How the calculator works

The calculator estimates the size, cost, running-cost savings and carbon reduction of a heat pump, solar PV system, battery, or any combination, for a specific UK property. It is an estimation tool, not a quote, design or MCS survey. It doesn't account for listed-building constraints, planning considerations, grid-connection issues, or site-specific plumbing.

Three accuracy tiers run the same underlying model but with different inputs:

• Simple – postcode and property basics only.
• Building-specific – uses your EPC certificate if we can find one.
• Professional – uses your actual energy bills.

The tier describes what the estimate is based on, not how accurate it is. More of the household's own data means fewer assumptions, but the calculation is the same. See how accurate is this?

Every result comes out of four stages, in order:

1. Energy profile – how much electricity, heating and hot water the home uses in a year.
2. Sizing – what capacity of kit fits that demand and the roof/site.
3. Costs – install cost, running cost before vs after, payback.
4. Carbon – annual and lifetime emissions saved.

Rather than asking the user lots of questions, we look up a building archetype: a combination of property type (detached, semi, terraced, flat, bungalow) and construction era (pre-1930, 1930–70, 1970–90, 1990–2006, post-2006). Each archetype has a base electricity intensity and heating intensity, expressed in kWh per m² per year, derived from UK housing-stock research.

That base figure is then adjusted by:

• Floor area – your input, or archetype average for your bedroom count.
• Occupancy – ±15% per person vs a 2.4-person baseline for electricity; hot water scales more directly with people.
• Region – from postcode: heating demand, wind exposure, solar irradiance, regional performance gap.
• Wall construction – solid brick, cavity (un/insulated), modern, timber frame.
• Glazing – single, double, triple, low-E – weighted by the archetype's glazing ratio.
• Performance gap – a factor accounting for the well-documented tendency of homes to miss their theoretical performance.

Checked against metered gas use. When there is no EPC and no bills, the gas estimate is then multiplied by four factors – region, property type, age and bedrooms – fitted to the median gas use DESNZ records for homes like it (National Energy Efficiency Data-Framework, June 2026 edition (2024 gas use); England, Wales and Scotland). Without them the archetypes came out at about 0.7× the measured median in London and 1.3× in Wales, and flats at about 0.6×.

The bedroom factor is left out when you give the floor area, because most of what it corrects is the floor area guessed from bedrooms. Northern Ireland has no NEED data, so estimates there are not calibrated. The measured figures include gas used for cooking, which the calculator counts as heating, so the calibrated estimate is slightly high on that account.

If you supply energy bills, we skip the estimation entirely and use your annual kWh, splitting gas 75/25 between space heating and hot water (a standard ratio).

With an EPC, the gas estimate comes from metered use rather than the archetypes: the certificate's floor area × the median gas use per m² of houses in its band (DESNZ National Energy Efficiency Data-Framework, 2024) × three factors from the same data: size, region and property type. The age of the home only counts through its band. Electricity still comes from the archetypes, scaled by band.

Smaller homes use more gas per m², which is most of why flats use more per m² than houses. Measured use barely changes from band D to G, so a low rating alone doesn't raise the estimate much. NEED has no per-m² figures for flats, so their factor is fitted to the median gas use of flats. NEED's floor areas come from the Valuation Office rather than EPCs; if a certificate records a larger area, the estimate is that much higher.

Worked example

A 1930s semi-detached, 90 m², 3 bedrooms, 2 occupants, Manchester (NW), cavity uninsulated walls, double glazing, no EPC or bills.

Electricity:

Heating + hot water:

Numbers rounded for clarity. The live calculator carries full precision to the final step.

Heat pump. Peak heating demand is the annual heat delivered (gas for heating and hot water × boiler efficiency 0.82) ÷ 2,000 (approximate full-load hours over a UK heating season), floored at 3 kW, then oversized by 10% for the coldest days. For 380 UK heat pump homes with a year of metered heat, the median result was within 3% of the installer's design heat loss. When the estimate comes from an energy certificate for a home built before 1900, the heat is multiplied by 1.2 for sizing only: those homes used about a quarter less gas than their heat pumps later delivered, probably because they were kept cooler. This figure comes from 33 homes and will be revisited.

Solar PV. We size to the smaller of: electricity consumption ÷ specific yield (kWh/kWp), or available roof area ÷ 6 m²/kWp. Clamped 2–20 kWp. Specific yield depends on region and orientation (south = 1.00, north = 0.60).

Battery. With solar, sized to store roughly 50% of daily generation (capped at 40% of daily consumption). Without solar, 20% of daily consumption for time-of-use arbitrage. Clamped 5–20 kWh.

What the home already has. If you say the home already has solar panels, a battery, an electric car or a heat pump, nothing new is invented for it. When you choose your home's energy certificate and it records solar panels or a battery, they are ticked as already there, with the panels' size when the certificate gives one, and a note says so; untick anything that has gone since. Panels the certificate shows as not connected to the home's own meter, such as a block's shared panels, are left out. A heat pump, panels or battery the home has are not costed again. A battery next to existing panels is sized from their yearly generation, worked out with the same yield as above (south-facing, the best case) when you give their size in kWp; without a size it is sized as if there were no panels, and the results say so. Electric car charging, when you give it from a year's bill, is taken off the electricity that new solar and a battery are sized from, because most of it happens overnight when panels make nothing; without it nothing is taken off. An existing heat pump's electricity is already in a bill, but not in the estimates from the basics or an energy certificate, which leave it out, and the results say so. With their size given, the savings count the battery letting the home use more of the panels' output: the same rise from 40% to 70% of generation used as for a battery with new panels, less the 40% the home already uses, and never more than the electricity the home still buys or the battery can deliver in a year.

