Sample Pilot M&V Report: An Annotated Template

Key takeaways

• This is a template: every bracketed value is a placeholder, not a result.
• A credible report states its boundary, baseline, normalization method, instruments and uncertainty.
• Results come in kWh, kW and kW per ton first, with uncertainty on each line, then dollars.
• Annual projections are estimates and should be labeled separately from measured savings.
• Agree the report format and the outcome categories before installation.

This page is a template. It shows the structure a pilot measurement and verification (M&V) report should follow, what each section must contain, and the red flags to look for when you review one. Use it to write a pilot M&V plan, to brief a third-party reviewer, or to check a report a vendor has sent you, including one from CryogenX4.

The structure follows the reporting principles of the Efficiency Valuation Organization's International Performance Measurement and Verification Protocol (IPMVP), the DOE Federal Energy Management Program (FEMP) M&V Guidelines, and ASHRAE Guideline 14's retrofit-isolation approach. A pilot report is shorter than a full performance-contract report, but it should cover the same ground. If you are new to reading these reports, start with how to read an M&V report; this template goes section by section.

Cover page and document control

Must contain: project name, site address, report version and date, author and their qualifications, independent reviewer (if any), the date and version of the signed M&V plan, and the IPMVP option used.

Pilot M&V Report, [site name], [unit IDs]. Version [x.x], [date]. Prepared by [name, credential]. Reviewed by [name, organization]. M&V plan: version [x.x], signed [date]. Method: IPMVP Option [B], retrofit isolation.

Red flags: no signed plan, or a plan dated after installation. FEMP's guidance is blunt: once a measure is installed, the baseline no longer exists to be re-measured.

1. Summary

Must contain: what was treated, the measurement boundary, the baseline and reporting periods, the headline result with its uncertainty and confidence level, and which pre-agreed outcome category applies.

[Number] [unit type] units were treated on [date]. Measured at the unit, under conditions matched to the baseline, efficiency changed from [baseline kW/ton] to [post kW/ton], a change of [x]% ± [y]% at [90]% confidence. Under the criteria in the M&V plan, this result is [a measurable improvement / no measurable effect / a measurable decline].

Red flags: a percentage with no uncertainty; a result stated for "the building" when only units were metered; the word "guaranteed."

2. Scope and units

Must contain: a table of every unit in the pilot and every comparison unit, and a description of the boundary.

The boundary statement should say what is inside the measurement (for example, compressor and condenser fan power) and what is outside (supply fan, pumps, other units). IPMVP asks for any interactive effects beyond the boundary to be described.

Red flags: units changed after the plan was signed; maintenance done between baseline and treatment; a comparison unit of a different type or exposure.

3. Baseline period and conditions

Must contain: start and end dates, the range of conditions covered, the static factors recorded, and why the period is representative.

EVO's principles say the baseline should represent a normal operating cycle, from maximum to minimum demand, and should sit immediately before the retrofit. FEMP's guidelines say chiller baselines should cover "a full range of loads during the cooling season," not spot readings.

Red flags: a baseline of a few mild days compared with a hot reporting period, or the reverse; reporting-period conditions that fall outside the baseline range. See baselines and weather normalization.

4. Treatment record

Must contain: date and time, product and quantity, technician name and certification, access point, system readings immediately before and after, and the settling period defined in the plan before reporting-period data counted.

Treated [date, time] by [technician, certification no.]. Quantity [amount] via [access point]. Readings before: [suction/discharge pressure, superheat, subcooling]. Readings after: [same]. Settling period per plan: [duration]; reporting period begins [date].

Red flags: a settling period chosen after the data were seen; no record of who installed the product.

5. Measurement points and instruments

Must contain: every measured point, the instrument, its stated accuracy, its calibration date, the logging interval, and the share of data that was captured and used.

Accuracy matters more than it looks. FEMP shows that when the temperature difference across a chiller is small, small temperature-sensor errors grow into large errors in calculated kW per ton. LBNL's 2006 chiller test cross-checked its flow meters with a separate ultrasonic measurement, which is the kind of check a good report documents. For the full list of points by equipment type, see what a CryogenX4 pilot measures.

Red flags: no accuracy or calibration data; cooling output "assumed" rather than measured; large unexplained data gaps.

6. Normalization method

Must contain: the model or comparison method, the independent variables, the fit statistics, the range over which the model is valid, and whether savings are avoided energy (under reporting-period conditions) or normalized savings (under a standard year).

Baseline model: unit kW = f([outdoor temperature], [load], [other variable]), fitted on [n] intervals of [15] minutes. R² = [value]; CV(RMSE) = [value]; net mean bias error = [value]. Valid for outdoor temperatures [min] to [max] °F and loads [min] to [max]. Savings reported as [avoided energy].

