When I signed on the dotted line for a £22,000 air source heat pump installation in July 2023, I thought I was doing everything right. I did my research. I checked credentials. I picked an installer that, at the time, was listed on Heat Geek’s own website as a certified/verified company. I wanted to do my bit for the planet and secure a future-proof heating system for my home.
Instead, I’ve been left with a cold house, spiralling bills, a drawn-out Section 75 battle with my credit card provider and an education in just how unaccountable this industry can be when things go wrong.
This is the story of my installation with EPC Improvements Ltd, why it failed and why homeowners across the UK need to start asking much harder questions before signing up to a heat pump.
The Installer I Trusted
I chose EPC Improvements Ltd because they were visible, endorsed and, at the time, listed as a Heat Geek certified installer. That felt like a safeguard. Surely if you appear on a platform that preaches technical quality and consumer trust, you’re competent?
The system they sold me was a Samsung AE120BXYDEG, 12 kW, for just over £22,000 (including the £5,000 Boiler Upgrade Scheme grant). That’s not pocket change. It’s the kind of money families sink into a kitchen extension or a car. You expect it to work.
From day one, it didn’t.
A System That Never Delivered
The signs were immediate. Rooms that never warmed up. Radiators lukewarm even when the heat pump was supposedly running flat out. Bills climbing instead of falling.
EPC’s response? “It’s working at 350% efficiency.”
But the numbers don’t lie. A year’s worth of data showed me what was really happening: approx. 19,703 kWh heat out from about 8,850 kWh in. That’s a seasonal COP of around 2.2, almost half of what was promised and far below what’s needed for cost-effective low-carbon heating.
The causes were glaring once independent eyes looked at it:
- Undersized radiators that physically couldn’t deliver the heat.
- A badly arranged buffer tank that dumped 8-10C of temperature before water even reached the radiators.
- Pumps that weren’t fit for heat pumps, one of them installed in the wrong place and literally pumping back into the buffer.
- Glycol dosed into the system, choking flow rates further.
- Sloppy commissioning and controls left on default, with no clear optimisation for the property.
This wasn’t cutting-edge renewable technology. It was bodging with expensive parts, and it’s not hard to understand why many families simply give up long before this point.
Living With the Consequences
Imagine paying £22,000 to make your home greener and more comfortable, and ending up colder than you were with your old boiler.
Imagine trying to tell your installer something’s wrong and being told it’s “all in spec.”
For months, we lived with a system that couldn’t heat the house. We plugged in fan heaters to stop freezing corners. One radiator even tore off the wall and caused an injury because of poor fixing. EPC brushed it off.
When the Industry Looks Away
Here’s what really shocked me: there’s no clear route to redress when this happens.
I went to MCS. They sent polite acknowledgements but no practical path forward. I tried RECC. I tried Trading Standards. I tried my home legal cover, but denied because the contract value was “too high.”
Even Heat Geek, whose badge EPC had traded on, initially told me they couldn’t help unless it was an install going through their platform. So much for consumer safety nets. Fine print everywhere.
I paid £800 out of pocket for an independent assessment from Andrew Millward, a Heat Geek Elite installer. Only then, with his technical report in hand, did Heat Geek HQ start engaging. By then, I’d been living with this mess for a year.
Section 75: A Lifeline With Strings
My only real weapon has been Section 75 of the Consumer Credit Act… the right to claim against your card provider for breach of contract.
But even here, the system isn’t built for heat pumps. Card providers want “2-3 comparable quotes” for remedial work. That works if you’re replacing a bathroom. It doesn’t when you need bespoke hydronic redesign.
I prepared a detailed Statement of Works with two options:
- Option 1: minimal remedial work (flush system, relocate pumps, fix/remove buffer, resize only undersized rads).
- Option 2: full overhaul (replace every emitter, overhaul hydraulics, possibly replace the heat pump).
Despite sending identical evidence packs to three firms, only one (Sun-lite Group) engaged constructively. They, alongside Heat Geek HQ, have been willing to talk about a realistic repair path.
But my card provider still insists on “quotes.” Installers are understandably reluctant to provide them without site visits, which the card provider won’t fund. So I remain stuck in limbo.
What My Case Exposes
This isn’t just my story. It’s a snapshot of an industry problem.
- Accreditation means very little. EPC were “certified” when I hired them. I found the quality of their work to be dreadful. The certification didn’t protect me.
- There’s no consumer escalation. MCS and RECC don’t offer practical remedies. Heat Geek’s map once listed dozens of installers, but course completion clearly isn’t the same as competence.
- Servicing is a Wild West. My annual “service” was basically an invoice. No checks, no documentation, no refrigerant testing. Unlike gas boilers, there’s no statutory service standard.
- Remedial work is toxic. Many competent engineers avoid it. It’s risky, messy, time-consuming. Everyone talks quality on social media, but few step up when a family’s freezing in a badly designed house.
- Finance processes don’t fit. Section 75 assumes like-for-like quotes. Heat pumps don’t work that way. Without independent, recognised assessors, consumers are stranded.
