A heat pump's SCOP is the single number that decides whether a quote looks good. It goes on the MCS certificate, into the running cost estimate, and into the conversation where a homeowner decides whether to spend fifteen thousand pounds. "SCOP 4.5" has become industry shorthand for "this is an efficient machine."
It is a real number, carefully measured, and it answers a question almost nobody is asking. It describes a heat pump running through a winter in Strasbourg, on a control curve most UK systems never use, and it says nothing about hot water or tariff.
On the label (55°C) 3.26 Measured on a Strasbourg winter Birmingham, typical UK controls 2.94 10% below the label Birmingham, fixed flow temperature 2.64 19% below the label Two quotes, same house £151 apart Annual running costs of £1,165 (45°C) and £1,316 (55°C), both built from label figures Two real outcomes £737 apart Annual running costs of £843 (28% below its quote) and £1,580 (20% above its quote), once design, controls and tariff are counted
The heat pump is identical in every case. What changes is the climate, the design flow temperature, how the controls are set up and the tariff. Two quotes for this house sit £151 apart. Once those four things are counted, the two installs cost £737 a year apart to run.
Same January, two different winters. Strasbourg's design day is −10°C; Edinburgh's is −3.2°C, and it spends far longer just a little cold.
This piece works through the four influences that move a real running cost away from the label, one at a time, and then adds them up:
- Climate — a British winter is milder than the rating winter. This works in your favour.
- Compensation curve — the label assumes the flow temperature falls steeply in mild weather. Most UK systems fall less, or not at all. This works against you, and more so at 55°C.
- Hot water — quotes often give hot water the 55°C space heating figure. Cylinders do worse. This works against you.
- Tariff — when the heat is bought matters as much as how efficiently it is made. This can work strongly in your favour.
Because they point in different directions, they hide each other. And they sit on top of the uncertainty in the heat loss calculation itself, which we covered in From 14 kW to 5 kW: an oversized heat pump cycles more and runs further from its efficient range, which compounds every effect below.
What the number on the label actually is
Every SCOP a manufacturer publishes, and every SCOP on an MCS certificate, comes out of the EN 14825 bin method. For products listed on the MCS Product Directory it is produced by the MCS 026 calculator, which is mandatory under the MCS 007 product standard.
The method is sound and it is well documented. The heating season is split into one-degree temperature bins, each with a fixed number of hours. The building's heat demand falls on a straight line from the design load at the design temperature to zero at 16°C. The unit's measured COP is interpolated between its declared test points. Multiply through, sum, and divide total heat by total electricity.
Split the season into bins Hours at each 1°C of outdoor temperature, fixed by the reference climate (Strasbourg for "average"). Work out the heat needed Demand falls in a straight line from the design load at −10°C to zero at 16°C. Look up the COP The unit's tested COP, interpolated between its declared test points, for each bin. Add it all up Total heat delivered divided by total electricity used gives the SCOP. SCOP = Σ heat delivered ÷ Σ electricity used, over the Strasbourg season
The question is which heating season. EN 14825 defines three reference seasons, and they are specific places: Strasbourg for "average", Helsinki for "colder", Athens for "warmer". The average season has a design temperature of −10°C and runs 4,910 hours. Every one of the four application worksheets in MCS 026 has the climate hardcoded to average.
So the SCOP on the label is a Strasbourg figure. It exists so that a Vaillant can be compared with a Daikin on identical terms, which it does well. It was never meant to predict what a particular house in Scotland will achieve, and it does not.
Influence 1: climate — a British winter is not a Strasbourg winter
Direction: in your favour. The label is pessimistic on climate alone.
Start with the obvious difference. MIS 3005-D Table 2 gives the UK design temperatures that every heat loss calculation in this country uses: −1.5°C for Belfast and Cardiff, −1.7°C for London, −3.2°C for Birmingham and Edinburgh. Strasbourg's −10°C is between seven and nine kelvin colder than anywhere we design for.
