# Who Checks the ESCO's Numbers? Independent Review of Energy Performance Contracts > On an energy performance contract the ESCO measures the savings its fee depends on. Where an owner's review effort pays, stage by stage, under Abu Dhabi's M&V Protocol. **Source:** https://www.esgweise.com/insights/who-checks-the-esco-savings/ **Author:** Sumit Agarwal **Published:** 2026-10-11 **Frameworks:** IPMVP, Abu Dhabi M&V Protocol **Countries:** UAE --- ## Key points - On an energy performance contract the ESCO designs the measures, often finances them, and also measures the savings that its payment and guarantee depend on. That is a governance question for the owner, not a technical footnote. - Under the Abu Dhabi M&V Protocol (DoE/PD/P04/005, version 0, effective 2 January 2022), the annual M&V report becomes the basis of the ESCO's invoice once the customer has reviewed and approved it. Approval is the owner's control point. - The Protocol does not require third-party review. Following IPMVP, it recognises that an owner with less experience than the party determining savings may seek help reviewing the savings reports. - Most review value sits in a few places: the baseline period and its conditions, the choice of IPMVP option, stipulated values under Option A, and the adjustments. Most disputes on an EPC are adjustment disputes. - The cheapest point to review is before signature, when the M&V plan is still being negotiated. After that, each annual report is the next opportunity, and the last one before money moves. --- ## Introduction An energy performance contract is an elegant piece of finance. An energy service company (ESCO) identifies efficiency measures in a building or facility, installs them, often funds them, and is repaid out of the savings. If the savings do not arrive, the ESCO does not get paid in full. Risk sits with the party that controls the engineering. A second feature gets less attention at signature. The same ESCO usually writes the measurement and verification (M&V) plan and prepares the annual savings report. The number that decides its payment is a number it calculates. None of this implies bad faith. Determining savings involves real judgement, and the ESCO often knows the measures best. But a board that signs off a multi-year payment stream should be able to say who, on its side, checked the arithmetic and the assumptions behind it. ## The arrangement, as Abu Dhabi's protocol describes it In Abu Dhabi the reference document is the **Abu Dhabi Measurement and Verification Protocol**, a Department of Energy guidance document, **document no. DoE/PD/P04/005, version 0, effective 2 January 2022**. It gives guidance on M&V for energy performance contracting projects carried out by ESCOs, treats energy and water savings interchangeably, and is distributed to **Abu Dhabi Energy Services (ADES)** and **all ESCOs operating in Abu Dhabi**. It is built on **IPMVP Core Concepts, EVO 10000-1:2016**, and the **US Department of Energy FEMP M&V Guidelines, version 4.0, November 2015**. Its definitions describe the commercial structure plainly. An ESCO is an organisation that designs, finances, procures, installs and possibly maintains the measures. Under an EPC, the project is repaid from the stream of cost savings, and the ESCO is not paid unless the expected savings are delivered. Then comes the clause an owner should read twice. The Protocol defines the **annual report** as the document that records the year's M&V activities and results, and states that in an EPC the savings it documents serve as the **basis for the ESCO's invoice, after the customer has reviewed and approved the report**. Under the Protocol, the owner's review and approval of each annual report is what turns a calculated saving into an invoice. That approval is the owner's main control over the contract for its whole term. An approval given without the capacity to test the report is a signature, not a control. ## What the Protocol says about review Less than people assume. The section headed third-party verification is a single sentence, quoting IPMVP to the effect that where the firm determining the savings has more experience than the owner, the owner may seek assistance in reviewing the savings reports. The annex on the M&V plan adds that the plan may be developed by the ESCO and may be reviewed by a qualified third party, giving a Certified Measurement and Verification Professional (CMVP) of the Association of Energy Engineers as an example. Both provisions are permissive. Nothing in the Protocol obliges an owner to appoint a reviewer, and nothing prevents it. The decision is a governance one, and it belongs to the owner. ## Why this matters in Abu Dhabi now The retrofit market in the emirate is being built deliberately. The **Abu Dhabi Demand Side Management and Energy Rationalisation Strategy 2030** targets a 22% reduction in electricity and 32% in water by 2030 against a 2013 baseline, and the Building Retrofits Programme is one of its nine programmes. **ADES, created in 2020 by TAQA**, is the emirate's Super ESCO for government building retrofits, with a stated scope of 3,000 government buildings and targets of 2.7 TWh of electricity and 9 million cubic metres of water saved by 2030, funded on the performance contracting model. That makes M&V contractual rather than advisory across a large and growing portfolio. For why Abu Dhabi creates audit work through contracts rather than a general audit mandate, see [what Abu Dhabi actually mandates on energy audits](/insights/abu-dhabi-energy-audits-what-is-mandated/). ## Where review effort pays Savings cannot be measured directly. The Protocol says so in terms: a saving is the absence of energy use, so it is determined by comparing measured use before and after the measures and adjusting for changed conditions. Every saving is therefore an estimate built on a baseline, a model and a set of adjustments. Review effort pays where judgement enters that chain. ### The baseline period and its conditions The baseline is the counterfactual against which every future year is measured. The Protocol's M&V plan outline requires the baseline period to be identified with the reasons for choosing it, together with baseline consumption and demand, the independent variables and the **static factors**: the equipment, schedules and operating modes