The Human Cost
This isn’t just about kilowatts and flow rates. It’s about trust.
I thought I was making a responsible choice for my family and the planet. Instead, I’ve spent two years fighting, documenting, emailing, chasing, all while paying for plug-in heaters to do the job a £22,000 heat pump couldn’t.
The emotional toll has been huge. The financial toll is worse. And I’m one of the lucky ones. I had the time, the technical curiosity and the persistence to push this far. Many families just give up.
What Needs to Change
If the UK is serious about scaling up heat pumps, the following must happen:
- Accreditation must mean accountability. If an installer fails, there must be a clear path for consumers to get redress, not just a polite email.
- Independent assessments must be formalised. We need a network of engineers who can provide recognised, insurer-grade reports on bad installs. At the moment, they barely exist.
- Servicing must be standardised. Annual service should mean something, not just a hose-down and an invoice.
- Finance providers must adapt. Section 75 and insurers need to recognise that heat pumps aren’t bathrooms. They should fund expert site surveys and accept technical reports in place of “like-for-like quotes.”
- Buffers must stop being a crutch. In my case, the buffer tank masked a bad design. It looked fine on paper, but in reality it was strangling flow and dumping energy. Buffers have their place in specific contexts, but in most domestic installs they’re just disguising poor design.
Where I Am Now
As I write this, I’m still chasing quotes for Section 75. I’m still living with a system that under-delivers. And I’m still waiting for the industry to show me it cares more about homeowners than about paper compliance.
If you’re considering a heat pump, demand clear design calculations based on your desired room temperature (in our case 22.5-24C, installers often assume 18-21C suits everyone) and emitter schedules that use MARC-confirmed radiators. You don’t need to do the maths yourself, but be aware of one crucial point: a radiator quoted at 1,000 W at ΔT50 will only deliver about 295 W at typical low-temperature operation. So I’ve found that a room with a 600 W heat loss needs about 2,500 W of radiator capacity to be safe. In multi-storey homes, aim for a slight over-delivery on the ground floor because heat rises. Insist on a proper commissioning plan, and if an installer tells you “the buffer will sort it,” walk away, because no family should have to pay £22,000 to sit in a cold house and fight for years to put it right.
And as it stands, it looks like I might have to pay another £10,000 to get it rectified, which takes me north of £30,000 for this installation.
Editor’s note
This article represents the personal account and views of the homeowner. Renewable Heating Hub does not independently verify or endorse every claim made, and we recognise that there are always multiple sides to any installation. However, the homeowner provided more than 190 supporting documents, including correspondence and technical reports, which we reviewed before publication and which substantiate the issues described. On that basis, we considered the account to be in the public interest to share.
As I was still trying to perfect my radiators, I ran into a few interesting facts. Using the Stelrad online heat-loss calculator, I was recommended a 1,881 BTU radiator to cover a ~554 W heat loss. This brings me to the following information, there’s a bit of maths involved so brace yourselves, but I do hope it helps shed light on what so often goes wrong before the installation even commences. It also shows why potential heat pump owners must make sure the radiators their installers recommend actually meet the demand at their chosen room temperature. This isn’t written to prove towel radiators are “not fit for purpose”, only to illustrate that in my case, the typical bathroom towel rail is unable to cope with the calculated demand.
Using a MARC-listed brand (Stelrad), I ran their basic heat-loss calculator:
• Installer’s HL for the bathroom: ~535 W
• Stelrad HL calculator: ~554 W (close enough)
Why catalogue watts don’t equal real watts at low temps
Radiator brochures quote outputs at a reference ΔT (temperature difference) typically ΔT50 (50 K between mean water temp and room). Real low-temp systems run with much smaller ΔT, so you must scale catalogue outputs with the EN442 correction:
Actual output = Catalogue output × (ΔT_actual / 50)¹·³
(or, equivalently, Catalogue W ÷ (50/ΔT_actual)¹·³).
(Exponent 1.3 is the standard EN442 convector factor.)
Assumptions (typical for heat pumps):
45/40 °C flow/return → MWT = 42.5 °CRoom setpoint examples: 21 °C and 23 °CSo ΔT_actual = 42.5 – 21 = 21.5 K (at 21 °C) or 42.5 – 23 = 19.5 K (at 23 °C)Conversion factors from ΔT50:(50/21.5)¹·³ ≈ 2.996(50/19.5)¹·³ ≈ 3.401
Quick rule of thumb from that:
At 21 °C, 1,000 W @ ΔT50 → ~334 W actual (1000/2.996)At 23 °C, 1,000 W @ ΔT50 → ~294 W actual (1000/3.401)
Bathroom example (shows the problem with towel rails)
Target HL ≈ 535–554 W.
