By 5 December 2027, every large undertaking caught by ESOS Phase 4 must notify compliance, and a certified energy management system (EnMS) is one of the routes[1][1]Regulation and standardsGOV.UK, “How to comply with the Energy Savings Opportunity Scheme (ESOS) Phase 4”, Environment Agency guidance (qualification, compliance routes, ISO 50001 route)See in sources ↓. ESOS is based on size, not energy use: 250 or more employees, or turnover above £44m and a balance sheet above £38m, assessed at group level[1][1]Regulation and standardsGOV.UK, “How to comply with the Energy Savings Opportunity Scheme (ESOS) Phase 4”, Environment Agency guidance (qualification, compliance routes, ISO 50001 route)See in sources ↓. ISO 50001 is never mandatory, but the EnMS route requires a certificate to BS EN ISO 50001:2018+A1:2024[2][2]Regulation and standardsBSI Knowledge, “BS EN ISO 50001:2018+A1:2024 Energy management systems. Requirements with guidance for use”, current UK editionSee in sources ↓. An ISO 50001 auditor will expect indicators that hold up.
A raw kWh/m² is not one of them: a raw ratio proves nothing. An energy performance indicator (EnPI) relates energy consumption to what drives it. The method: choose, normalise, make the data reliable, then link every drift to a decision.
What is an energy performance indicator, and which EnPIs suit offices, hospitals or factories?
A ratio between consumption and what drives it
In ISO 50001 terms, an energy performance indicator (EnPI) is a measure or unit of energy performance defined by the organisation, expressed as a simple metric, a ratio or a model[3][3]Standards and protocolsISO, “ISO 50001 — Energy management”, definition of an EnPISee in sources ↓. This quantitative value relates the energy consumed to what drives it: weather, occupancy or production. Without that denominator, the figure describes a bill, nothing more.
A building’s heating is related to heating degree days (HDD), which reflect outdoor temperature. On a production line, electricity is related to the units produced. This principle underpins the monitoring required by ISO 50001 and any defensible energy performance measurement.
What ISO 50001 and ISO 50006 require
The ISO 50001 standard requires the EnMS to compare its EnPIs with a quantified baseline. In the UK it is published by BSI as BS EN ISO 50001:2018+A1:2024, the 2018 edition amended in 2024[2][2]Regulation and standardsBSI Knowledge, “BS EN ISO 50001:2018+A1:2024 Energy management systems. Requirements with guidance for use”, current UK editionSee in sources ↓.
That baseline is the energy baseline (EnB). EnMS software calculates it but does not choose it for you. To build it, a sister standard serves as a guide:
Did you know? ISO 50001 requires indicators without saying how to build them. ISO 50006, in its 2nd edition of 2023, governs how they are built, along with the EnB[4][4]Standards and protocolsISO, “ISO 50006:2023 Energy management systems — Evaluating energy performance using energy performance indicators and energy baselines”, 2nd editionSee in sources ↓.
Why is kWh/m²/yr not enough?
The simplest ratio is still the most quoted: kWh/m²/yr, the unit of UK benchmarks such as CIBSE TM46 and the Better Buildings Partnership REEB benchmarks[5][5]UK benchmarksCIBSE TM46 (2008), “Energy benchmarks”, General office benchmark cited in a UK government paper | https://assets.publishing.service.gov.uk/media/5a7e1520e5274a2e8ab45b9e/EED_regs_-_benchmark_heat_demand_paper_-_261114_.pdf ; Better Buildings Partnership, “2025 Real Estate Environmental Benchmarks” (REEB) | https://www.betterbuildingspartnership.co.uk/sites/default/files/media/attachment/REEB%20Benchmarks%202025.pdf ; Display Energy Certificates: Energy Performance of Buildings (England and Wales) Regulations 2012, SI 2012/3118, regulation 14See in sources ↓. TM46 also underpins Display Energy Certificates (DECs), which rate public buildings on metered energy, whereas an EPC rates modelled energy. An EnMS works with metered data, and so do NABERS UK ratings for offices. The same ratios feed sustainability reporting, as explained in ISO 50001 and ESG reporting. The table below sets each ratio against what it reveals and what it hides; filter it by site type.