FEMP describes the logic: a model built on baseline data adjusts for expected changes such as weather (routine adjustments), and anything unexpected that changes energy use is handled as a non-routine adjustment. If the report compares condition bins rather than using a regression, it should give the number of data points per bin in each period. See IPMVP options explained for the option choice.

Red flags: no fit statistics; a model used outside its valid range; a different model for each period with no explanation.

7. Results

Must contain: results per unit, in energy units first, then dollars, with uncertainty on each line.

Any annual projection belongs in a separate, clearly labeled table, with assumed run hours and the utility rates used. EVO's principles state that IPMVP-adherent savings apply only to periods measured with adherent procedures, so a projection is an estimate, not verified savings. For converting kWh and kW to dollars, see how to calculate HVAC energy savings.

Red flags: dollar savings without kWh; an annual figure presented as measured; the comparison unit missing from the table.

8. Non-routine events and data exclusions

Must contain: a dated log of every change to setpoints, schedules, occupancy or equipment, every service visit, every sensor failure, and every block of excluded data with the reason.

[Date]: [event, e.g. setpoint changed from x to y]. Effect: [adjusted / excluded / none]. [Date range]: [n] hours excluded, reason [sensor fault].

Red flags: an empty log on a pilot lasting months; exclusions that remove mostly unfavorable periods.

9. Uncertainty

Must contain: the sources of uncertainty, how they were combined, the confidence level, and a plain statement of whether the savings exceed the uncertainty.

FEMP lists four sources: instrumentation and data collection, sampling, estimation and modeling, combined as the square root of the sum of their squares when they are independent. The LBNL report shows the honest format: savings of −2.02 kW ± 10.93 kW at 90% confidence, read as "no effect could be measured."

Combined uncertainty: ± [value] at [90]% confidence (instrumentation [value], model [value], sampling [value or n/a]). The measured change [does / does not] exceed the uncertainty.

Red flags: no uncertainty at all; a confidence level chosen after the fact to make a result look significant.

10. Oil analysis

Must contain: baseline and follow-up oil samples per unit, with total acid number, moisture and wear metals, and the lab used. LBNL ran these tests in its chiller study to check for decomposition.

Red flags: no baseline sample; samples from different points or labs before and after.

11. Conclusions

Must contain: the outcome category agreed in the plan, the limits of the result, and recommendations.

On [units], under conditions between [min] and [max] °F, the measured result is [category]. This result applies to these units and conditions. It does not predict results on [other equipment types]. Recommendation: [scale up / extend monitoring / do not proceed].

Red flags: conclusions that reach beyond the units measured; a recommendation that ignores a "no measurable effect" finding.

Appendices to request

• Raw interval data for both periods, in an open format
• Model files or spreadsheets, so a third party can reproduce the result
• Calibration certificates and sensor installation photos
• The signed M&V plan and any amendments
• Oil analysis lab reports

EVO's transparency principle is the test: an outside reviewer should be able to follow the data and analysis from start to finish.

Next step

Use this template as the reporting section of your pilot M&V plan, and agree it before installation. Our pilot program guide covers the steps that lead up to the report. When you are ready, request a pilot and ask for a report in this format.

Frequently asked questions

Why are there no example numbers in this template?

Numbers in a sample report are easily read as typical results. Placeholders keep the focus on what the report must contain. For a real example of honest reporting, see the LBNL chiller study cited on this page, which stated its result with a confidence interval.

Who should write the pilot report?

The vendor or its M&V provider usually writes it, but the plan and format should be agreed with you first. For larger decisions, ask an independent engineer to review the raw data and model.

How long should the baseline and reporting periods be?

Long enough to cover the same range of loads and weather in both periods. FEMP's guidance for chillers is to measure through a full range of loads during the cooling season, not to rely on spot readings.

Can a pilot report support a utility incentive application?

Possibly, but programs set their own M&V and pre-approval rules, sometimes including when measurement may begin. Check with the program before the baseline starts.

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 5.0 (September 2024) — U.S. Department of Energy, Federal Energy Management Program
3. Guideline 14-2023: additional material (contents, retrofit-isolation metering and whole-building model forms) — ASHRAE Journal
4. International Performance Measurement and Verification Protocol (IPMVP) — Efficiency Valuation Organization (EVO)
5. De-Scaling the Peak: The Impact of a Polarized Refrigerant Oil Additive on Chiller Performance at a U.S. Postal Service Processing & Distribution Center (2006 ACEEE Summer Study) — Lawrence Berkeley National Laboratory / ACEEE