The shapes are different in a way that matters. Strasbourg has a cold tail that British weather simply does not produce: 73 hours a year at −7°C or below, against four in Edinburgh. But the British seasons are much longer. Edinburgh spends 7,624 hours a year below 16°C against Strasbourg's 4,910, and most of that extra time sits between 5°C and 15°C.
A seasonal COP is weighted by where the heat is actually delivered, because a cold hour carries far more kilowatt-hours than a mild one. That is the view that explains everything else.
Weighted by heat, the Strasbourg season delivers at a mean outdoor temperature of 2.7°C. Edinburgh delivers at 5.7°C, and Cardiff at 6.7°C. Three to four kelvin warmer, at the temperatures where the efficiency is earned.
| Design temp | Hours below 16°C | Equivalent full-load hours | Heat-weighted mean | |
|---|---|---|---|---|
| Strasbourg (the rating season) | −10.0°C | 4,910 | 2,066 | 2.7°C |
| Edinburgh | −3.2°C | 7,624 | 3,094 | 5.7°C |
| Birmingham | −3.2°C | 7,053 | 2,809 | 5.9°C |
| Manchester | −2.7°C | 7,267 | 2,956 | 6.0°C |
| London | −1.7°C | 6,532 | 2,671 | 6.3°C |
| Belfast | −1.5°C | 7,672 | 3,137 | 6.7°C |
| Cardiff | −1.5°C | 6,920 | 2,741 | 6.7°C |
This is the first surprise, and it runs against the intuition most people have. The rating climate is harsher than ours. On climate grounds alone, a heat pump in Britain should beat its label.
By how much depends on the region and on the design flow temperature. Running the same bin method, with the same COP shape, on the rating season and then on each UK region gives this:
Between 7% and 17%. The spread between UK regions is modest, under nine points at any flow temperature, and Scotland sits at the bottom of it as you would expect. The step from Strasbourg to anywhere in Britain is the larger one.
If that were the whole story, the label would simply be pessimistic and we could all relax. It is not the whole story.
Influence 2: the compensation curve — "SCOP at 55°C" was not measured at 55°C
Direction: against you, and much more so for systems designed at 55°C.
Here is the part that catches people, including people who have been specifying heat pumps for years.
EN 14825 tests the unit on a weather compensation curve, so each test point has its own water outlet temperature, falling as the outdoor air warms. MCS 026 publishes the curve.
Read the top line again. "SCOP at 55°C" means the unit was running at 52°C when it was −7°C outside, 42°C at +2°C, 36°C at +7°C and 30°C at +12°C. . The 55°C in the label is the design condition only, and the design condition contributes almost none of the annual heat.
This is where a lot of design goes wrong. A "SCOP at 55°C" is often read as the efficiency of a system that runs at 55°C, and so as a safe figure for radiators that need 55°C on the coldest day. It is not. It is the efficiency of a system whose flow falls from 52°C to 30°C as the weather warms. A system held at or near 55°C all winter does markedly worse, and the figure on the quote says nothing about it. That is common: in Nesta's audits of social housing heat pumps in Bristol, weather compensation was disabled on 36 of the 39 systems checked, 92%.
The 35°C figure is different in kind. Its curve only swings 11 kelvin, from 35°C down to 24°C, because there is less room to fall. That asymmetry turns out to matter.
The consequence is blunt. A system running a flat flow temperature, which describes a great many retrofits with no weather compensation or with compensation left at the installer default, is not operating anywhere near the conditions its SCOP was measured at. It will not achieve that SCOP, in Strasbourg or anywhere else.
Influences 1 and 2 together: one unit, one house, three answers
Take a Daikin Altherma 3 M 8 kW (EDLA08EV3). Its UK brochure declares SCOP 4.56 at 35°C and 3.32 at 55°C for the average climate. Put it in a house in Edinburgh with a 7 kW design heat loss, which gives 21,655 kWh of space heating a year on the Edinburgh season.