treated as fixed. If any static factor later changes, the baseline has to be adjusted. A reviewer asks whether the chosen period was representative and whether the static factors were recorded in enough detail to recognise a change when it happens. A baseline that overstates pre-retrofit consumption inflates every year of savings that follows. A thin record of baseline conditions makes every later adjustment an argument. ### The choice of IPMVP option The Protocol sets out the four IPMVP options. Options A and B isolate the retrofitted equipment, Option C uses whole-facility meter data and regression against variables such as weather and occupancy, and Option D uses a calibrated simulation. Its selection diagram, taken from EVO, turns on whether the measure can be isolated with meters and whether the expected savings are large enough to show at the main meter. The question for a reviewer is whether the option fits the measure and the risk. Option C on a site where savings are small relative to total consumption buries the saving in noise. Option A on a measure whose operating hours are the main uncertainty stipulates away the variable that matters. The detail of each option is in our companion piece on [the Abu Dhabi M&V Protocol and IPMVP Options A to D](/insights/abu-dhabi-mv-protocol-ipmvp-options/). ### Stipulated values under Option A Option A measures the key parameter and stipulates the rest. A stipulated value is held constant whatever actually happens, so the measurement process for that parameter is eliminated. The Protocol lists stipulation of parameters as one of three sources of savings risk, alongside usage risk and performance risk. Stipulation is legitimate, but it is a commercial term disguised as an engineering one: whoever sets the stipulated operating hours has fixed part of the saving in advance. A reviewer checks what each stipulated value rests on and who bears the consequence if it is wrong. The Protocol's own principles say M&V should under-estimate savings where judgements are made about uncertain quantities. ### Routine and non-routine adjustments The Protocol defines two kinds. **Routine adjustments** account for expected variation in the independent variables, and are calculated through the baseline model. **Non-routine adjustments** compensate for unexpected changes unrelated to the measures: equipment added or removed, a change in operation, new floor area. Most disputes on an energy performance contract are adjustment disputes. The baseline is agreed once. Adjustments are argued every year. Routine adjustments are mechanical once the model is accepted, so the review point is the model itself. The Protocol's plan template asks for the baseline regression model for each site, with IPMVP-recommended thresholds for the coefficient of determination, the coefficient of variation of the root mean squared error and the t-statistics. A reviewer recalculates these rather than accepting them, using cooling degree-day data from the UAE National Center of Meteorology, the source the template points to. Non-routine adjustments are where judgement enters. Each should be traceable to a documented change in a static factor, quantified by a stated method, and marked temporary or permanent. A reviewer also looks at the direction of every adjustment over the life of the contract. Changes consistently treated as adding to baseline energy are worth a conversation. ### Meters, calibration and data gaps The Protocol treats the utility's own main meter as compliant without further validation. Every other meter is not. For non-utility meters the M&V plan has to specify the meter characteristics, the reading and witnessing protocol, commissioning, routine calibration and the method for dealing with collection errors and lost data. That last item decides more savings than people expect. A logger that fails for six weeks in summer leaves a gap someone has to fill, and the method is a choice. A reviewer checks calibration records against the meter list and tests whether each gap was filled as the plan specified. ### Interactive effects Interactive effects are changes in one system's energy use caused by changes to another. A lighting retrofit reduces heat gain and therefore cooling load. Under retrofit isolation these effects sit outside the measurement boundary unless the plan deals with them. The question is whether they were assessed, and whether the treatment is the same every year. ### Guaranteed against verified savings The Protocol distinguishes three figures. **Expected savings** come from the post-installation report and are what year one should deliver given as-built conditions. **Verified savings** are what the annual report calculates for that year. The **guaranteed** amount is the contractual commitment. The annual report has to compare verified savings with the guaranteed amount, and the Protocol's report format includes cumulative savings from the start of the project, year by year. A reviewer reads that comparison across years rather than in isolation. A contract that meets its guarantee every year only after progressively larger non-routine adjustments is telling the owner something. ### Persistence over the term Measures degrade. Controls are overridden and setpoints drift. The Protocol expects M&V at regular intervals, typically annually, with operational verification that each measure is still operating as intended. A reviewer checks that this actually happened, and that the O&M summary in each annual report is consistent with the savings claimed. ## What a reviewer looks at, by stage The Protocol sets out five steps to determine and verify savings: allocate risks and responsibilities, develop the project-specific M&V plan, install and commission, conduct post-installation verification, and perform regular-interval M&V. Review effort maps onto them as follows. | Stage | What the owner's reviewer tests | |---|---| | **Contract and M&V plan** | Risk and responsibility allocation (the Protocol's annex gives a matrix covering financial, operational and performance risks), option choice and measurement boundary, stipulated values, adjustment rules, meter specifications, data gap method, expected accuracy, energy prices used to value savings | | **Baseline** | Representativeness of the period, completeness of baseline conditions and static