Skip to content

findings.md

Latest commit

 

History

History
47 lines (27 loc) · 4.48 KB

findings.md

File metadata and controls

47 lines (27 loc) · 4.48 KB
Raw

Key Findings — Joule Ledger

Three non-obvious findings surfaced by the dashboard across six years of EfficiencyOne reporting (2019–2024). Detail and supporting visuals are available in the Power BI model.

Finding 1: Electric programs beat their 2024 target across both energy and demand

In 2024, EfficiencyOne reported 172.8 GWh of electricity savings against a filed target of 156.56 GWh — a beat of approximately +10.4%. Demand savings reached 30.7 MW against a target of 26.42 MW — a beat of approximately +16.2%.

Demand response programs contributed an additional 8.1 MW of available capacity on top of the structural demand savings figure.

Total electric program investment was approximately $67.1 million, and an additional $79.1 million from Natural Resources Canada was distributed through the Canada Greener Homes Grant and Oil to Heat Pump Affordability programs.

What the dashboard makes visible: The Plan vs. Actual page shows the program-by-program contribution to the overall beat. The dominant contributors are the Heat Pump Program (strong uptake driven by federal co-funding) and the Home Energy Assessment program (referral pipeline maturation). Two programs — Commercial HVAC and Small Business — fell short of target, which is masked in the portfolio-level headline.

Implication for refresh users: When a new Annual Report is published, the variance page will immediately show which programs drove the aggregate result and which offset it — without re-reading the PDF.

Finding 2: Low-income programs are a structurally significant share of cumulative impact

Cumulatively since 2011, EfficiencyOne reports $5.6 billion in lifetime energy savings, of which $628 million has accrued to low-income homeowners and renters — roughly 11.2% of the total.

The equity page traces how this share has evolved year over year. Key observations:

  • The low-income share of annual spend has grown from approximately 8% (2019) to over 14% (2024), driven by the introduction of Province- and federally-funded streams alongside the DSM-Electric funded programs.
  • The dashboard separates DSM-funded electric low-income work from Province- and federally-funded programs, making the funding-source breakdown explicit. This matters because the two streams have different oversight mechanisms and reporting requirements.
  • Participant counts in dedicated low-income programs grew 3× over the six-year period, even as the electric savings per participant declined slightly — consistent with deeper weatherization work in more challenging housing stock.

What the dashboard makes visible: The Equity lens page shows share-of-spend, share-of-savings, and share-of-participation for low-income programs year by year, disaggregated by funding source.

Finding 3: Program taxonomy drift materially affects long-term trend analysis

Across six reporting years, this project identified 41 distinct program name variants resolving to 28 canonical programs. Programs have been renamed, bundled together, or split out as new federal funding streams created reporting carve-outs.

Examples of drift observed:

  • "Low Income Program" → "Efficiency NS Low-Income" → "Low-Income Efficiency" (same program, three names)
  • The Canada Greener Homes Grant and Oil to Heat Pump Affordability programs were co-delivered by EfficiencyOne but reported under federal program names from 2021 onward, creating a structural break in the residential category totals
  • "Industrial Efficiency" absorbed what was previously reported as a separate "Process Improvement" sub-program in 2022

Consequence: Without the reconciliation table, naïve year-over-year comparisons systematically overstate the apparent volatility of individual programs. A program that appears to have grown 400% may simply have been renamed; a program that appears to have vanished may have been merged into another.

What the dashboard makes visible: The Methodology page documents every mapping decision and preserves the original names for traceability. Every metric on the Plan vs. Actual page uses the reconciled program_id — so a 2019-to-2024 trend line for "Low-Income Efficiency" is actually comparable.

Recommendation for future refresh: When a new Annual Report introduces an unfamiliar program name, check sql/program_mapping.csv for a plausible canonical match before adding a new program_id. The pipeline will flag unmatched names with __UNKNOWN__ rather than silently accepting a false new program.