EvaluateProof, Testing & Measurement6 min readUpdated

Key takeaways

  • A credible report states the IPMVP option, measurement boundary, baseline and reporting dates, data, adjustments and calculations.
  • Savings should appear in energy units (kWh, kW) before dollars, with the utility rates used shown separately.
  • Look for model statistics or bin counts, instrument accuracy, and how missing data were handled.
  • Red flags: no baseline, a percentage with no kWh, unexplained exclusions, undisclosed changes, or claims of IPMVP certification.
  • A lab test report and a field M&V report answer different questions; do not accept one in place of the other.

An M&V report is where a savings claim either holds up or falls apart. You do not need to be an engineer to review one. You do need to know what should be in it and where weak reports usually cut corners. This guide follows the structure set out by the Efficiency Valuation Organization (EVO) for IPMVP-adherent reporting and the Federal Energy Management Program (FEMP) for federal performance contracts.

What a complete report contains

EVO's Generally Accepted M&V Principles list the minimum contents of an M&V report:

  • Project background and a description of the energy conservation measure (ECM)
  • The IPMVP option chosen in the M&V plan
  • Reporting period start and end dates, and measurement period start and end times
  • Energy use data
  • Data for independent variables (such as weather) and static factors (such as set points and occupancy)
  • A description of inspections carried out
  • Verified savings calculations, with a detailed description of the analysis method
  • An updated list of assumptions and data sources
  • Details of all routine and non-routine adjustments
  • The utility costs used to calculate savings
  • Verified energy and cost savings, compared with the proposed savings

FEMP's annual report requirements for federal energy savings performance contracts in the M&V Guidelines add results of performance measurements and inspections, a comparison of verified with guaranteed savings, the commodity rates used, a summary of operations and maintenance activities, and any performance issues that need attention. A report on a short pilot will be lighter, but the same logic applies.

Review it in this order

1. Boundary and option

What exactly was measured: one compressor, a chiller, a plant or the whole building? Which IPMVP option was used? If the report states a percentage saving, check what it is a percentage of. Savings measured at a treated unit are not savings on the facility bill, and one figure should not be presented as the other. EVO also requires the plan to describe interactive effects beyond the boundary and their likely size.

2. The baseline

Look for the baseline dates, the range of weather and load it covered, and the static factors recorded at the time. Ask whether known equipment faults were repaired before the baseline. EVO allows a baseline to be adjusted to reflect operation after needed repairs, but only if this is disclosed. If the baseline covered only mild weather and the reporting period only hot weather, look closely at how the two were compared. See baselines and weather normalization.

3. The adjustment method

If a regression model was used, EVO requires the report to include its coefficients, statistical metrics (CV(RMSE), mean bias error, R-squared, t-statistics) and the range of conditions over which the model is valid. If periods were compared by condition bins, ask for the number of data points per bin in each period and how bins were weighted into an overall result. Check whether the result is avoided energy (under actual reporting conditions) or normalized savings (under typical weather). The report should say which.

4. Data quality

EVO's plan requirements cover meter type, make, model, accuracy, calibration and commissioning, and the method for handling lost data. FEMP points out that measurement uncertainty starts with instrument accuracy and that data management can add error through omitted, adjusted or lost data. A good report states how much data was missing or excluded and why.

5. Uncertainty

FEMP explains that every statement of measured savings includes some uncertainty, coming from measurement, sampling, estimation and modeling, combined statistically. For chillers, FEMP shows how small temperature-sensor errors grow into large kW-per-ton errors when the chilled-water temperature difference is small. A report does not always need a formal uncertainty calculation (FEMP notes it is not required for federal projects), but it should show that the savings are clearly larger than the measurement error.

6. Energy first, then dollars

Savings should be calculated in kWh and kW first, then converted to dollars using stated rates. EVO requires the plan to specify the tariffs used and how cost savings will be adjusted if prices change. Dollar figures that mix rate changes with efficiency changes are not verified savings. For the conversion math, see how to calculate HVAC energy savings.

Red flags

Red flagWhy it mattersWhat to ask
No measured baseline, or a baseline taken from a different unit or siteSavings cannot be determined without a "before" on the same equipment"Where and when was the baseline measured?"
A percentage with no kWh, kW or operating hoursPercentages hide the size of the base and the boundary"What were baseline and reporting-period kWh, and what boundary do they cover?"
Before-and-after bills with no weather adjustmentWeather alone can swing cooling energy substantially"How were differences in weather and load adjusted for?"
Selected days or "typical" days onlyCherry-picked periods can show anything"What share of the collected data is in the analysis, and what was excluded and why?"
Changes during the test not mentionedSet-point, schedule or maintenance changes can create or erase savings"Please provide the change log and any non-routine adjustments."
"IPMVP certified" resultsEVO states IPMVP does not currently provide formal certification of project M&V plans"Which IPMVP option was used, and can we see the adherent M&V plan?"
Future savings projected from a short test and labeled IPMVP-adherentEVO says adherent savings apply only to reporting periods measured with adherent procedures"Which periods were measured, and which are projections?"
No instrument detailsSensor error can be as large as the claimed savings"What instruments were used, what is their accuracy, and when were they calibrated?"
Lab results presented as site savingsLab conditions are controlled and differ from your equipment and load"What was measured on our equipment?"