factors, regression model and its statistics, source and quality of independent variable data | | **Post-installation** | Installed equipment against the contract, commissioning results kept separate from M&V results, performance measurements, the basis of expected first-year savings | | **Each annual report** | Recalculation of savings, every adjustment and its evidence, calibration and data gaps, verified against guaranteed savings, cumulative trend, O&M issues raised | The M&V plan is negotiated with the EPC contract, and the Protocol calls it the most important single item in a savings guarantee. Every later report is calculated on its terms. A review at plan stage can change the terms. A review at year three can only test whether the terms were followed. ## What independence means Independence here is a commercial fact, not a credential. A reviewer is independent of the ESCO if it has no stake in the measures or the savings: it does not design, sell, finance or install them, and its fee does not rise or fall with the verified figure. A reviewer paid a share of savings, or one hoping to win the next phase of works, is not independent however competent it is. ## Questions a board should ask For an owner with an EPC in force or in negotiation, these are the questions to put to management. 1. Who wrote the M&V plan, and who on our side understood it well enough to approve it? 2. Why was this baseline period chosen, and what were the baseline conditions? 3. Which values are stipulated, on what basis, and who bears the risk if they are wrong? 4. How many non-routine adjustments have been made since acceptance, in which direction, and who proposed each one? 5. Which meters feed the savings figure, who owns them, and when were they last calibrated? 6. How were gaps in the data filled, and is that the method the plan specified? 7. What were verified savings against the guarantee in each year, and what is the trend? 8. Has anyone with no financial interest in the outcome reviewed the plan or the reports? If the answer to the last question is no, the board is approving an invoice on the word of the party issuing it. ## How ESGweise helps ESGweise works on the owner's side. We carry out [energy audits](/services/energy-audits/) and [chiller audits](/services/chiller-audits/) that establish an independent baseline before an EPC is tendered, review M&V plans against the Abu Dhabi Protocol and IPMVP before signature, and review ESCO post-installation and annual savings reports for the owner. We do not sell, finance or install energy conservation measures, and our fees do not depend on the savings reported. That work sits within our [assurance practice](/services/assurance/), and where an owner wants the gains held after the contract ends, within an [ISO 50001 energy management system](/insights/iso-50001-energy-management-system/). ## References and sources - Abu Dhabi Department of Energy, *Abu Dhabi Measurement and Verification Protocol, Guidance Document*, document no. DoE/PD/P04/005, version 0, effective 2 January 2022 ([doe.gov.ae](https://www.doe.gov.ae/)) - Efficiency Valuation Organization, *IPMVP Core Concepts*, EVO 10000-1:2016 (the edition the Protocol references; EVO 10000-1:2022 is the current edition) - US Department of Energy, Federal Energy Management Program, *M&V Guidelines: Measurement and Verification for Performance-Based Contracts*, version 4.0, November 2015 - Abu Dhabi Department of Energy, *Abu Dhabi Demand Side Management and Energy Rationalisation Strategy 2030* ## Conclusion An energy performance contract moves engineering risk to the ESCO. It does not move the duty to check what the owner is paying for. Abu Dhabi's Protocol makes the owner's approval of each annual report the step that turns a calculated saving into an invoice, and leaves it to the owner to decide whether that approval is informed. The checking is concentrated in a few places, above all the adjustments. It is cheapest before signature, and worth doing every year after. --- ## Frequently asked questions ### Does the Abu Dhabi M&V Protocol require independent review of an ESCO's savings? No. The Protocol's section on third-party verification is short and permissive. Following IPMVP, it recognises that where the firm determining the savings has more experience than the owner, the owner may seek assistance in reviewing the savings reports. Separately, its outline of the M&V plan says the plan may be developed by the ESCO and may be reviewed by a qualified third party. Neither is a requirement. Whether to use a reviewer is the owner's decision. ### Who prepares the M&V plan and the savings reports on an energy performance contract? Usually the ESCO. The Protocol describes the project-specific M&V plan as being developed during EPC contract negotiations, and post-installation verification as an activity the ESCO conducts. The customer's role is to review and approve the reports, and in an EPC the approved annual report is what the ESCO invoices against. ### What is the difference between a routine and a non-routine adjustment? A routine adjustment accounts for expected variation in the independent variables named in the M&V plan, such as cooling degree days or occupancy, usually through the baseline model. A non-routine adjustment compensates for an unexpected change unrelated to the energy conservation measures, such as added floor area, new equipment or a change in operating hours. Non-routine adjustments are where judgement enters, and therefore where most disputes start. ### When is the best time to bring in an independent reviewer? Before the contract is signed, while the M&V plan is still being negotiated. The baseline period, the measurement option, the stipulated values and the adjustment rules are all fixed in the plan. Once it is signed, every annual report is calculated on those terms, and a reviewer can only test whether the terms were followed. ### What makes a reviewer independent of the ESCO? Having no financial interest in the measures or the savings. An independent reviewer does not design, sell, finance or install the measures, and is not paid on the size of the verified savings. Their only interest is whether the reported figure is correctly calculated from the agreed plan and the evidence. --- Published by ESGweise Global LLC, Dubai. https://www.esgweise.com