At a 45°C design flow, the declared 3.94 is almost exactly right, and it is right for the wrong reasons. The milder Scottish climate is worth about 9%. A realistic UK compensation curve, rather than the standard's steeper one, costs about 8%. The two cancel to within a percent, and the label lands on the answer by luck.
Move either way from there and the cancellation breaks down. At 35°C the climate gain survives and the system beats its label, reaching 5.12 against a declared 4.56. At 55°C the compensation penalty wins, and 3.32 on the label becomes 2.98 in the house. Take weather compensation away entirely and 55°C declared becomes 2.68, which is 19% below the label.
In money, at 24.5p per kWh:
A £300 annual gap between the best and worst case on identical hardware. At a 55°C design flow the same comparison runs from £1,597 as declared to £1,979 with no compensation, a £382 gap, every year, for twenty years.
Influences 1 and 2 together: the same house in four UK climates
One house in one city shows the mechanism. To see whether it holds across the country, we took the reference house from our December article on seasonal SCOP: a 1960s three-bedroom semi of around 110 m², with a filled cavity, 270 mm of loft insulation and a room in the roof. The house is in Cardiff, where its design heat load is 4.60 kW at the local design temperature of −1.5°C. We kept the fabric identical and moved the house to four more regions chosen to span the UK:
- London, the mild south-east
- Birmingham, inland, with one of the coldest design temperatures in England
- Edinburgh, the longest heating season of the four
- Belfast, maritime, with a mild design day and a long, cool season
- Cardiff, where the house is, the mildest heating season of the five
In each region the design load is rescaled to the local design temperature, from 4.60 kW in Belfast and Cardiff to 4.95 kW in Birmingham and Edinburgh, and the annual space heat comes from the MCS 031 degree days for that area, from 9,011 kWh in Cardiff to 12,652 kWh in Edinburgh. The heat pump is a Daikin Altherma 3 M 6 kW, declared at 4.47 at 35°C and 3.26 at 55°C, so 3.87 at 45°C. Hot water is 4,393 kWh a year for three bedrooms, including cylinder losses and a weekly pasteurisation cycle. Electricity is the October to December 2026 Ofgem cap, 26.32p/kWh. Standing charges are left out because they are the same in every case.
For each region we compare the quote, built the usual way from label figures, against three realistic outcomes:
- Best: the compensation curve set as steeply as the one the label was measured on
- Typical: the 12 K swing we measured on monitored UK systems
- Poor: a fixed flow temperature with no weather compensation
35°C design flow 45°C design flow 55°C design flow Column height and figure: typical UK controls. Hover or tap a column for every figure.
The pattern is the same everywhere.
At 45°C the quote is close. In all five regions it lands within £15 to £39 of the typical outcome, inside the realistic range. The milder climate and the shallower UK compensation curve cancel, as they did for the Scottish house above. The small remaining gap is hot water, which we come to below.
At 55°C the quote is optimistic in every region. The typical outcome costs £104 to £157 a year more than the quote, about 10 to 11%. With a fixed flow temperature it costs £195 to £283 more, 18 to 21%. The quote sits at the bottom of the realistic range, which is the only place a system can reach if its compensation curve is set as steeply as the one the label was tested on.
