Program administrators, implementers, evaluators and measure developers
DeepGarden for energy efficiencyAvailable through implementation
Compare measures across jurisdictions and program years. DeepGarden reconciles Technical Reference Manuals with your program data so savings, cost-effectiveness and adoption sit on consistent definitions — built with your team during implementation rather than maintained as a DeepGarden dataset.
The manuals are public. That is not the same as comparable.
That is good disclosure. None of it is designed to be compared across the boundaries anyone actually works across.
Measures are revised, superseded and retired between program years, so a portfolio compared naively against last year is comparing two things that share a name and nothing else.
Where the answer is currently spread
- Statewide and utility Technical Reference Manuals
- Measure workpapers and the documentation behind each assumption
- Program-year versions of the same manual, and the revisions between them
- Avoided-cost inputs used in TRC and PAC tests
- Your program tracking system: projects, measures installed, incentives paid
- Evaluation reports and their realization rates
Who is involved
- Measure development and regulatory affairs
- Getting a measure characterized and approved, and what precedent exists next door.
- Program design and portfolio management
- Which measures carry the portfolio, and which are cost-effective under current avoided-cost assumptions.
- Evaluation, measurement and verification
- Whether claimed savings reconcile to evaluated savings, and which measures realization rates keep punishing.
- Implementers and trade allies
- What a measure pays, under which program, in which territory, this year.
Questions this makes routine
These all cross a system boundary, which is why they currently get asked once a quarter rather than once a week.
- How is this measure characterized in each jurisdiction, and where do baselines differ?
- Which measures were revised between program years, and which comparisons does that invalidate?
- Which measures deliver the most claimed savings per incentive dollar?
- Which measures are close to failing cost-effectiveness under the current curve?
- Where does tracked installation volume diverge from what the deemed values assumed?
- Which measures are approaching the end of their effective useful life?
What DeepGarden does about it
What changes, specifically, and what lands in front of the person who asked.
The manuals, reconciled rather than collected
Measures resolved across jurisdictions and program years, so a comparison is between the same thing and a revision reads as a revision. The entity-resolution problem the platform already solves, pointed at measures.
Your tracking data beside the public record
Connect your program tracking system through Roots and the question crosses the boundary: what you installed against what the manual assumed, incentives paid against savings claimed.
An answer with the arithmetic shown
A savings figure arrives with the algorithm, the inputs, the manual version and the effective date. The destination is frequently a filing someone is paid to contest.
Questions asked in your own vocabulary
Measure, baseline, EUL, net-to-gross, realization rate, avoided cost. Agreed with your team and certified in Grove.
Why the output holds up
The parts of the governed path that matter most in this particular setting.
- Every figure opens onto the executed query, the manual and version, and the timestamp.
- A measure revised between program years is reported as revised, not as a change in savings.
- Where a jurisdiction defines a measure differently, the difference is surfaced rather than averaged away.
- Assumptions behind a derived figure are stated, because the number is contestable.
- DeepGarden reads and only reads.
How implementation would work
Eight to twelve weeks for a first jurisdiction set, depending on how many manuals are in scope
Bring a cross-jurisdiction measure question.
The kind that currently means opening several manuals and a spreadsheet. We will walk through how it gets resolved, and be straight about which parts are built and which are implementation work.