Lab reports and field reports answer different questions

Vendors in HVAC-R efficiency often present laboratory reports alongside or instead of field results. Lab tests have a real role. A sealed-tube chemical stability test, for example, addresses whether a product is compatible with refrigerants and oils. But a lab test of compatibility or wear performance does not measure what your equipment will save. Field M&V on your own equipment is the evidence for savings. See laboratory testing standards for what each type of test can prove.

Questions to send back to any vendor

  1. Which IPMVP option did you use, and where was the measurement boundary?
  2. What are the baseline and reporting-period dates, and what range of outdoor conditions does each cover?
  3. How were weather, load and operating differences adjusted for? Please share the model or the bin counts.
  4. What instruments were used, with what accuracy and calibration dates?
  5. How much data was missing or excluded, and why?
  6. Were there any changes to set points, schedules, occupancy or maintenance during the test?
  7. Are the savings avoided energy or normalized to a typical year?
  8. Can we have the raw interval data so our own engineer or a third party can check the calculation?

The last question matters most. EVO's transparency principle says data, methods, assumptions and calculations should be documented so that any outside reviewer can follow them. A vendor confident in its results should be comfortable sharing the data. For a broader due-diligence list, see questions to ask efficiency vendors.

A short example of reading the headline correctly

Suppose a report (hypothetical numbers) says "Unit kW reduced 12% under matched conditions." Read it carefully:

  • "Unit kW": the boundary is the equipment, not the building. The facility bill will change by less, depending on the unit's share of total use.
  • "Reduced 12%": ask whether this is lower kW at the same cooling delivered (an efficiency change) or simply lower kW (which could mean less cooling).
  • "Matched conditions": ask how many matched intervals there were, what temperature and load range they covered, and how they translate to an annual figure.

Once those three answers are in hand, a 12% unit-level improvement can be turned into annual kWh and dollars with a defensible method.

How CryogenX4 approaches reporting

CryogenX4 describes a field method that captures indoor and outdoor temperature, humidity, chilled-water or air flow and energy use every minute, over a testing period of usually 3 to 9 months. A baseline is captured before the one-time treatment, and performance is compared under similar weather and building load conditions. The company reports savings of up to 30%, and results vary by equipment condition. The questions above are the right ones to ask of any CryogenX4 result as well. A sound report on your own equipment should answer each of them. See what a CryogenX4 pilot measures for the data behind it.

Next step

Take the most recent savings report you have from any vendor and check it against the contents list and red-flag table above. Note which answers are missing and ask for them in writing. If you would like a CryogenX4 pilot reported to this standard, get in touch.

Frequently asked questions

Should an M&V report include raw data?

It should at least make raw data available on request. EVO's transparency principle calls for data, assumptions and calculations to be documented so an outside reviewer can follow them, and EVO's Option D requirements call for measured data to be archived and made available.

What is the difference between proposed, expected and verified savings?

In FEMP's terms, expected savings are those in the post-installation report, based on as-built conditions, and verified savings are those reported after measurement in the annual report. EVO also asks reports to compare verified savings with the savings originally proposed.

Is a higher R-squared always better?

Not on its own. R-squared shows how much variation a model explains, but a report should also give CV(RMSE) and bias, and state the range of conditions the model is valid for. A model used outside that range can mislead even with a high R-squared.

Can I hire someone to review a vendor's M&V report?

Yes. FEMP suggests engaging a third-party M&V expert to help develop the plan and protect the customer's interests. An independent engineer can recheck the baseline, the adjustments and the calculations from the raw data.

Sources

  1. IPMVP Generally Accepted M&V Principles (October 2018) — Efficiency Valuation Organization (EVO)
  2. M&V Guidelines: Measurement and Verification for Performance-Based Contracts, Version 4.0 (November 2015) — U.S. Department of Energy, Federal Energy Management Program
  3. M&V Guidelines: Measurement and Verification for Performance-Based Contracts, Version 5.0 (September 2024) — U.S. Department of Energy, Federal Energy Management Program
  4. Guideline 14-2023: additional material (contents, retrofit-isolation metering and whole-building model forms) — ASHRAE Journal
  5. International Performance Measurement and Verification Protocol (IPMVP) — Efficiency Valuation Organization (EVO)

Keep reading

See what your equipment could save

CryogenX4 is a one-time treatment installed while your system runs. Start with a pilot on a few units, measured against a baseline, before you commit to a building or a portfolio.