COVID-19 has created drastic challenges in utility revenue and collections forecasting:

• C&I load profiles have changed dramatically, with non-essential businesses using much less electricity while many essential businesses have increased usage.
• With the incredible rise in work-from-home, residential load has increased – not enough, however, to offset the decline in C&I load.
• Making matters worse is the unprecedented rise in unemployment, which is predicted to reach 30% and, undoubtedly, affecting customers’ ability to pay.

Because of the medium and long-term consistency of electricity demand and the vital importance of the service they provide, utilities have traditionally used very long-term forecasts to understand how best to work with regulators, customers, and other stakeholders to set prices, rate structures, and collection policies that fairly balance the needs of shareholders & lenders and customers.

What is so unique about the current environment is the incredibly rapid pace of change and the requirement for utilities and their stakeholders to deal with much shorter time horizons and more real-time data flows.  Utilities with smart meters collecting interval data will be much better positioned to analyze the current situation and make informed recommendations and decisions.

To accurately forecast revenue and collections, utilities will need to:

• Measure what has happened in just the last 6 to 8 weeks and be able to compare it to previous periods and most recent expectations.
• Develop a range of load forecasts based on best-case and worst-case scenarios for the projected end of stay-at-home orders.
• Reclassify C&I customers as essential or non-essential businesses to accurately forecast load and ability to pay.
• Modify residential historical load forecast assumptions based on shelter-in-place;
• Update customer payment models based on best-case and worst-case unemployment projections.

Supporting these requirements will present several challenges to utilities. Most utilities rely on a sampling approach for load research and rate making.  While utilities have relied on the long-term stability of demand to establish that sampling can accurately forecast a customer group’s behavior, that sampling does not account for the COVID-19 Black Swan.

Also, as mentioned previously, existing utility industry forecasting technology was not designed for rapid adjustments based on new data, data models and assumptions.  Developing new forecasts is typically a multi-month or multi-week process, which is not sufficient in times of crisis.

Finally, and related to the previous point, utilities typically have multiple planning applications – one application for load forecasting, one for revenue forecasting based on the load forecast, and another for projecting cash collections based on revenue generated.

Current forecasting technologies may not be able to quickly adapt to these new realities.

Whether a C&I customer is an essential or non-essential business is now a leading indicator for load forecasting; a criterion not included in current models.  Without knowing the full extent of COVID-19 load implications, individual customer load profiles should be used as data for new post-COVID-19 forecasts. Existing utility technology is not capable of using each customer’s load profile. Additionally, COVID-19 data is changing daily. Forecast models, assumptions and data need to be updated daily with new forecast results. To quickly produce updated forecasts based on the latest information, utilities would benefit from a single application capable of developing load forecasts, revenue forecasts, and collections forecasts.

Utilities nolonger need to rely on sampling.  Formerly unknown or undiscovered COVID-19 implications that would be masked by sampling will be included in new load forecasts. Whole population bill calculations can produce more accurate revenue projections. With something like an Enterprise Rating Engine, every customer bill line item is calculated with revenue-grade accuracy for the revenue forecast. This enables the utility to segment forecasted revenue based on customer class, tariff, rider, and tax, or any combination. While the effects of the crisis are projected to last for years, utilities must find a solution for this new reality.

The timeline for companies to react to the coronavirus has shrunk dramatically. We are in unprecedented times. All previous planning assumptions need to be questioned.  And new information is becoming available daily.  Being able to quickly update and process forecasting models for load, revenue, and collections can be a strategic differentiator for utilities.