Use the good roof first
Maximise sensible, productive roof space before paying for panels on a weak roof face simply to increase the panel count.
One homeowner. Five very different answers. One very useful rabbit hole.
Buying solar in the UK shouldn't feel this complicated.
I started with a simple question: what solar and battery system makes sense for our house? The quotes that came back varied hugely in panel count, roof faces, battery size, inverter power, backup capability and claimed payback.
This is a personal project diary, not financial or electrical advice. Prices and tariffs are a September 2026 snapshot and every property is different.
Why I made this
The quote price was only part of the decision.
I quickly learned that two systems at a similar price can be completely different products. A smaller array with a bigger battery is not the same as a larger array with a smaller battery. A 5 kW inverter is not the same thing as a 10 kW inverter. “Backup” might mean a single emergency socket, selected house circuits, or a much more comprehensive setup.
Most importantly, the software assumptions matter. The same roof can be modelled with different slopes, shading factors, consumption figures, tariffs and future energy inflation. Those assumptions can move the advertised payback by years.
Our priorities
What we were actually trying to achieve
The “best” system is the one that fits the household. These were the things that shaped our choice.
Size for the future
We already have a PHEV and expect a heat pump later, so inverter headroom and expandable storage mattered more than optimising only for today's loads.
Make backup meaningful
Keeping Wi‑Fi, smart-home control, lighting, refrigeration and selected sockets alive in a power cut was a real requirement, not a brochure extra.
Model the tariff we use
We can buy very cheap overnight electricity, so the battery can earn its keep in winter by shifting cheap grid energy as well as storing solar.
How I actually chose
The winning quote did not come from comparing prices in a spreadsheet.
The biggest change in my thinking was moving from “which package looks best?” to “who has understood this particular house, and can explain every design choice?”
Use the big names as benchmarks
National and online quotes were genuinely useful for learning the market, seeing different hardware and establishing a price range. I just stopped treating the first desktop design as the final engineering answer.
Compare the design, not the logo
The same house was offered dramatically different panel counts, roof faces, inverter sizes, batteries and paybacks. Those differences were more important than the brand name at the top of the PDF.
Interrogate contradictions
If one company says ten panels and another says twenty-four, or one assumes heavy shading while another assumes none, that is not noise to ignore. It is the point at which to start asking “why?”.
Search for certified installers locally
I eventually went to the MCS installer directory and looked for firms operating locally rather than relying only on national sales funnels and lead-generation sites. That is how EcoGlow came onto my shortlist.
Triangulate the reviews
I checked several review sources, searched specifically for solar/battery comments, read the poor reviews and looked at how companies responded. A single impressive Trustpilot number was never enough on its own.
Pick up the phone
Talking to people was incredibly revealing. Could they answer technical questions without reading a script? Would they challenge my assumptions? Did they explain the trade-offs, or simply keep trying to close the sale?
Push for a physical site survey
This became non-negotiable for me before the final design. Roof measurements, Velux windows, shading, access, consumer unit, equipment positions and cable routes are physical things. A satellite image cannot see all of them.
Get the awkward bits in writing
Scaffolding, DNO work, bird protection, backup circuits, EV charging, concealed cable routes, making good, warranties and aftercare all became part of the decision — not “extras” to discover on installation day.
Why I moved away from a national-only approach: I am not saying a national installer is automatically worse. My own confidence changed when I met an installer at the property who could physically look at the roof, talk through cable routes and backup, and discuss how the work would actually be delivered. For a project of this value and lifespan, I would now push hard for a proper site visit before treating any design as final.
A separate decision
Do you even need backup?
I did. Plenty of people may not. A battery and a backup system are not the same thing, and resilience should be a deliberate choice rather than something you assume is included.
Cheapest and simplest
If power cuts are rare and there is nothing important you need to keep alive, spending extra on outage protection may not make financial sense. Solar and a battery can still work perfectly well day-to-day without being configured to run the house when the grid fails.
