What GridScore measures
GridScore answers a simple question: when the grid needs capacity most, is this resource delivering?
Specifically, GridScore is the percentage of a distributed energy resource’s monthly energy savings that occur during the top 10% of net demand hours on the grid. Net demand is total grid demand minus solar and wind generation, reflecting the hours when the grid needs the most support. It is expressed as a percentage from 0 to 100.
A GridScore of 80% means that 80% of the resource’s energy savings happen during the hours when net grid demand is highest. A GridScore of 10% means the savings are concentrated during off-peak hours.
Why it matters
Not all kilowatt-hours are equal. A kilowatt-hour saved during the evening ramp, when solar generation drops off and the grid needs capacity, is worth significantly more than one saved at midday when abundant solar is already serving load.
Traditional energy metrics treat all savings the same. GridScore differentiates resources by when they deliver value to the grid’s capacity needs, which is exactly what grid operators, utilities, and capacity markets care about.
Using net demand rather than total demand means GridScore correctly identifies the hours when the grid is actually stressed. In regions with high solar penetration (like CAISO), total demand peaks midday but the grid has ample solar generation to serve it. The real constraint is the evening ramp after solar drops off, and winter mornings before solar ramps up. Net demand captures this.
This has practical implications:
- Resource adequacy. Utilities need to know which DERs can be counted toward planning reserve margins. A high GridScore means the resource reliably delivers during the hours that matter for reliability, not just during hours that happen to have high total demand.
- Program design. Program administrators can use GridScore to evaluate which types of DERs deliver the most capacity value per dollar spent, accounting for the grid’s evolving generation mix.
- Competitive differentiation. Energy service providers with high GridScore portfolios can demonstrate that their assets deliver value when the grid actually needs help, not when renewables are already doing the job.
How it’s calculated
The calculation has three steps:
Step 1: Identify peak net demand hours. For a given month, calculate net demand for each hour (total grid demand minus solar and wind generation in the resource’s region). Rank all hours by net demand, descending. The top 10% are classified as peak hours. For a 30-day month with 720 hours, this is the 72 highest net-demand hours.
Step 2: Measure verified savings. Using meter-based M&V (via Aristotle), calculate the resource’s actual energy savings for every hour of the month. This uses a statistical baseline model that accounts for weather, occupancy, and other variables.
Step 3: Calculate the ratio. Divide the savings that occurred during peak net demand hours by total savings for the month.
GridScore = (savings during top 10% net demand hours) / (total savings) × 100If a resource saved 1,000 kWh in a month and 800 of those kWh were during peak net demand hours, its GridScore is 80%.
What drives high and low scores
Different resource types tend to have characteristic GridScore profiles:
Typically high GridScore:
- Demand response (dispatched specifically during peak net demand events)
- Battery storage (discharged during evening ramp and peak net demand periods)
- Heat pumps in heating-dominant climates (savings during winter mornings align with high net demand)
Typically moderate GridScore:
- Air conditioning efficiency improvements (summer cooling loads partially overlap with evening net demand peaks, but midday cooling savings fall outside peak net demand when solar is abundant)
- Heat pumps in mild climates (depends on whether savings concentrate in morning/evening hours)
Typically lower GridScore:
- Lighting efficiency (savings spread across operating hours, not concentrated during net demand peaks)
- Baseload efficiency measures (constant savings regardless of grid conditions)
- Solar-only resources (generation peaks when net demand is lowest)
These are generalizations. Actual GridScores depend on the specific asset, its location, and the local grid’s net demand profile. That’s why meter-based verification matters: modeled estimates give a rough idea, but measured data from Aristotle gives the real number.
Modeled vs. verified scores
Assets enrolled in Aristotle receive verified GridScores calculated from actual meter data. These are updated continuously as new data arrives.
Assets not yet enrolled can receive modeled GridScores estimated from the asset type, location, and typical performance characteristics. Modeled scores are useful for initial assessment but carry less certainty.
On the DER Registry, verified scores are marked with a “Verified” badge and modeled scores with a “Modeled” badge. Enrolling an asset in Aristotle upgrades its score from modeled to verified.
Open methodology
GridScore is not a proprietary metric. The calculation methodology is published as open methodology by the OpenEAC Alliance, where it is subject to peer review and public comment. Any stakeholder can review the approach, validate the math, and propose improvements.
This transparency is deliberate. A scoring system that influences investment decisions and program design needs to be auditable and trustworthy, not a black box.