential input to rate design calculations, delivering more effective rate structures that better satisfy utility objectives.
New Hampshire Electric Cooperative’s (NHEC)’s experience developing AMI analytics-based rate design provides an example of the power and potential of this approach.
Incentives vs. Rewards
While many utility rates are designed to encourage customers to modify their behavior by shifting energy use to lower cost hours (i.e. time of use rates), NHEC approached Bidgely with a different objective. NHEC sought to offer existing segments of its member base (NHEC’s customers are ‘members’ of the cooperative) lower overall rates because their current time of use consumption was less expensive to serve. In other words, NHEC sought to proactively reward members who demonstrated positive power use behaviors rather than incent those who demonstrated negative behaviors to use energy differently.
To achieve NHEC’s objective, Bidgely employed an innovative data science technique known as unsupervised learning.
Starting with our patented AI-powered load disaggregation solution to itemize 100 percent of residential member loads, Bidgely was able to provide NHEC with visibility into member’s usage patterns at both whole-home and appliance levels. This advanced data science enabled NHEC to explore opportunities to design rates that better aligned with member usage patterns.
Bidgely’s Analytics Workbench business intelligence platform analyzes AMI meter data and supplemental utility-contributed member data to generate up-to-date analyses of major appliance ownership, usage in kWh (including time of use) and additional attributes such as major appliance power draw in kW.
Using Analytics Workbench, Bidgely expanded upon disaggregation to deploy unsupervised learning algorithms to identify clusters of members who fell into patterns of electricity use by annual consumption, seasonal consumption, and even daily consumption.
For example, the advanced data science identified members who had high summer loads or high winter loads. It also identified members who fell into certain daily patterns, such as early birds, afternoon peakers and night owls.
In total, the unsupervised learning clustering analysis identified 14 unique lifestyle-specific clusters among NHEC’s membership.
Cost to Serve
We then worked with NHEC to ingest their cost-to-serve data into the platform, which included cost-per-hour data for the course of an entire year. With this data, Analytics Workbench was able to determine the cost to serve each member cluster at a specific time.
Next, NHEC looked at whether they could offer one or more of the low-cost-to-serve member clusters a discount on their rate without unfairly affecting the remainder of their service population. Bidgely’s analysis revealed that NHEC could offer four member clusters a lower rate without adversely impacting other members.
Applying NHEC’s Learnings in a Regulated Environment
New Hampshire Electric Cooperative is a cooperative, but there are still lessons to be learned for traditional IOUs. The same principles of rate design can apply as regulated utilities work toward approval of a rate case.
First, it’s essential to understand the unique segments that exist within a service population, including their time of use and usage patterns. It is also important to take into consideration the cost to serve each individual customer. No matter the business model, leveraging AMI data empowers utilities to design innovative and modern rate structures.
Building Customer Relationships
Rate design has always been the foundation of the relationship between utility and customer. However, before AMI data became available, the number of rate structures was limited. Now with the benefit of AMI-derived customer intelligence, it is possible to design a wide-range of rate structures to better fit diverse customer segments. As a result, customers are better able to understand the value they receive from their energy provider, and utilities are better able to achieve their business objectives.