A socket or two
A lower-cost middle ground can be dedicated emergency sockets for a fridge, router, lamps or phone charging. Useful, but very different from the house continuing to behave normally.
Selected circuits
I wanted automatic backup for the network/Control4 rack, lighting, refrigeration and useful house sockets, with other circuits included only where the electrician confirms the loads make sense. That is why I paid extra for a dedicated backed-up consumer-unit arrangement.
Do not assume it
Cookers, heat pumps, EV chargers and other large loads change the engineering. Ask what the inverter, battery and EPS output can actually supply in an outage and which circuits will be isolated. A future heat pump should not simply be assumed to be backed up.
The key distinction
Backup is mainly a resilience purchase.I would not force it into the solar payback calculation as though every pound spent on backup has to generate a return. For me, keeping the smart home, internet and core circuits functioning in a power cut has value in its own right. Someone else may quite reasonably spend that money elsewhere.
The comparison
Five proposals, five different designs
These are snapshots from quotes/configurators I received for the same property. They are not perfectly like-for-like, which is exactly the point.
Swipe sideways on mobile.
| Installer / stage | Solar | Inverter | Battery | Backup / EV | Price snapshot | What stood out to me |
|---|---|---|---|---|---|---|
| EcoGlow Chosen |
24 × 465W DMEGC 11.16 kWp Modelled 10,097 kWh/yr |
Fox KH10 10 kW |
Fox EP12 Plus 11.52 kWh nominal ~10.36 kWh usable in proposal |
Zappi included Selected-circuit automatic backup added |
£14,520 base £15,720 incl. £1,200 backup upgrade |
Largest sensible array, 10 kW inverter, expandable heated battery, EV charger included, physical site survey and practical cable/making-good discussion. |
| Professional Energy Solutions Detailed proposal |
18 × 470W DMEGC 8.46 kWp Modelled 7,639 kWh/yr |
Fox H1 G2 6 kW |
Fox EP12 Plus 11.52 kWh nominal |
Backup +£1,100 EV charger +£1,150 |
£12,750 base £13,850 incl. backup |
A strong, sensible design that deliberately avoided the north-facing roof. The main trade-off for me was the 6 kW inverter versus the larger future-facing option. |
| Glow Green Tesla proposal |
16 × 475W DMEGC 7.60 kWp Modelled 6,107 kWh/yr |
Tesla Powerwall 3 configured to 5.0 kW |
Tesla Powerwall 3 13.5 kWh |
Tesla Gateway included | £13,740.33 | The design used three roof faces, including a very northerly one. It also assumed a shading factor of 1.00. That prompted me to challenge whether every proposed panel was really earning its place. |
| Heatable 20-panel comparison option |
20 × 485W LONGi 9.70 kWp |
Sigenergy SigenStor EC 8.0 kW |
SigenStor BAT 10.0 9.04 kWh 8.76 kWh usable shown |
Integrated system proposition | £15,605 | High-quality-looking package and a large array, but at a materially higher price than some alternatives. It reinforced how much the roof-layout assumptions drive the result. |
| Octopus Energy Early online estimate |
10 panels Online layout stage |
Fox EVO 5 package shown | Configurable portal recommended smaller storage initially |
Online package estimate | £10,149 shown for 10 panels + Fox EVO 5 package | Useful as a national benchmark, but the early automated design was much smaller than later site-specific proposals. I treated it as a starting point rather than a final engineered answer. |
Quote snapshot dates: September 2026. Prices may have changed, and optional extras differ between suppliers. This table describes what was proposed to one house; it is not a ranking of installers.
Installer due diligence
The star rating is not the review.
This became one of the most important parts of my search. A great score on one website can be reassuring, but I would never choose a solar installer from one review platform alone.
How I found the installer I chose
MCS search → local shortlist → reviews → phone calls → site visit.
I did not arrive at EcoGlow because an advert told me to. I searched the MCS directory for installers working locally, sifted the shortlist through several review sources, made contact and spoke to people. The deciding factor was then the quality of the conversation and the physical survey at the house. MCS certification was a starting filter, not the entire due-diligence process.