| Indicator | Unit | Site type | What it shows | What it hides |
|---|---|---|---|---|
| Energy use per floor area Raw ratio | kWh/m²/yr | Offices, commercial buildings | The unit of CIBSE TM46, REEB and DEC benchmarks [5]. | Weather, occupancy and end uses merge into a single figure. |
| Weather-normalised heating Weather-normalised | kWh/HDD | Any heated site | How sensitive heating is to outdoor conditions. | The drifting zone, lost in the total. |
| Weather-normalised floor-area ratio Weather-normalised | kWh/(m²·HDD) | Comparing two sites | The gap between sites in different climates. | Differences in operating hours and end uses. |
| Out-of-hours baseload Operations | kW | Offices, care homes | AHUs running at night. | The cause: schedule, manual override or faulty sensor. |
| Energy per occupied hour or per bed Activity-normalised | kWh/occupied hour, kWh/bed | Offices, hospitals, care homes | The effect of actual activity. | Everything, if occupancy is poorly counted. |
| Energy per unit produced Activity-normalised | kWh/unit, kWh/tonne | Industrial site | The efficiency of a line. | Changes in product mix. |
| COP of a chiller or heat pump Utility | kWh cooling/kWh electrical | Commercial and industrial [6] | The efficiency of the machine. | Nothing, if the controls force the equipment on. |
| Compressed air Utility | Wh/Nm³ | Industrial site [6] | The efficiency of the compressor house. | Network leaks, produced at good efficiency. |
| Steam Utility | kWh GCV/tonne of steam | Industrial site [6] | The efficiency of the boiler house. | Distribution losses downstream. |
The first row of the table sums up the problem: kWh/m²/yr smooths everything out. A mild winter, an empty floor and an air handling unit (AHU) left on manual override all end up in the same energy use per square metre. Normalised ratios, utility ratios[6][6]Studies and reportsCertinergy, industrial energy performance indicators page (in French), utility indicatorsSee in sources ↓ and the night-time baseload say more, provided you take into account the controls that produce them.
Offices, hospitals, care homes: energy indicators for a commercial building
For an office commercial building or a care home, four ratios are enough: heating per heating degree day, out-of-hours baseload, consumption per occupied hour and breakdown by end use. The BMS (building management system) already trend-logs these points, domestic hot water included. The link between energy use and occupancy rate is covered separately.
Hospitals are the exception. Occupied day and night, their baseload is high by nature and loses its value as a warning. Per-bed ratios then take over.
Factories and utilities: consumption ratios for an industrial site
On an industrial site, each utility calls for its own ratio. Refrigeration is judged by its coefficient of performance (COP), compressed air in Wh per normal cubic metre (Wh/Nm³), steam per tonne produced[6][6]Studies and reportsCertinergy, industrial energy performance indicators page (in French), utility indicatorsSee in sources ↓ and domestic hot water in kWh per cubic metre. On the production side, kWh per unit produced or the electricity used per line completes the picture. A whole-plant total, on the other hand, never tells you which utility is drifting.
How do you build a reliable EnPI in four steps?
Step 1: start from the energy review and significant energy uses
The energy review first identifies each significant energy use (SEU): a use that weighs heavily in consumption or offers strong potential for improvement.
The rule fits on one line: one EnPI per SEU, at least. A ratio worth keeping ticks three boxes:
- A relevant variable that can be measured.
- Data already available in the metering plan.
- Genuine sensitivity to the actions taken.
Step 2: qualify the data before calculating
Before turning it into an indicator, check for yourself what the history of a meter or a BMS point is worth. Here are the seven checks an expert runs before any calculation.
Expert tip For each point, what to check and the error the check prevents:
- Continuity of the history → gaps read as drops.
