Most measurement and verification is a report. Ours is a process. CMV2.0 maintains a weather-normalized baseline for every building and scores each 15-minute interval against it, following IPMVP Option C — so savings are proven on an ongoing basis, not estimated once a year.
Measurement and verification is how an energy-savings claim becomes a defensible number instead of a promise. The standard exists. The problem is how rarely it gets applied.
Under IPMVP Option C — the whole-facility approach — a statistical baseline is built from the building's own interval (IDR) meter data and adjusted for weather. Actual consumption is then compared against what the model says the building would have used under the same conditions. The gap between the two lines is the savings: not an estimate, but a measured difference at the same meter that generates the bill.
BOT's engine for this is CMV2.0. It maintains a weather-normalized baseline for each building and verifies actual performance against it every 15 minutes. Savings are quantified per interval, per day, and per measure. When an operating change stops paying — a schedule creeps, a setpoint gets overridden, an economizer fails quietly — the baseline-to-actual gap narrows and gets flagged the interval it opens.
The stakes justify the rigor. Commercial buildings burn roughly 40% of U.S. electricity, and most of the waste hides in operations — the kind that returns silently the moment nobody is watching the meter. Continuous M&V is the watching. For the broader software landscape this sits in, see our EMIS guide.
Most M&V today is a $10,000–25,000 engagement performed once a year. An engineer builds a baseline, issues a report, and leaves. Everything that drifts after the report ships goes unmeasured until the next one.
The annual study isn't wrong — it follows the same IPMVP methodology we do. It's just blind for 364 days. Savings from operational measures decay by nature: schedules get extended for one tenant and never pulled back, overrides accumulate, sensors fail toward "safe" and expensive. A yearly snapshot can tell you that savings eroded. It cannot tell you when, where, or why.
| Dimension | Traditional annual study | Continuous M&V (CMV2.0) |
|---|---|---|
| Methodology | IPMVP Option C, whole-facility | Same standard — IPMVP Option C, whole-facility |
| Cadence | Once a year | Every 15-minute interval, automatically |
| Cost structure | $10–25K per study, per building | Runs continuously inside the RMS platform |
| Drift detection | Discovered months after the fact | Flagged the interval the gap opens |
| Granularity | Annual or monthly totals | Per interval · per day · per measure |
| Model validation | Checked once at baseline construction | CV(RMSE), NMBE, R² gated continuously |
Both columns follow the same published methodology. The difference is frequency — and frequency is what turns M&V from an audit document into an operating instrument.
A savings number is only as good as the baseline model behind it. So the model gets judged before the number gets published — with the industry's statistics, not ours.
Every CMV2.0 baseline is validated ASHRAE Guideline 14-style: CV(RMSE), NMBE, and R² are computed and gated before any savings claim is made. If a building's model doesn't meet the bar, we don't publish a savings number for it — we fix the model. Our production baselines hold R² above 0.96, which is the difference between defending a savings figure and merely illustrating one. Weather normalization uses outside-air-temperature regressors, so a mild winter can't masquerade as an efficiency win.
We're equally strict about qualifiers. When we say 800 Wilcrest cut its annual energy roughly 50% from 2015 to 2025, we say it is time-normalized, not weather-normalized — because that decade-long comparison is a year-over-year record, not a regression output. An M&V shop that blurs that line will blur others.
One thing we deliberately won't publish: how the engine computes its baselines internally. Naming the standards we follow is accountability; publishing the implementation is just giving away the machine. You get the statistics, the methodology name, and the verified gap — the three things an auditor, a lender, or an ESG report actually needs.
We are owner-operators first: four buildings, ~421,000 SF, with 19+ years of operating expenses held below CPI. Every number below passed through the same verification engine we license. We eat our own cooking.
11200 Richmond is the cleanest recent test. We acquired it in September 2025, put it on the loop, and roughly eight months later it shows a 22.4% verified energy reduction year-to-date — verified meaning scored against its weather-normalized baseline, interval by interval, not read off a bill. On another building, peak demand dropped 36%, from 372 kW to 239 kW — the lever that moves demand charges, covered in our demand-charge guide and ERCOT 4CP explainer. And because the same interval math yields emissions, one building's ledger shows 135,841 kg of CO₂ avoided — IPMVP-grade, not a marketing estimate.
The longer arcs hold up too. 800 Wilcrest: roughly 50% energy reduction 2015–2025 (time-normalized, not weather-normalized), with a further ~25% year-over-year after our own BAS beta went in around March 2025. 11104 West Airport: ~36% annual electricity reduction 2018–2025, achieved across 27 rooftop units. Part of why those savings stick to the owner's ledger: the Atenea BAS runs on edge microcontrollers with zero licensing fees — where other landlords pay $50–200K a year in BAS licensing, we built our own for under $50 per controller.
And the verification engine now runs well beyond our own walls: the RMS platform serves roughly 600 K-12 schools plus Harris County, with 44 validated M&V reports across 15 buildings to date. One school district case: $13,140 per year saved with zero capital — pure conservation, verified.
No site visit required to start, and no capital. The whole process begins with data you already have.
Twelve months of interval data is the ideal input — a full cycle of seasons and occupancy for the baseline to learn from. CMV2.0 builds the weather-normalized model, gates it against the Guideline 14 statistics, and only then starts scoring live intervals. From your side it's simpler still: send a recent utility bill and we'll show you what's hiding in your interval data. Where operations are the low-hanging fruit — and they usually are — the first measures come straight from the HVAC optimization playbook, with a human operator reviewing every change. The verification loop tells you, in kW and dollars, whether each one held.
Send a recent utility bill and we'll show you what's hiding in your interval data: where the kWh, demand, 4CP, and power-factor money is going, and what continuous verification would hold onto.