Check more than one platform
I would look at Google, Trustpilot and, where relevant, a trade or local-review platform. Each has a different mix of customers and review collection methods, so the pattern across them matters more than one headline number.
Review the service you are actually buying
A company may sell boilers, heat pumps, EV chargers and solar. Thousands of excellent reviews for another part of the business do not automatically tell you how good its solar design, installation and aftercare are. Search within reviews for words such as solar, battery, inverter and aftercare.
Read the newest and the worst
The average score hides the useful detail. Sort by recent reviews, then deliberately read the lower-rated ones. I was looking for repeated themes: missed appointments, subcontractor problems, poor communication, installation quality, commissioning delays and how warranty issues were handled.
How a firm responds matters
Problems happen. A calm, specific response that explains how an issue was fixed can be more reassuring than a page of perfect reviews. Repeated complaints with defensive or generic replies were much more concerning to me.
Then get on the phone
Reviews tell you about other customers; a conversation tells you how the company deals with your questions. I would speak to two or three serious contenders and ask the same awkward questions about roof layout, backup, DNO, cable routes and aftercare. The contrast can be striking.
Look for local evidence
Recent installations near you are useful because they tell you the company is actively working with similar roofs, local grid conditions and access constraints. If I were unsure, I would ask for one or two recent local customers who are happy to be contacted.
Reviews are only one layer
I also checked whether the installer could support the technical promises in the quote: MCS/NICEIC credentials where applicable, warranties, DNO process, written scope, payment protection and who would actually attend site. Reviews should reinforce that evidence, not replace it.
My rule of thumb
Three sources beat one score.Google for local/current experience, Trustpilot for scale and trend, plus a trade/local source or direct recent references. Then search specifically for the product you are buying and read the low-scoring reviews as carefully as the five-star ones.
A lesson from my own shortlist: one national installer looked very reassuring from its overall Trustpilot presence, but the picture was less clear when I checked Google and focused specifically on solar rather than the company's other services. That did not automatically make it a bad installer; it simply showed why a single headline score can be misleading.
What changed my thinking
The useful questions were not “which panel is best?”
Most mainstream panels are good enough that the design, electrical architecture and assumptions can matter more than obsessing over a tiny difference in module efficiency.
kW and kWh are different
kW is power — how fast the system can supply or absorb energy. kWh is capacity — how much energy is stored or consumed over time. A big battery behind a small inverter and a smaller battery behind a large inverter behave very differently.
A battery does not automatically mean backup
Grid-connected batteries can still shut down with the grid unless the system has the necessary EPS/gateway/transfer arrangement. Decide whether you want outage protection first, then get the exact backed-up circuits and power limits written down.
Export limit ≠ house-use limit
A DNO may approve a system with an export limit. That does not automatically mean the house can only use that same amount internally. Ask the installer to explain the approved export limit, inverter capability, battery output and backup/EPS output separately.
A satellite layout is not a roof measurement
Desktop tools are excellent for early design, but they can miss real-world constraints. One of my draft layouts even put panels across a roof window. Final panel count should follow physical measurement, access and installation clearances — not the other way around.
North-facing is not automatically wrong
It can still generate electricity, particularly on shallow roofs, but the marginal generation must justify the panel, labour, inverter input and roof space. Ask for the predicted annual output of each roof face separately.
Shading assumptions can swamp small spec differences
If one proposal assumes zero shading and another applies a large shading penalty to the same roof, do not just average the two numbers. Ask how each figure was derived and whether anyone has physically inspected the property.
Tariffs change the battery maths
A home with very cheap overnight power has another use for a battery: charge cheaply and discharge through expensive daytime periods. A payback model that ignores your actual tariff can be misleading.
You do not have to buy all the battery on day one
I deliberately preferred an expandable system. One battery now gives us real import/export data; if a second battery later proves worthwhile, it can be added. Expansion capability can be more valuable than paying upfront for capacity you may not use.
