UK Domestic Energy Benchmarks (NEED) — By Dwelling Type, EPC Band & Floor Area

These are the meter readings, not the energy rating. The National Energy Efficiency Data-Framework (NEED) matches actual gas and electricity meter data to housing records for the whole of Great Britain, so what you get is how much energy homes really used — not how much a model says they should.

This page publishes the 45 NEED benchmarks held in MasterBrain: mean annual consumption by dwelling type, by EPC band and by floor-area band, plus the mean floor area used to scale them. They are the default when you need a credible figure for a home whose actual bills you cannot see.

What NEED is

NEED is a Department for Energy Security and Net Zero statistical framework that links metered gas and electricity consumption to property attributes — dwelling type, age, floor area, EPC rating, installed measures — across the domestic stock of Great Britain. Because it starts from meter data rather than from a model, it is the closest thing available to ground truth for how British homes actually consume energy.

That distinction is the whole point of the dataset. An EPC tells you how a home should perform under standardised assumptions about occupancy, heating patterns and weather. NEED tells you what happened. The gap between the two is well documented and runs in both directions, which is why a footprint built on modelled ratings and one built on metered benchmarks can differ substantially for the same property.

Three cuts of the data are published here, and they answer different questions:

• By dwelling type — the right cut when you know what kind of home it is but nothing about its fabric.
• By EPC band — the right cut when you hold an official energy rating, which is the common case for a mortgage lender.
• By floor area — per square metre rather than per dwelling, which removes size from the comparison entirely.

By dwelling type

Mean metered consumption per dwelling per year. Aggregate rows are shown in bold.

By EPC band

Mean metered consumption per dwelling per year, by the property’s official energy rating. Note the ordering is not monotonic — see the warning above and the methodology note below.

By floor area

The same population normalised per square metre, which removes dwelling size from the comparison. Read this way the pattern is clean and steep: the smallest homes use roughly 2.6× the electricity and 1.7× the gas per square metre that the largest do, because heat loss scales with surface area while floor area scales faster.

The per-dwelling tables above are published on an average-sized home. This is the reference used to scale them:

Application — worked examples

Estimating a home’s energy use from its EPC band

The EPC benchmarks are published per dwelling, on an average-sized home. Where you know the floor area, scale them against the reference above:

consumption = EPC-band benchmark x (floor area / 97 m2)

A semi-detached house, EPC band D, 120 m²:

scale       = 120 / 97 = 1.2371
electricity = 3,328 kWh x 1.2371 =  4,117 kWh/yr
gas         = 11,500 kWh x 1.2371 = 14,227 kWh/yr

Convert those to emissions with the grid electricity factor for the reporting year and the DEFRA natural gas factor, and add well-to-tank if you are reporting Scope 3 Category 3 as well.

PCAF mortgages — Option 2a, data quality score 3

These benchmarks are the reference data behind PCAF Option 2a for financed emissions on a mortgage book: energy estimated from the property’s official energy label and floor area, which scores 3 on the PCAF data-quality scale. Every row in this dataset carries that note.

The financed-emissions chain runs:

property energy (above) -> property emissions (grid + gas factors)
    -> x attribution factor (outstanding loan / property value at origination)
    = financed emissions for that mortgage

Score 3 sits in the middle of the PCAF scale — better than a floor-area-only estimate (score 4) or a stock average (score 5), and weaker than actual metered consumption for the specific property (score 1–2). If you can obtain real meter data for part of the book, use it for those properties and these benchmarks for the remainder, then report the weighted average data-quality score across the portfolio.

Scaling by floor area is what earns score 3 rather than score 4. Applying the per-dwelling benchmark unscaled, when floor area is known, discards information PCAF expects you to use.

Calculators that use these factors

The worked examples above are automated in these GreenCalculus tools, which read this dataset live from the MasterBrain so results always reflect the current vintage:

• Heat pump vs gas boiler calculator — domestic heating options against these benchmarks.
• Personal carbon footprint calculator — a household footprint including home energy.
• Small business carbon footprint calculator — energy use for small premises.

Common reporting errors

• Treating EPC band as a consumption ranking. It is a modelled rating per square metre, not measured total energy. Band A averages more than bands B and C in this data, and band G less than band F.
• Forgetting to scale by floor area. The dwelling-type and EPC benchmarks are per average dwelling (97 m²). Applying them unscaled to a 200 m² house understates it by roughly half, and drops the PCAF data-quality score from 3 to 4.
• Mixing the per-dwelling and per-m² tables. The floor-area band table is already per square metre. Multiplying those figures by floor area again and then scaling is a double application.
• Assuming every home has gas. These are means across properties on the relevant meter. A home with no gas connection — electric heating, heat pump, oil, off-grid — is not described by a benchmark that assumes a gas supply, and roughly one in seven British homes is not on the gas grid.
• Applying GB benchmarks to Northern Ireland. NEED covers England, Wales and Scotland. Northern Ireland has a materially different heating mix, with heavy reliance on heating oil.
• Using a stale vintage across a time series. Consumption moves with weather, prices and efficiency measures. Apply the NEED edition contemporaneous with the reporting year rather than restating history on the newest figures.

Methodology, boundaries & uncertainty

What is measured. Mean annual metered gas and electricity consumption, weather-corrected by DESNZ, matched to property attributes. Consumption only — no emission factor is applied at this stage, so these figures are fuel-agnostic inputs rather than carbon results.

Geography. Great Britain: England, Wales and Scotland. Northern Ireland is outside the framework.

Means, not distributions. Every figure is a mean across a large population. Individual homes vary enormously around it, driven by occupancy, heating behaviour and equipment far more than by fabric alone. These benchmarks are appropriate for portfolios and for properties where nothing better exists; they are poor predictors for any single home.

Why the EPC ordering is not monotonic. EPC is modelled efficiency per square metre under standard assumptions; NEED is measured total consumption per dwelling. Larger homes tend to hold better ratings and still consume more in absolute terms, and the highest and lowest bands hold far fewer properties than the middle ones. Read the per-m² table when you want efficiency, and the per-dwelling tables when you want totals.

Gas coverage. Gas benchmarks describe properties with a gas meter. Where a home has no gas connection the gas figure does not apply and its heating energy will appear — or not appear — elsewhere.

Implementation & provenance chain

Every row traces to the published NEED tables. Nothing is hand-transcribed: the tables are generated from the same MasterBrain rows the REST endpoint serves, and the generator asserts that all 45 source keys render exactly once before the page is built.

Data access — REST API & CSV

The full 45-row dataset is available as a machine-readable REST endpoint and as a flat CSV download. Both serve exactly the rows shown above, versioned by MasterBrain edition and citable.

Citation guidance

If you use this dataset in a published tool, report or academic work, cite the DESNZ primary source. The GreenCalculus compilation reference is optional but appreciated.

Department for Energy Security and Net Zero (2025). National Energy Efficiency Data-Framework (NEED). London: DESNZ. Contains public sector information licensed under the Open Government Licence v3.0.

NEED primary citation

Cite this dataset (GreenCalculus compilation). A versioned, machine-readable snapshot of these factors is archived on Zenodo with a citable DOI:

Say, Jeremiah (2026). UK domestic energy benchmarks by dwelling type, EPC band & floor area (NEED) (machine-readable) (v2026.107). GreenCalculus. Zenodo. https://doi.org/10.5281/zenodo.21831188

GreenCalculus dataset DOI

Frequently asked questions

Version history