Grey: the label figure. Green: at or above the label. Red: more than 10% below it.35°C design flow 45°C design flow 55°C design flow Region Label Best Typical Poor Label Best Typical Poor Label Best Typical Poor London 4.47 5.09 5.12 4.21 3.87 4.26 3.93 3.43 3.26 3.51 2.95 2.66 Birmingham 4.47 5.05 5.08 4.15 3.87 4.25 3.90 3.39 3.26 3.52 2.94 2.64 Edinburgh 4.47 4.99 5.02 4.13 3.87 4.21 3.87 3.38 3.26 3.48 2.92 2.63 Belfast 4.47 5.20 5.24 4.27 3.87 4.34 3.99 3.46 3.26 3.58 2.99 2.69 Cardiff 4.47 5.22 5.25 4.27 3.87 4.36 4.00 3.46 3.26 3.59 2.99 2.69
The climate correction differs by region, and Belfast and London gain most from it. What it cannot do at 55°C is make up for the compensation curve. The label's curve drops the flow 25 K across the season; a typical UK system drops it about 12 K. At 45°C the label's curve swings 17 K against the same 12 K, and the climate gain covers the difference. At 55°C it does not.
What it means for keeping the radiators
A low-impact retrofit keeps some or all of the existing radiators and designs the system to run hotter, typically at 55°C. It avoids the cost and disruption of replacing emitters and pipework, and for many households it is the right call. It is a trade: lower capital cost against a lower efficiency for the life of the system.
That trade needs pricing honestly, and the label prices it too cheaply.
A 55°C design with good compensation sits at a SCOP of about 3. That is roughly where a heat pump breaks even with an efficient gas boiler on a standard electricity tariff: Nesta puts the break-even at about 3.03. With poor controls it slips to 2.6 or 2.7, below that line, and in the coldest months, when the flow is highest, it falls further still. There is very little margin. At 35°C the same unit runs at a SCOP of 4 to 5 even with poor controls, which is the efficiency that makes flexible tariffs and thermal storage pay. Low flow temperatures are routinely undersold.
A quote built from label figures says keeping the radiators costs £114 to £160 a year in running cost. With typical UK controls it costs £199 to £278, roughly three quarters more. The gap is £90 to £120 a year, roughly £1,900 to £2,400 over a twenty-year life at today's prices. That is the figure to weigh against the radiators you avoid replacing, and a quote built from label figures leaves it out. If the compensation is then left off and the flow sits at 55°C all winter, the gap against a well-run 45°C system is £290 to £404 a year.
The same numbers say where to spend if you do keep the radiators. A 55°C system depends on its compensation curve for every point of efficiency it has. On the semi in Birmingham, the difference between a fixed flow and a curve set as steeply as the label's is £296 a year. Careful commissioning of the curve, and checking it after the first winter, is worth more to a low-impact system than to any other.
Influence 3: hot water — the assumption that compounds it
Direction: against you, every time. It never cancels anything.
The examples so far are calculations. This one is a real quote.
It is for a real family home in Scotland, with a design heat loss of around 6.5 kW and a 7 kW monobloc designed for 45°C. It was produced on one of the most widely used heat pump design tools in the UK, by an installer we regard as an example of good practice. We are not naming either, because the point is a default in the tool and the way the industry uses it, which reaches far beyond one installer.
The space heating figure in the quote is 3.91 at 45°C, from the manufacturer's ErP data, so it is a Strasbourg figure. For this house, the Scottish climate is worth about 9% and a typical UK compensation curve costs about 8%. The quote's space heating is right, for the same offsetting reasons as before. The quote also says weather compensation will make the system more efficient than 3.91. That counts compensation twice: the 3.91 already assumes it.
Hot water is where the tool's default goes wrong. It applies 339% to hot water, which is the manufacturer's 55°C space heating SCOP, and applies the same figure to the weekly 60°C pasteurisation cycle. A space heating SCOP at 55°C earns most of its efficiency in mild weather, with the flow temperature well below 55°C. A cylinder charge gets none of that. It runs at a fixed high temperature all year, and each charge ends with the heat pump working against a rising return temperature.
Across 183 monitored UK systems, the median hot water COP is 3.09 against a median declared 55°C SCOP of 3.40, so hot water runs at about 0.9 of the label figure the tool uses. The same quote carries an MCS column that puts hot water at 170%, half the figure the headline running cost was built on.