Future loads can justify inverter headroom
A heat pump, full EV, induction cooking or more electrification can increase instantaneous demand. Bigger is not always better, but future loads should be discussed before locking in an inverter architecture.
The wallbox rating is not necessarily the car's charge rate
A 7 kW charger does not make every car charge at 7 kW. The vehicle's onboard AC charger, the phase arrangement and the supply can all be limiting factors. Size the infrastructure for the future, but understand what today's car will actually accept.
Outdoor battery behaviour matters
If the battery is outside, ask about operating temperature, charging limits in cold weather and whether battery warming is built in. A capacity figure on a brochure does not tell you everything about winter behaviour.
Cable routes and making good are part of the product
The neatest equipment in the world can still create a poor job if cables are surface-run where you did not expect them or nobody owns the making-good. Agree routes, penetrations, equipment positions and who repairs finishes before work starts.
Panel count is not a quality metric
“We can fit 24” and “we recommend 24” are different statements. I wanted each roof face to have a reason for being used, not simply the largest headline number.
Handover paperwork matters
DNO approval/notification, MCS certification, electrical certificates, warranties, app ownership, settings and commissioning information are part of the installation. I want the system documented well enough that somebody other than the original salesperson can support it years later.
Where we landed
Our chosen system
We chose EcoGlow's larger Fox-based design, then added the more comprehensive selected-circuit backup option. The hardware mattered, but the decision was also about confidence in delivery: they came to the property, discussed the real roof, cable routes, equipment positions, making good, backup and future heat-pump plans rather than treating the house as a remote desktop exercise.
The final roof count is still subject to precise scaffold-level measurement; if a panel does not physically fit, it should come off rather than be forced into the design. I also chose one expandable battery initially rather than automatically buying a second before we have real usage data.
Solar24 × DMEGC 465W / 11.16 kWp
InverterFox KH10 / 10 kW single phase
BatteryFox EP12 Plus / 11.52 kWh nominal
EV chargingZappi tethered charger
BackupAutomatic selected circuits
Price£15,720 including backup upgrade
Simple sense-check
Rough payback calculator
This deliberately simple model makes the assumptions visible. It is useful for challenging a sales forecast, not replacing a full design model.
Illustrative result
7.2years simple payback
Solar self-use value£1,137
Export income£600
Cheap-rate battery shifting£458
Total annual benefit£2,195
Defaults mirror my conservative planning case, not the installer's advertised 5-year payback.
Read the inputs, not just the payback headline: one of my final proposal models was using 12,000 kWh/year of household consumption while our real run-rate was closer to 9,200 kWh. It was still useful as a design document, but I would not take the advertised payback at face value without correcting the assumptions.
Avoid double-counting: the “cheap-grid energy shifted” box is intended for grid energy bought cheaply and used later. If you are modelling solar that charges the battery, do not count the same solar kWh again as both self-use and battery arbitrage.
Before signing
The questions I would now ask every installer
- Will you come to the property before I treat the design as final? A desktop quote is useful; a physical roof, access and electrical survey is better.
- What annual electricity usage have you modelled? Is it my real usage or a default?
- Show me each roof face separately. Orientation, pitch, panel count and predicted annual generation.
- How did you calculate shading? Desktop estimate, horizon survey, physical site survey or optimiser/microinverter assumption?
- Why this many panels? Is the layout physically measured, and what happens if the scaffold survey finds less space?
- Where will everything physically go? Panel access, inverter, battery, isolators, EV charger, consumer unit changes, cable routes and who is responsible for making good.
- What limits the power? Inverter AC output, battery charge/discharge output and EPS/backup output are separate numbers.
- What exactly is the DNO application for? What has been submitted, what export limit is requested/approved, and is export limitation part of the design?
- Do I actually need backup? If yes, is it emergency sockets, selected circuits or something broader — and what will definitely stay live in a power cut?
- How much battery is genuinely usable? Ask about depth of discharge, continuous output, cold-weather behaviour, battery warming, warranty, cycles and expansion options.