- Duplicate points → consumption counted twice.
- Meter never recalibrated → a drift mistaken for a trend.
- Silent disconnections → zeros recorded with no alarm.
- Archived time step (10 min, 1 h, 1 day) → an invisible night-time baseload.
- Actual physical scope → a meter that also supplies the neighbour.
- Main meter versus sub-meters → a sum that does not add up.
Skip these checks and you end up with 500 meters and nobody knowing which ones are reliable. Only actual consumption counts, not what a silent point suggests.
The resulting rule: a kWh/HDD that improves because a meter stopped is not a saving. Designing the measurement system belongs to the measurement plan.
Step 3: normalise the indicator for weather and activity
Heating depends on outdoor temperature, captured by heating degree days (HDD), which in the UK are usually published at a base of 15.5 °C for most buildings. Occupancy and production play that role for other uses. The method relies on a regression between consumption and each relevant variable. Take an illustrative example, unrelated to any real site: 420,000 kWh for 2,100 HDD, i.e. 200 kWh/HDD. The following winter shows 380,000 kWh for 1,750 HDD, i.e. about 217 kWh/HDD. Consumption falls, performance worsens: the winter was simply milder.
The scatter plot below places both winters against the same baseline line.
All seven months of the second winter sit above the line: each one signals a drift to explain, which the falling annual total concealed.
Step 4: set the energy baseline and the calculation frequency
The EnB is the quantified reference against which energy performance is compared. It is based on a representative year, with no works or prolonged shutdown, normalised for the same factors as the indicator. Our case studies show how an EnB set on a real site changes the reading of each ratio.
Once the EnB is set, the calculation frequency follows the purpose: real time to detect, daily or weekly to diagnose, monthly for management reporting.
How do you track EnPIs and turn an indicator into action?
Spreadsheet, EnMS software or BMS: which tracking tool?
For three monthly ratios, a spreadsheet is enough. Once an ISO 50001 programme starts, EnMS software structures it with the EnB, the action plan and audit evidence. Energy monitoring goes further by adding alerts. Then there is the BMS supervisor, which already trend-logs meters and setpoints: it is often the forgotten source.
Whatever the tracking tool, a dashboard looks at the past. An EMS calculates the indicators; it does not replace the person who acts.
From meter to setpoint, then to proof
People often believe the tool does the work. Yet a dashboard is only one link in five. The diagram below walks through the complete energy management chain, link by link, with the breaking point of each.
In the field, it is the fourth link that fails most often. The diagnosis is made and the action plan exists. Yet nobody has access to the BMS to change a setpoint or clear a manual override.
Dashboards then show the same drift every month. To improve efficiency, the indicator must lead to a changed setpoint, then to a measurement.
What Foobot does with your EnPIs on an existing BMS
Make meters reliable before the first calculation
With BMS advisory, we go through the history point by point: disconnections, data gaps, duplicates, invisible drifts and unexplained consumption peaks. Energy monitoring then rests on actual consumption. This clean-up is worth doing for its own sake. Many sites stop there, and rightly so.
Tackle at source the drifts the indicator reveals
A rising night-time baseload often tells the story of a VIP complaint, from a senior executive or a priority occupant, settled with a manual override that was never cleared.
It can also betray an AHU stuck in degraded mode. In our operations advisory work, we remove these manual overrides, restore consistent sequences and keep only the useful alarms.
Control HVAC and prove the savings with AI Control
When the building’s complexity justifies it, AI Control trains a digital twin on the site’s historical data. It then controls heating, ventilation and air conditioning (HVAC) every 15 minutes, 24 hours a day. At Valeo, the result reached −28.6% HVAC consumption, measured according to the IPMVP protocol[7][7]Standards and protocolsEfficiency Valuation Organization, “International Performance Measurement and Verification Protocol (IPMVP)”, options A to DSee in sources ↓. This control never starts before the BMS is reliable.
