Topic:"Demand Side Management" in M12920

June 29, 2026 Crystal Henwood Clerk of the Board Nova Scotia Energy Board 1601 Lower Water Street, 3rd Floor Halifax, NS B3J 3S3 Re: CI C0021839 IT – Customer Energy Management 2025 Evaluation, Measurement and Verification Report – Extensi...

AI summary Nova Scotia Power Inc. (NS Power) requested an extension to file its 2025 Customer Energy Management (CEM) Evaluation, Measurement and Verification (EM&V) Report due to a cybersecurity incident impacting data availability and the unavailability of the My Energy Insights platform. The report was filed in two parts due to these data limitations.

EM&V Metrics The Year 4 report reports on the metrics established in the CEM EM&V Plan and approved by the Board for which data was available for 2025. Those metrics are as follows: 2 M12251, Board Decision, CI C0021839 IT – Customer Energ...

AI summary The Year 4 report discusses metrics from the CEM EM&V Plan, approved by the Board, with data available for 2025. It references a prior Board decision (M12251) related to the CEM 2024 (Year 3) report.

Evaluation, Measurement and Verification (EM&V) Report, July 8, 2025. CEM Effectiveness Category Measurement Metric Measurement Description What Is Measured How it Is Measured Measurement Approach To Be Included in 2025 Year 4 Evaluation C...

AI summary The EM&V report evaluates the effectiveness of the Customer Energy Management (CEM) program by analyzing billing impact, energy usage reduction, and load shifts for high and low user groups. Metrics include billed amounts, peak period usage, and load profiles, with a focus on residential and business rate classes.

CEM users generally participated in EfficiencyOne programs at a higher rate than non-CEM users. › The additional annual savings were deducted from the savings captured by the DID approach so that no savings already claimed by EfficiencyOne...

AI summary CEM users showed higher participation in EfficiencyOne programs. Additional annual savings from CEM were adjusted to avoid double-counting with EfficiencyOne's DID approach.

The impact evaluation on a subset of electrically heated customers served to capture savings attributable to the CEM platform. - › High users under the Residential Standard and Residential TOD rate codes achieved statistically significant...

AI summary The impact evaluation on a subset of electrically heated customers shows statistically significant energy and load savings for some residential users under specific rate codes, particularly those on the CPP rate code. However, not all customer subsets showed significant savings, and some results were not statistically significant, indicating the need for further analysis.

Table 1: Impact Evaluation Metrics CEM Effectiveness Category Measurement Metric Measurement Description What Is Measured How it Is Measured Measurement Approach To Be Included in 2025 Year 4 Evaluation Customer Load and Energy Billing Imp...

AI summary The document outlines evaluation metrics for assessing the effectiveness of the Customer Energy Management (CEM) program, focusing on customer load, energy usage, and cost-effectiveness. It also discusses unrealized impacts due to a cyber incident and how these should be interpreted as directional estimates rather than formal evaluations.

Table 3: CEM Logins per Commercial Customer in 2024 and 2025 2024 2025 Statistical Parameters General Small General General Small General MURB All Eligible Customers Total Number of Eligible Customers 10,079 24,067 10,371 24,947 10 Number...

AI summary The table shows that commercial customers use the Customer Energy Management (CEM) tool significantly less than residential customers. As of March 2025, 66% of eligible commercial customers had activated CEM, but only 4% logged in at least once. Most commercial customers never logged in more than once, and none of the three MURB customers who activated CEM logged in.

Table 4: CEM Usage Level Thresholds per Rate Code, Based on MDef3 Metric Rate Code Low CEM User Threshold High CEM User Threshold Residential Standard ≥1 ≥2 Residential TOD ≥1 ≥2 Residential CPP and TOU ≥1 ≥3 Small General and General ≥1 ≥...

AI summary Table 4 outlines CEM usage level thresholds for different rate codes based on MDef3 login data. Econoler analyzed the data and determined thresholds, with high CEM users defined as those with ≥2 logins for most rate codes, except for Residential CPP and TOU, where the threshold is ≥3.

2 Evaluation Methodology The EM&V Plan establishes a general methodology for measuring the impacts of the CEM platform. Econoler further refined that methodology before implementing it. The EM&V Plan establishes how impacts will be estimat...

AI summary The EM&V Plan uses a difference-in-difference (DID) approach to evaluate the impact of the Customer Energy Management (CEM) platform by comparing usage data between CEM users and non-users. Due to a cyber incident and lack of AMI data, the evaluation period for Year 4 is limited to January to March 2025, and baseline periods are adjusted accordingly.

2.1 Available Data NS Power provided the following data to calculate CEM impacts: - › CEM web user login information: Number of logins by month and by customer from November 2021 to March 2025. - › Monthly electricity consumption for each...

AI summary NS Power provided data on customer energy usage and program participation to calculate CEM impacts. EfficiencyOne program participation data were used to identify confounding effects in the evaluation of CEM savings. This includes data on login activity, electricity consumption, rate codes, weather, and paperless billing.