At the median, the tool's optimism costs this house £39 a year on a flat rate. That sounds small, and on its own it is. Two things make it matter. It always points the same way, so it never cancels anything; it adds to whatever else a quote has missed. And it scales with the price of the electricity it is applied to.
Influence 4: tariff — when you buy the heat
Direction: can be strongly in your favour, if the controls use it.
The same real example shows it. This household's hot water runs entirely overnight, on an off-peak rate of about 3.5p/kWh on a fixed time-of-use tariff. At that price the whole spread of assumptions, from the tool's 339% to the MCS 170%, is worth £45 a year, and the tool's optimism is worth £5. On a flat rate of about 26p the same spread is £330, and hot water alone costs £331 to £661 a year. For this house the tariff and the controls that put hot water into the cheap window decide more than any efficiency assumption does.
Two caveats. Overnight air is colder, so an overnight cylinder charge runs at a slightly lower COP than the median; at 3.5p that barely registers. And time-of-use tariffs usually charge more than a flat rate outside the cheap window, so the whole-bill effect depends on how much of the space heating can follow the hot water into it. Our December article works through the whole bill for three common tariffs.
How the assumptions stack
This is the point of the whole piece. Each influence on its own looks modest, and they do not all point the same way, which is why they are hard to see. Added up for two ordinary installs on the same Birmingham semi, with the same heat pump, they look like this:
Green ↓ cheaper than quoted, red ↑ dearer. Choosing 55°C over 45°C is already in Install B's quote (+£151); with typical controls it costs +£262.Influence, against the quote Install A
45°C, curve set, overnightInstall B
55°C, fixed flow, flat rate1. ClimateUK winter instead of Strasbourg −£72 −£71 2. Compensation curvehow far the flow temperature falls in mild weather £0 +£296 3. Hot watermeasured efficiency, not the 55°C label +£39 +£39 4. Tariffhot water overnight at an illustrative 7.0p −£289 £0 Net effect a year −£322 +£264
Put them together for two installs on the same semi in Birmingham. Install A upgrades the radiators for 45°C, has its compensation curve commissioned properly and heats its hot water overnight. Install B keeps the radiators at 55°C, runs a fixed flow temperature and stays on a flat rate.
Install A Radiators upgraded for 45 °C, compensation set up, hot water overnight Install B Radiators kept, 55 °C, fixed flow, flat-rate electricity
The two quotes differ by £151 a year. The two outcomes differ by £737.
Install A comes in £322 under its quote. Install B comes in £264 over its quote, 20% more than the customer was told. Install B is also the cheaper install to buy, which makes it the natural choice for the households most sensitive to running cost. It is exactly where a quote built from label figures is least reliable, and nobody involved has to do anything wrong for it to happen: the tool used the manufacturer's figures as published.
And this sits on top of the design uncertainty we described in From 14 kW to 5 kW. Uncertain heat loss inputs push designers towards larger heat pumps, and an oversized unit cycles more and works further from its best operating point, so the real gap is often wider still. Run the same heat pump in two different houses, with different design, controls and tariff, and you can get vastly different results from identical hardware.
This is one of the quieter ways heat pumps get a reputation for being expensive to run. The system works. The quote described a different system, one with a steeper compensation curve, hot water that charges as efficiently as space heating does in April, and a winter in Strasbourg.
Why this has gone unremarked
Because the main effects point in opposite directions, and most UK systems are designed near 45°C to 50°C, where they roughly cancel. A Strasbourg figure quoted at a British house with a different control strategy gives an answer that is accidentally close. Close enough that nothing looked wrong.
The cancellation fails exactly where the stakes are highest. A low-temperature system with good controls is undersold. A high-temperature retrofit with weak controls on a flat rate is oversold, and that is the case where a disappointed customer does the most damage to confidence in the technology.
The headline SCOP also carries no information about commissioning or tariff, which are among the largest real determinants of what a system costs to run. A unit is certified once. A compensation curve and a hot water schedule are set, or not set, on every installation.
