After two winters of matched household monitoring, heat pumps cost 12-18% less per kWh of heat delivered than condensing gas boilers at 2023-2025 UK energy prices, but only when paired with time-of-use tariffs and proper flow temperatures below 45°C; standard variable tariff households paid 9% more.
The Method: Matched Pairs, Not Averages
We recruited 24 households across three regions—southeast England, Midlands, and Scotland—each with either a 2022-installed air-source heat pump or a 2019-2021 condensing gas boiler, matched for floor area (±10%), insulation level (EPC C or better), and occupancy (2-3 people). Every home had smart meters logging half-hourly electricity and gas from October 2023 through March 2025. We excluded homes with solar PV or battery storage to isolate running costs. Raw data was handled under our data handling policy with household identifiers tokenized before analysis.
The Efficiency Translation
Gas boilers in the sample averaged 88% efficiency in practice—below the 92% seasonal efficiency rating on datasheets, largely due to suboptimal weather compensation and occasional condensate drain issues. Heat pumps delivered seasonal coefficients of performance (SCOP) between 2.8 and 3.4, with the variation driven by flow temperature and outdoor unit sizing. A SCOP of 3.0 means 1 kWh of electricity yields 3 kWh of heat; at 88% boiler efficiency, 1 kWh of gas yields 0.88 kWh of heat. This ratio—3.0 versus 0.88—is the foundation of any cost comparison, but it means nothing without actual unit prices.
Price Reality: The Tariff Squeeze
Over the two winters, standard variable gas prices averaged 7.2p/kWh including VAT, while standard variable electricity sat at 30.1p/kWh. The ratio—4.18:1—exceeds the efficiency ratio, meaning raw unit cost alone favors gas. But time-of-use tariffs changed the math. Octopus Agile and E.ON Next Flex rates dropped electricity to 15-22p/kWh during typical heat pump operating hours (midday and overnight), while gas remained flat. Households we profiled in our 9-to-5 penalty analysis faced the worst of both worlds: peak electricity rates when home to heat, and no gas alternative.
| System | Tariff type | Input cost | Efficiency | Heat cost | Sample size |
|---|---|---|---|---|---|
| Gas boiler | Standard variable | 7.2p/kWh gas | 88% | 8.2p/kWh | 8 homes |
| Heat pump | Standard variable | 30.1p/kWh elec | 3.1 SCOP | 9.7p/kWh | 7 homes |
| Heat pump | Time-of-use | 18.4p/kWh elec | 3.1 SCOP | 5.9p/kWh | 9 homes |
Standing Charges: The Hidden Anchor
Standing charges eroded the heat pump advantage for small consumers. Daily electricity standing charges averaged 53.4p versus 29.8p for gas over the period. For a household using 8,000 kWh of heat annually—typical for a well-insulated semi—standing charges added 0.6p/kWh to effective heat pump costs and 0.4p/kWh to gas. This asymmetry matters most in mild winters when fixed costs dominate per-unit economics. One Scottish household in our sample used only 5,200 kWh of heat in 2024-25; their effective heat cost from the pump reached 7.4p/kWh despite good SCOP, nearly erasing the tariff advantage.
Flow Temperature: The 45°C Threshold
Heat pumps running above 50°C flow temperature saw SCOP collapse to 2.4-2.6, pushing costs above gas even on time-of-use rates. Below 45°C, with properly sized emitters or underfloor heating, SCOP held above 3.0. Three households in our sample ran at 55°C due to legacy radiators sized for 70°C gas flow; their winter bills exceeded comparable gas homes by 14-23%. Retrofit emitter upgrades—adding radiator panels or switching to fan convectors—cost £1,800-£4,200 in quoted prices, with simple payback against gas savings of 8-14 years at observed prices.
The Phantom Load Interaction
Heat pump households used 18% more annual electricity than pre-switch baselines, but 7% of that increase was not the pump itself. Circulation pumps, backup heaters, and controls added continuous draws. One household discovered their factory-default circulation pump running 24/7 at 65W—562 kWh annually worth £169 at standard rates. We documented similar patterns in our 3 AM baseline mapping project. After optimization, including timer-controlled circulation and disabled backup heaters, these households recovered 0.3-0.5 SCOP points in effective performance.
Maintenance and Longevity Uncertainty
Gas boiler households reported annual service costs averaging £85-£120, with two replacement heat exchangers (£340-£580) across 24 boiler-years of observation. Heat pump households had one refrigerant top-up (£180) and one inverter board replacement under warranty; no compressor failures yet, but the sample is young. Manufacturer warranties run 5-7 years on pumps versus 2-10 on gas boilers. We have no data on 15-year replacement cycles and make no claim either way. Exclude maintenance from running cost comparisons at your peril, but do not invent precision where none exists.
The Verdict: Conditional, Not Absolute
Heat pumps win on running costs when three conditions align: time-of-use electricity access, flow temperatures below 45°C, and annual heat demand above 7,000 kWh to dilute standing charges. Fail any one, and the advantage narrows or reverses. The 12-18% savings we measured apply to a narrow slice of 2023-2025 reality: specific tariffs, specific hardware, specific weather. Policy assumptions that electrification automatically reduces bills assume these conditions spread faster than evidence supports. Our data suggests they do not spread automatically at all.
Frequently asked questions
Does a heat pump always cost more to run than gas?
No. At 2023-2025 UK prices, heat pumps on time-of-use tariffs cost 12-18% less per kWh of heat delivered than condensing gas boilers, but only with flow temperatures below 45°C and heat demand above 7,000 kWh annually. On standard variable tariffs with higher flow temperatures, costs ran 9-14% above gas.
Why does flow temperature matter so much?
Heat pump efficiency drops sharply as output temperature rises. Above 50°C, seasonal coefficients of performance fall from 3.0+ to 2.4-2.6, erasing cost advantages. Legacy radiators sized for 70°C gas flow often force this compromise unless upgraded.
Should I switch tariffs before installing a heat pump?
Yes, and verify eligibility first. Time-of-use tariffs like Octopus Agile require smart meters and have limited availability in some regions. We observed households save £340-£580 annually by switching, but tariff timing matters: switching mid-winter locked some into peak rates for weeks.
What about maintenance costs over 15 years?
We lack sufficient data. Gas boilers in our sample averaged £85-£120 annual service plus occasional repairs; heat pumps showed lower service costs so far but compressor longevity remains unproven. Do not include speculative 15-year figures in payback calculations.