Methodology

Demand Charges

Commercial and institutional electricity bills typically consist of two components: energy charges (cents per kWh consumed) and demand charges (dollars per kW of peak demand recorded during the billing period). Energy charges scale with how much you use; demand charges scale with how fast you use it at your single worst moment of the month.

The National Renewable Energy Laboratory surveyed more than 10,000 commercial and industrial utility tariffs across 48 states and found that demand charges commonly comprise 30–70% of a commercial customer's total electric bill (NREL, 2017). The tariff-by-tariff demand charge rates behind that survey are published as an open dataset by the U.S. Department of Energy.

Where your institution falls in that range depends on your specific rate schedule and load profile. Utilities publish their tariff sheets — see, for example, Puget Sound Energy's electric tariff library. The fastest check is your own bill: the line items priced per kW are demand charges.

Reduction Potential

Our 10–15% reduction figure for billed peak demand sits inside the range published field research reports for demand management in commercial and institutional buildings:

• Lawrence Berkeley National Laboratory field tests of demand shifting with building thermal mass demonstrated peak-load flexibility of roughly 10–25% over multi-hour peak windows in large commercial buildings.
• In an LBNL pre-cooling field study (LBNL-55800), an office building shed 80–100% of chiller demand during the 2–5 pm peak window with no occupant comfort complaints — pre-cooling is one of the highest-leverage no-capital strategies Intervalwise recommends.
• American Council for an Energy-Efficient Economy analysis finds well-run demand response programs reduce peak demand by about 10% on average across utility programs, with wide variation in results.

Results vary based on building type, existing controls infrastructure, and the consistency of operational changes. Intervalwise provides the data visibility required to identify and act on reduction opportunities; actual savings depend on the actions your team takes. Our savings attribution is deliberately conservative: we price only the marginal reduction of the actual billed peak, and when a flagged peak does not coincide with the interval that set your bill, we attribute zero dollars to it.

Payback Period

The 8–12 month typical payback period is calculated as: annual software cost divided by annual demand charge savings at the lower bound of the 10–15% reduction range. For a $12,000/year Campus plan customer with $120,000/year in demand charges, a 10% reduction ($12,000/year) yields a 12-month payback. A 15% reduction yields an 8-month payback.

This calculation does not account for operational changes that cost nothing to implement but require behavioral adjustment (e.g., staggering HVAC startup times). Institutions achieving these low-cost changes can see payback periods under 6 months.

How Intervalwise Calculates Peak Demand

Intervalwise ingests 15-minute interval data from your utility account (via PG&E's GreenButton OAuth integration or CSV upload for Pacific Power, PSE, Avista, and Idaho Power customers). We identify the single highest 15-minute reading in each billing period per meter as the peak demand basis.

Billing demand definitions vary by tariff: many schedules bill on the highest 15-minute interval, while others use 30-minute averages, seasonal ratchets, or time-of-use demand windows — PSE's Schedule 49 (which defines billing demand on 30-minute intervals) is one example. Intervalwise verifies its calculated charges against your uploaded utility bills, models rolling ratchets where parameterized, and — rather than silently guessing — explicitly flags reduced accuracy whenever your tariff includes structures we do not yet model.

Demand peaks are attributed to buildings by aggregating the meters assigned to each building. Building-level attribution is an approximation when a single utility account covers multiple buildings; accuracy improves as meters are individually assigned.

Emissions Factors & Weather Data

Greenhouse gas figures in Intervalwise exports use the U.S. EPA's eGRID subregion output emission rates (currently eGRID2023 Rev 2, covering all 27 subregions), the same factors published in the EPA's GHG Emission Factors Hub. Your subregion is resolved from your campus locations, and every export records the exact factors used at the time it was generated.

Weather-normalized comparisons use heating and cooling degree days computed against the standard 65°F base temperature (see EIA's degree-day explainer) with a regression model trained on your own metered history. When the model's fit falls below our reliability threshold (R² of 0.75), reports omit the normalized figure and say why, rather than publish a number the data cannot support.

Questions

If you have questions about our methodology, sourcing, or calculations, email us at hello@intervalwise.com.