An indicator is only as good as the meters that feed it
Choose few indicators, normalise them for weather and activity, make the data reliable, then link each drift to a verified action: the method comes down to four moves. Without this tracking, no saving can be proven. The first move is to review your first three meters: continuity, scope, consistency with the main meter.
For weather normalisation, our article on baselines and degree days details the calculation. To calculate and track energy performance indicators, the free Foobot EMS takes over. If your meters cannot answer these questions yet, let’s talk.
FAQ: energy performance indicators
What is an energy performance indicator (EnPI)?
In ISO 50001 terms, it is a measure or unit of energy performance defined by the organisation, in the form of a metric, a ratio or a model[3][3]Standards and protocolsISO, “ISO 50001 — Energy management”, definition of an EnPISee in sources ↓. This definition remains theoretical: an indicator only becomes useful once it is normalised for its relevant variables and compared with a baseline.
Why put EnPIs in place in a company?
Putting EnPIs in place gives the company a common basis for decisions: where to invest, what to correct in operation, how to prove a result. Without them, the action plan rests on impressions and every claimed saving remains unverifiable.
How do you implement energy performance indicators?
Implementation follows four steps: start from significant energy uses, qualify the data, normalise for weather and activity, then set the baseline. Skipping the second step distorts all the others, even with a good tool.
How do you track EnPIs effectively?
Tracking indicators requires a named owner, a set frequency and an alert threshold. Real time detects; the monthly review decides. A deviation that nobody deals with after the alert has no value: it comes back in the next report.
How do you analyse EnPIs?
Analysing an indicator means comparing it with the EnB normalised for each relevant variable. If the gap persists, you trace it back to the equipment, then the zone, then the time slot that is drifting. A gap with no identified cause remains a hypothesis.
How often should they be tracked?
It depends on the purpose. To detect a drift, track continuously. To diagnose, daily or weekly. To report to management, monthly. A single time step serves all three purposes poorly.
What is the ISO 50001 standard?
It is the international standard that sets out the requirements for an energy management system. The current edition is ISO 50001:2018 with Amendment 1 of 2024, published in the UK by BSI as BS EN ISO 50001:2018+A1:2024[2][2]Regulation and standardsBSI Knowledge, “BS EN ISO 50001:2018+A1:2024 Energy management systems. Requirements with guidance for use”, current UK editionSee in sources ↓.
How many indicators should you track?
At least one per significant energy use. Beyond that, never more than the team can actually use each month. A few ratios read and discussed in committee are worth more than a full page of curves that nobody opens.
Sources and references
7 sources.- [1]GOV.UK, “How to comply with the Energy Savings Opportunity Scheme (ESOS) Phase 4”, Environment Agency guidance (qualification, compliance routes, ISO 50001 route)30/07/2026↑ 1 2
- [2]BSI Knowledge, “BS EN ISO 50001:2018+A1:2024 Energy management systems. Requirements with guidance for use”, current UK edition28/02/2025↑ 1 2 3
- [3]ISO, “ISO 50001 — Energy management”, definition of an EnPI↑ 1 2
- [4]ISO, “ISO 50006:2023 Energy management systems — Evaluating energy performance using energy performance indicators and energy baselines”, 2nd edition05/2023↑
- [7]Efficiency Valuation Organization, “International Performance Measurement and Verification Protocol (IPMVP)”, options A to D↑
- [5]CIBSE TM46 (2008), “Energy benchmarks”, General office benchmark cited in a UK government paper | https://assets.publishing.service.gov.uk/media/5a7e1520e5274a2e8ab45b9e/EED_regs_-_benchmark_heat_demand_paper_-_261114_.pdf ; Better Buildings Partnership, “2025 Real Estate Environmental Benchmarks” (REEB) | https://www.betterbuildingspartnership.co.uk/sites/default/files/media/attachment/REEB%20Benchmarks%202025.pdf ; Display Energy Certificates: Energy Performance of Buildings (England and Wales) Regulations 2012, SI 2012/3118, regulation 14↑