2.3 Comparison of CEM and TVP Evaluation Methodologies The methodology used for the CEM evaluation differs from that used for the TVP evaluation, and savings measured for the CEM and TVP are not cumulative. The non-cumulative nature of the...

AI summary The evaluation methodologies for CEM and TVP differ, particularly in the composition of treatment and control groups and the reference periods used. Savings from CEM and TVP are not cumulative, especially for residential customers. While there may be some overlap in savings, the methodologies make it difficult to quantify this overlap, and thus the savings should not be combined.

3 Control Group Selection As part of the CEM Year 2 evaluation, Econoler concluded that performing the DID electricity consumption analysis on CEM users and using non-CEM users as a control group without applying further selection criteria...

AI summary The CEM Year 2 evaluation used a modified DID approach to account for electrification effects by focusing on CEM users already using electrical heating and matching residential customers on the standard rate. For TVP, control groups were defined based on electrical heating in both periods, while commercial rate codes faced limitations in sample size or participation.

3.1 Control Matching Algorithm This subsection describes the algorithm used to establish a matched control group for customers on the Residential Standard rate to gather customer load and energy usage metrics. As previously mentioned, only...

AI summary This section describes the control matching algorithm used to create a matched control group for residential customers on the Residential Standard rate. Customers using electrical space heating were considered, and matches were based on consumption similarity within clusters. Econoler set a threshold for acceptable matches, excluding less than 1% of treatment customers.

3.2 Confounding Effects Confounding effects refer to the CEM evaluation savings that should be attributed to increased participation in EfficiencyOne programs rather than to CEM usage. Econoler calculated the sum of the savings achieved th...

AI summary The text discusses confounding effects in the CEM evaluation, where savings from EfficiencyOne programs may be incorrectly attributed to CEM usage. Econoler calculated EfficiencyOne savings and adjusted the CEM evaluation by subtracting these savings to ensure accurate measurement of CEM impacts.

4.1 Customer Load and Energy Usage This subsection summarizes customer load and energy usage savings per customer for residential and commercial rate codes for the period of January to March 2025. [Table 7](#page-30-2) and [Table 8](#page-...

AI summary This subsection summarizes customer load and energy usage savings for residential and commercial rate codes from January to March 2025. Tables provide load savings for morning and evening peaks and commercial peak demand charge savings. Non-statistically significant values are highlighted in grey, indicating that CEM effects cannot be distinguished from externalities and random variations.

Table 7: Summary of Customer Load and Energy Usage Savings - Residential Rate Code CEM Usage Post Period Energy Usage Savings (kWh) Average Load Savings During Morning Peak (kW) Average Load Savings During Evening Peak (kW) Post Period Bil...

AI summary The table highlights residential energy usage and load savings under various rate codes. High users on Residential Standard and TOD achieved significant energy savings, with TOD showing the greatest reductions. TVP rates showed significant load savings during peak times, especially for CPP customers, who shifted usage away from peak periods despite non-significant overall energy savings.

4.1.3 Billing Impacts For residential customers, billing impacts are generally positive when energy usage savings are positive. For Residential Standard rate and Commercial Small General customers, the bill impacts follow the same pattern...

AI summary The section discusses billing impacts for residential and commercial customers, noting that positive energy usage savings lead to positive bill impacts. For customers on non-Standard rate codes, peak period load affects bill impacts, with CPP high users showing significant savings even with non-significant energy usage savings.

4.1.4 Alert Impacts Econoler sought to determine if the CEM high and low user groups that subscribe to alerts achieved higher savings than those who do not subscribe to alerts. This analysis was performed for the high bill alert only, the...

AI summary Econoler analyzed the impact of high bill alerts on energy savings for CEM users, finding inconsistent and non-statistically significant results due to a small non-subscribed customer group. The analysis was limited to specific rate codes and showed mixed outcomes.

4.1.5 Load Shift Analysis The EM&V Plan requires a comparative analysis of the annual energy usage load shape be performed to determine if noticeable changes occurred following the introduction of the CEM. Therefore, Econoler calculated th...

AI summary The EM&V Plan requires a comparative analysis of annual energy usage load shapes before and after the introduction of the CEM. Econoler found that the curves for all rate codes were very similar in both periods, and therefore does not recommend using this analysis to draw conclusions about the effect of the CEM on electricity consumption.

4.2.2 Paperless Billing To determine the impact of the CEM on the uptake of paperless billing, Econoler compared the percentage of customers that used paperless billing in the baseline period and then in the post period. Given that paperle...

AI summary The analysis evaluates the impact of the Customer Energy Management (CEM) program on paperless billing uptake by comparing baseline and post periods. Econoler used subscription data from 2025 and calculated expected uptake based on linear annual increases, with different timeframes for TVP participants and others.

Table 11: Percentage of 2025 CEM Eligible Customers Using Paperless Billing per CEM Status Non-activated CEM Activated CEM, Not Used Activated CEM, Used Overall Rate Code % of Paperless Billing % of Rate Code Customers % of Paperless Billi...