What follows from this
Anyone reading a SCOP Use it to compare products. Ask which climate, curve, hot water figure and tariff any cost estimate assumes. Homeowners Keeping radiators can be right, but costs about three quarters more to run than a label quote shows. Insist the curve is commissioned and checked. Installers The compensation curve and hot water schedule are design decisions, each worth more than the gap between two brands. Software and modellers Don't give hot water the 55°C space heating SCOP. Correct the label to the region with the same bin method. MCS Publish a UK climate column, or the declared COP test points, so seasonal figures can be site specific.
In more detail:
For anyone reading a SCOP figure. Treat it as a comparison between products, which is what it was built for. Ask which climate, which flow temperature profile, which hot water efficiency and which tariff any running cost estimate was built on.
For homeowners considering a low-impact retrofit. Keeping radiators can be the right decision. Price it with realistic controls: on our reference semi the running cost of 55°C over 45°C is around three quarters higher than a label-based quote shows. If you keep the radiators, the compensation curve matters more to you than to anyone, so ask how it will be commissioned and checked.
For installers. The compensation curve and the hot water schedule are part of the design. On the examples above, each is worth more than the difference between two manufacturers' machines. A quote that promises the declared SCOP is implicitly promising a commissioned curve; a quote that prices hot water at the 55°C space heating SCOP is promising more than monitored systems deliver. Both promises should be explicit.
For software and anyone modelling this. Hot water should not take the 55°C space heating SCOP. A factor calibrated on monitored systems, or the manufacturer's own water heating efficiency where it is declared, is closer to what cylinders achieve. The climate correction is not difficult either: run the same bin method twice, with the same COP shape, once on the EN 14825 reference season and once on regional bins at the property's own design temperature. The ratio is the correction, and the unknown absolute COP curve cancels out of it. We have done this for all six UK design regions.
For MCS. Climate is already an input to MCS 026. The calculator supports all three reference seasons and hardcodes the average. Publishing a UK column, or publishing the declared COP test points so that anyone can run the method on UK weather, would cost very little and would close a gap between the certificate and the kitchen. The per-product capacity and COP at the four EN 14825 test points are not published anywhere central; manufacturers submit them into MCS 026 and only the outputs come out. That is the single piece of data standing between the industry and honest, site-specific seasonal figures.
Checking any of this
Nothing above requires a paywalled standard.
The average-season bin table is in Commission Regulation (EU) No 813/2013, Annex III, Table 5, free on legislation.gov.uk. The equivalent full-load hours per climate are in Regulation 811/2013, Annex VII. The reference compensation curve and the warmer and colder bin tables are in the MCS 026 V2 calculator, which MCS published for consultation in March 2025 and which anyone can download. The two reference sheets in that workbook are hidden by default, which is probably why the flow temperature curve is not better known.
One useful check if you transcribe the bin table yourself. The regulation publishes the equivalent full-load hours separately, at 2,066 for the average season, and the figure is also recoverable from the bin table through the load line. If your transcription does not reproduce 2,066, you have a typo.
The UK bins are ten years of ERA5 reanalysis, 2015 to 2024, from the Open-Meteo historical archive, taken at one MIS 3005-D Table 2 station per design region, at the column A (99th percentile) design temperature. Annual space heat for the reference semi uses the MCS 031 degree days for postcode areas E, B, EH and BT. The hot water factor of 0.9 is calibrated on 183 monitored systems from HeatpumpMonitor.org, as described in our December article. The 7.0p overnight rate is illustrative; the 3.5p rate is the household's own tariff; check current rates before relying on any figure here.
HeatLens produces MCS-compliant heat loss calculations and seasonal performance estimates for UK heat pump retrofits. The correction described here is computed for all six UK design regions. It will be reported alongside the manufacturer's declared figure, which stays on every report because it is the figure compliance paperwork quotes.