Stelrad Classic Towel Rail 1211×500 (brochure): 1965 BTU / 576 W @ ΔT50
Delivered at 21 °C room: 576 / 2.996 ≈ 192 WDelivered at 23 °C room: 576 / 3.401 ≈ 169 W
Even a much bigger rail, Stelrad Caliente White Double Towel Rail 1199 X 600mm 1154 W @ ΔT50, gives:
1154 / 2.996 ≈ 385 W (21 °C)1154 / 3.401 ≈ 339 W (23 °C)
Both miss the ~535–554 W requirement by a long way. So what are my options?
Why my “~295 W per 1,000 W @ ΔT50” line is fair
At 45/40 flow and 23 °C room, the factor is ~3.40, so 1000/3.40 ≈ 294–295 W delivered.
At 21 °C room it’s ~334 W per 1000 W. Pick the one that matches your setpoint.
Tip: I’d rather design for 23 °C and run cooler later than design for 21 °C and be forced to crank flow temps (and bills) in winter.
Buffers can devastate emitter output
Measured on my system (thermal camera + controller): HP flow ≈ 48 °C, but buffer → rads ≈ 40 °C (≈ 8 °C lost). Radiator output scales with ΔT¹·³, so dropping ΔT from (say) 27 K to 19 K cuts output to (19/27)¹·³ ≈ 0.63 — about 37% less heat from every emitter before you even consider sizing errors. That’s huge.
Constants used: 1 W = 3.412 BTU.
Scaling: EN442 radiator correction, n ≈ 1.3.
MWT: (Flow + Return)/2.
Bottom line: check room-by-room heat loss, set your real room temperature target, and convert every radiator’s ΔT50 brochure watts to actual watts at your MWT–room ΔT. Bear in mind that any buffer-induced temperature drop compounds the shortfall.
If I was to take a wild guess, I would say that if my buffer is repurposed as a volumiser, replaced with a better option, or removed (where hydraulically appropriate), I’d expect an uplift in seasonal efficiency of around 30%; which would be a huge.
Some additional information:
My whole life, I’ve run my gas boiler 24/7 with the downstairs thermostat set at 25°C. I manually adjusted the flow temperature using the dial on my Vaillant Combi Boiler (from 10°C to 75°C) depending on how warm I wanted the house. Most of the time, I ran it at 28°C, and in winter I’d increase it to around 40–50°C during the day, occasionally up to 60°C at night, though that was rare. Most winter days I’d sit comfortably in shorts and a T-shirt with the boiler running around 52–57°C.
Poor efficiency aside, maybe this explains why the heat pump “hype” didn’t really make much of a difference for us. If you run your gas boiler like a heat pump, your house can stay cosy and warm all the time, and it might even save you money. Might be worth a try this winter.
We’ve also been extremely unlucky on the financial side of going green, missing out on every possible saving. We received a £5,000 BUS grant instead of £7,500 because we applied a month too early, and we lost a further £2,000 “Green Living” reward from our mortgage provider due to technicalities (we used a broker in 2023, only for the rules to change in July 2024 to include such cases).
Taken together, this journey shows just how much cumulative financial harm and unfairness there can be. For us, “going green” has become a huge financial burden. If I had to make the choice again, knowing what I know now, I wouldn’t do it. If it costs £30,000+ per family to “save the planet,” perhaps that cost should fall on the companies cutting down forests and polluting on an industrial scale, rather than on individual households trying to do their bit.
And finally, one has to wonder: are these BUS and Green Grants truly “free,” or will we end up paying for them in one form or another anyway?
P.S:
1. Gas boiler run like a heat pump (always on, low flow temps)
This is like motorway driving: once you’re cruising steadily, you use fuel more efficiently per mile.
2. On/off heating schedule (boiler off in the day, back on as you get home from work)
This is like stop–start city driving: every time you accelerate from a standstill, you burn more fuel than if you’d just kept rolling.
3. So does it mean more savings?
I’d glanced through the post and comments but hadn’t fully read your crime sheet. This is terrible but common, all the accreditation type bodies refer you to the installer with any issues. MCS are the worst but HG also has some non complying members, 3 in the last week! I’ve now seen paperwork proof.
Your home or the work involved must be above average, the Samsung heat pump is £4000, 250L cylinder £2000, have you photos you can post?
No heat pump should be running at 50C to rads, that’s full capacity and may use electric boosters in colder weather. If a heat pump mostly runs above 50% compressor speed then there’s an issue with the install, or its -5C outside
@Mars has all my files, if he doesn’t mind maybe he can share the photos and design shortfalls I have highlighted. I cannot send private messages yet, still new member status, and I don’t want to post the link to all the files, as some of the information is private/confidential. For the sake of the article, I did share EVERYTHING, as I wanted to be completely transparent, with all the facts available, so no one can rebut any of it.
In regards to HG, from what I understand, they have now started to check the names on the contracts, installations and commissioning documents, and blacklist the people not just the companies. Which means that if they move on to other companies, there is a record of their antics.
In my case I provided them with all the relevant names, project manager, design engineer who signed off my system design and commissioning, head/lead engineer installer who did the testing and signed off the installation.
This is a much needed step forward, if indeed this is the case.