SCOP (seasonal performance: heat delivered per unit of electricity) is 2.8 for air source. This is the median from the Electrification of Heat trial's final report (Energy Systems Catapult, December 2024; 428 homes), including backup heaters and pumps (SPFH4). Ground source uses 3.3. This isn't from a field trial, because the trial had too few ground source heat pumps to publish a figure, so ground source results are less certain.

Unit rates

Install cost formulas

Grants: with a heat pump, the grant for the home's nation is taken off. In England and Wales that is the £7,500 Boiler Upgrade Scheme. In Scotland it is the £7,500 Home Energy Scotland grant, for owner-occupiers (it is higher in remote rural and island areas, which a postcode can't identify). Northern Ireland has no equivalent, so nothing is taken off. Without a postcode, the calculator assumes the BUS. All grants have eligibility rules, so check the scheme's own page.

Running cost

Current cost = kWh × unit rate (or your actual bills). Projected cost applies each technology in turn:

• Heat pump: gas cost → 0; add (gas kWh × boiler efficiency 0.82 ÷ SCOP) × electricity rate. Gas standing charge also saved. Boiler efficiency converts gas bought into heat delivered, which is what the heat pump has to provide. The figure is the real-world seasonal efficiency of a condensing boiler, from DESNZ's in-situ boiler trial.
• Solar PV: subtract generation × self-consumption rate (40% without battery, 70% with) × electricity rate.
• TOU variant – heat pump: heat-pump electricity priced at the blended off-peak rate; household electricity stays at the standard rate.
• TOU variant – battery: a portion of household electricity flows through the battery (charged off-peak, discharged at peak). That portion is repriced at the off-peak rate divided by round-trip efficiency (0.9), so storage losses are accounted for. Annual kWh shifted ≈ capacity × 0.85 × 1 cycle/day × 0.9 × 365, capped at the home's non-HP electricity use. If solar is also selected, cycles used for storing PV are subtracted first (they already save at the standard rate via the solar offset).

Emissions = kWh × carbon intensity. Current grid electricity is 0.131 kgCO₂e/kWh; natural gas is 0.182 kgCO₂e/kWh. Source: DESNZ/Defra GHG Conversion Factors 2026. The electricity figure covers generation only; grid losses would add about 0.013 kgCO₂e/kWh. Both factors are this year's. As the grid gets cleaner, a heat pump's lifetime saving is likely to be larger than shown.

Lifetime savings use a 25-year system life. The "trees planted" and "car miles" equivalences (÷22 kg/tree/year and ÷404 g/mile) are communication aids, not precise accounting.

If you supply actual bills and your heating costs look unusually low vs what the archetype predicts (below 85%), we assume you're heating selectively – rooms off, thermostat low – to save money. Running a standard projection off those numbers would flatter the heat pump artificially.

So we show two results:

• Current pattern – heat pump run the same way you heat now. Lower savings, same comfort.
• Full comfort – heat the whole home properly. Higher running cost than "current pattern" but a fairer like-for-like with a well-heated home.

Heat-pump size is held constant across both – only the run-hours change.

We haven't measured how closely each tier's estimates match what homes use after an installation, so the results don't give a ± figure. The tier label only says what the estimate is based on.

• Installation cost is shown as a range, from 85% to 125% of the central estimate. The range is the same for every tier, because quotes for the same job vary this much.
• Savings, payback and carbon are single central estimates. They are at least as uncertain as the energy use they're based on, and also depend on prices and how the home is heated.

For estimates from the basics alone (no EPC or bills) we can say how the gas estimate compares with metered use. Across 16,620,095 homes in England and Wales, grouped by region, type, age and bedrooms, the calibrated estimate is within 0.95–1.05× of its group's median use for half of homes (before calibration, 0.81–1.12×). But homes in the same group differ a lot: typically a quarter use less than 0.72× the median and a quarter more than 1.32×. Four facts about a home cannot narrow that. An EPC narrows it, and your bills remove most of it.

Internally, the model gives its energy-use estimate an uncertainty of 10–30%, depending on how many property details are known. This is widened by 30% for homes built before 1930, 15% for 1930–69 and 20% when the floor area is estimated. It's a judgement rather than a measured error, and it isn't shown in the results. Before any decision, the home needs a survey by an MCS-certified installer.

• Energy prices – Ofgem Price Cap
• Carbon intensities – DESNZ/Defra GHG Conversion Factors
• Heat pump field performance – Electrification of Heat demonstration: insights from heat pump performance data (Energy Systems Catapult, 2024)
• BUS grant – gov.uk/apply-boiler-upgrade-scheme
• Housing-stock archetypes – UK building performance research and EST publications
• Metered gas use (EPC bands and the calibration) – DESNZ National Energy Efficiency Data-Framework (NEED) consumption tables