AI summary In 2025, customers who activated and used the Customer Energy Management (CEM) program showed significantly higher uptake of paperless billing compared to those who did not activate or use CEM. The correlation is strongest among Residential TOU customers, with non-activated CEM customers predominantly using paper billing.

While it was not possible to evaluate CEM impacts for April to December 2025, Econoler estimated the unrealized CEM impacts for that period. These estimates are an indication of the impacts on customer energy consumption (savings) that cou...

AI summary The document discusses Econoler's estimation of unrealized CEM impacts from April to December 2025, which were not fully evaluated due to a cyber incident. These estimates aim to fill data gaps, maintain continuity in CEM assessment, and provide context for partial-year results.

formance; - › Allow for continuity in the assessment of CEM; - › Place evaluated results for January to March 2025 into a broader annual context, which help in interpreting these partial-year results. To estimate the unrealized CEM energy...

AI summary Econoler estimated the unrealized CEM energy usage impacts for April to December 2025 by comparing results from January to March 2025 with the same period in 2024. A weighted average ratio was applied to monthly impact results from 2024 to estimate the unrealized impacts for the remaining months of 2025, considering the non-uniform distribution of energy savings throughout the year.

CEM users generally participated in EfficiencyOne programs at a higher rate than non-CEM users. › The additional annual savings were deducted from the savings captured by the DID approach so that no savings already claimed by EfficiencyOne...

AI summary CEM users showed higher participation in EfficiencyOne programs. Additional annual savings from CEM were adjusted to avoid double-counting with EfficiencyOne's DID approach.

The impact evaluation on a subset of electrically heated customers served to capture savings attributable to the CEM platform. - › High users under the Residential Standard and Residential TOD rate codes achieved statistically significant...

AI summary The evaluation of the CEM platform on electrically heated customers showed significant energy and load savings for certain residential rate codes, particularly under CPP and TOD. While some savings were statistically significant, others were not, indicating that external factors may have influenced the results.

Monthly Savings 90% confidence interval Figure 42: Monthly Energy Usage Savings, Residential CPP – High Users Figure 44: Average Monthly Load Savings During Morning Peak, Residential CPP – High Users Figure 45: Average Monthly Load Savings...

AI summary The text presents a series of figures illustrating monthly energy usage savings, load savings during peak times, and bill savings for residential customers participating in the Conservation Program (CPP) and Time-of-Use (TOU) programs, categorized by high and low users.

Status of the Issues Raised for Prior Reports Weather / difference-in-difference (DID). We previously accepted Econoler's position that the month-by-month DID design controls for weather by differencing the treatment and control groups ove...

AI summary The report discusses concerns with the DID methodology due to a shorter evaluation window, acknowledges Econoler's approach to E1 savings, notes the retention of electrification controls, and highlights the discontinuation of the E1 Behaviour program for integrated evaluation.

July 28, 2026 VIA EMAIL Ms. Crystal Henwood Clerk of the Board Nova Scotia Energy Board 1601 Lower Water Street, 3rd Floor Halifax NS B3J 3 S3 Dear Ms. Henwood: Re: Ml2920 - Nova Scotia Power Inc. - CI C0021839 - IT - Customer Energy Manag...

AI summary The Small Business Advocate (SBA) comments on Econoler's 2025 Evaluation of the Customer Energy Management Platform, noting concerns about limited data due to a cyber security incident and the need to maintain activation levels for commercial customers.

decision on the Year 3 CEM EM& V Report to undertake additional work with 1 M12920 - Exhibit N-1 - Econoler's Evaluation of the Customer Energy Management Platform - Year 4 (the "Report"). 2 Ml2920 - Exhibit N-1- The Report, Page 24, Figur...

AI summary The document discusses the evaluation of the Customer Energy Management (CEM) platform's Year 4 report, highlighting negative energy savings for commercial customers. NSPI cites a cybersecurity event as a barrier to new initiatives, while the SBA emphasizes the importance of engaging commercial customers. Econoler suggests that electrification may have influenced the results despite limitations in the sample size.

been considered (as was the case for the Residential Standard rate code), but the relatively small sample of available CEM users in the Commercial rate codes made this option impractical 7 . The SBA notes that this was the same explanation...

AI summary The SBA questions the reliability of the CEM platform's data, citing limited sample sizes and statistically insignificant results in the Commercial General code. It references Econoler's report and requests NSPI to confirm whether Econoler's findings are accurate and how they should be addressed in future evaluations.

Clarification of Evaluation Methodology The SBA states that it is unclear how the statement in the Report that "The evaluation results for high and low users under the Commercial General code and for low users under the Commercial Small Ge...

AI summary The SBA questions the interpretation of statistically significant energy savings in the Commercial General and Commercial Small General rate codes, while GEEG highlights that peak load reductions under Critical Peak Pricing may be driven by pricing signals rather than the CEM platform. The report uses a difference-in-differences approach to control for rate signals in evaluating TVP rates.