- Could I start with one battery and add another later? Ask what expansion requires and whether the inverter/battery architecture supports it cleanly.
- What will my EV actually charge at? Separate the wallbox's rating from the car's onboard AC charging limit and the house supply.
- Which tariff assumptions produce the payback? Import rate, cheap window, export rate, energy-price inflation, degradation and whether grid-charging the battery is modelled.
- What changes if I add a heat pump or full EV? Ask before choosing the inverter, battery and backup architecture.
- What is included in the price? Scaffolding, bird protection, DNO work, MCS/electrical certification, EV charger, backup hardware, cable routes, making good and any network charges.
- Are you MCS certified for the work, and who is actually doing the installation? I used the MCS directory as a starting filter, then checked the company itself rather than assuming the badge answered every question.
- What do the reviews look like away from the headline score? Check several sources, search for solar-specific feedback, read recent low-rated reviews and look at how the installer responds when something goes wrong.
- Can I speak to someone technical before signing? Get on the phone. Ask the same awkward questions of the serious contenders and listen to how they explain the trade-offs.
- Have I told my home insurer? Confirm how the permanent solar, battery, EV charger and electrical work should be noted and keep the install/certification paperwork.
- Who supports it after installation? App setup, commissioning, fault diagnosis and warranty handling matter for a system expected to last decades.
- What documents will I receive at handover? DNO outcome, MCS certificate, electrical certification, warranties, commissioning settings and equipment details should not be an afterthought.
Plain English
Solar jargon in 60 seconds
kWpPeak rated size of the solar array under test conditions.
kWInstantaneous power — how fast energy is flowing.
kWhEnergy over time — battery capacity and electricity usage.
DNOYour local electricity distribution network operator.
G99 / export approvalThe connection process relevant to larger generation/storage systems. Exact requirements depend on the design and network.
EPS / backupEmergency power supply when the grid is unavailable, subject to the system and circuits configured.
DoDDepth of discharge — how much of a battery's nominal capacity is intended to be usable.
Round-trip efficiencyHow much energy comes back out compared with what went into the battery.
The boring bits
The installation is not finished when the panels go on.
A good handover should leave you with a system you understand, documents you can find and settings that match the way your household actually uses energy.
Keep the paperwork
Store the MCS certificate, electrical certification, DNO correspondence/approval, warranties, datasheets, final design and installer contact details together. Future you will thank you.
Tell the insurer
I would notify the home insurer about the permanent solar, battery, EV charger and associated electrical works and get any confirmation recorded rather than discovering a question at claim time.
Set the system up for your tariff
Battery reserve, overnight grid charging, export behaviour and EV schedules can materially change the outcome. In my case I do not want a car charge at the cheap overnight rate to pointlessly drain the house battery.
Close the loop with real data
Once installed, compare actual generation, import, export and battery cycling against the forecast. That is when you can make an evidence-based decision about another battery, a different tariff or future heat-pump integration.
About this site
Independent homeowner research, not an installer comparison business.
I built The Solar Minefield after discovering how dramatically different solar proposals for the same house could be. The aim is to make the questions, assumptions and trade-offs visible so other homeowners can have better conversations with installers — and to show the process I ended up trusting: compare widely, verify credentials, read reviews across sources, speak to people and get somebody competent to look at the actual property.
At launch, nobody has paid to appear here and there are no paid rankings. If advertising, affiliate links or referral arrangements are added later, they will be disclosed clearly rather than hidden inside the comparison.
For the full wording, see Independence & disclosure and Privacy.
Project status
September 2026 — installation phase
The EcoGlow proposal has been accepted. The DNO application has been submitted, scaffolding is being arranged, and the roof will be measured precisely before the final panel layout is locked down. The next useful test is reality: what actually fits, what the DNO approves, how neatly the system is installed and then what it really generates, imports, exports and saves. I will keep updating this page with that evidence rather than leaving it as a pre-installation sales comparison.