Definition
Measures the concentration of actual DR response among enrolled participants, revealing whether the AI is narrowing the effective pool by repeatedly signaling the same easy responders.
The Measurement Gap
Traditional DR tracking counts enrolled MW and program-average response: metrics that treat all enrolled customers as interchangeable. They do not reveal whether 300 enrolled customers are each contributing proportionally or whether 30 of them are shouldering the bulk of every event. This distinction is not academic; it determines how fragile the program is if those top contributors opt out.
The AI dispatch system's optimization logic is the proximate cause of concentration. When the platform forecasts available MW per customer, it weights customers who have responded reliably in the past more heavily. Customers with high confidence scores get dispatched more often, which drives fatigue and opt-out in exactly the customers the program most needs. The result is a self-reinforcing narrowing of the effective pool that aggregate metrics never expose.
The operational decision DRPCI enables is pool diversity management: when the index crosses a concentration threshold, the utility must implement a rotation strategy, signaling lower-confidence customers more frequently to build their response history and spread dispatch burden, even if this temporarily reduces per-event MW delivery.
What goes wrong without this measurement: the program reaches a grid emergency with 200 enrolled MW on paper, dispatches, and discovers that the 40 customers who historically delivered 170 of those MW have all opted out or reduced participation. The remaining 460 enrolled customers have been signaled so infrequently they're not conditioned to respond. This failure mode is entirely predictable from concentration trends. AI and AMI enable DRPCI by producing customer-level response attribution for every event, making the HHI computation straightforward.
Formula
DRPCI = Σ(share_i²) where share_i = MW_response_i / MW_response_total
Companion: Effective Pool Ratio:
EPR = Customers_responding_to_50%_or_more_of_events / Total_enrolled_customers_who_were_signaled
Variables:
- share_i: Customer i's share of total DR MW response across all events in the measurement period (dimensionless, 0–1). Source: AMI M&V reads per customer per event, DR management platform
- MW_response_i: Customer i's cumulative MW-hours of verified reduction delivered across all events (MWh). Source: AMI interval data
- MW_response_total: Fleet-wide verified reduction across all events in the period (MWh). Source: AMI, DR platform
- EPR numerator: Count of customers who responded to at least half the events they were signaled for. Source: DR platform signal log, AMI response records
Calculation Frequency: Seasonal (end of summer or winter program period); updated mid-season after 5+ events for early warning
Output: DRPCI as a normalized HHI, 0 to 1. Values below 0.05 indicate healthy diversity. Values above 0.15 indicate dangerous concentration. EPR below 25% means fewer than one in four enrolled customers are actively participating, a fragile capacity claim. Rising DRPCI trend across program years confirms pool narrowing.
Proposed formula, pending validation pilots.
At a Glance
Decision Supported
Is the DR program's effective MW capacity dependent on a shrinking core of reliable responders, or is participation broadly distributed?