Solar calculator methodology

Last updated: 2026-04-30. Tracking 26 SEG tariffs across 11+9+6 eligibility tiers.

The short version

We model your annual solar generation from system size, location, orientation, and tilt. We split it into self-consumption (what you use directly) and export (what goes to the grid). We then run that through every SEG tariff currently on the market, grouped by what you actually qualify for, and surface the best in each tier so you can see what’s realistic versus what’s marketing.

Every number on the results page comes from this calculation. Every assumption is documented below. Every supplier rate has a `verifiedAt` date and a clickable source URL.

Step 1: Annual generation

The core formula:

Annual generation (kWh) = System size (kWp) × Regional irradiance × Performance ratio × Orientation factor × Tilt factor

Regional irradiance

The amount of solar energy hitting a horizontal square metre in your region per year, measured in kWh/m²/year. We use 14 GSP regions matching the UK’s electricity distribution map. Values range from about 770 (Northern Scotland) to 1,020 (South-East England) - so location really does matter.

Source: PVGIS, the EU’s open photovoltaic geographical information system. This is the same dataset professional installers use.

Future improvement: V1.1 will integrate the PVGIS API for postcode-specific irradiance rather than regional averages, improving accuracy by 5-10%.

Performance ratio

Real-world losses including inverter efficiency, wiring losses, panel temperature, soiling, and module mismatch. We use 0.80 (80%) - the standard assumption for well-installed UK domestic systems. A new high-quality install might hit 0.85; an older or poorly-sited system might be 0.70.

Orientation factor

Multiplier vs perfect south-facing. Our values:

• South: 1.00
• South-East / South-West: 0.96
• East / West: 0.83
• North-East / North-West: 0.70
• North: 0.62

These are derived from PVGIS modelling for UK latitudes at optimal tilt.

Tilt factor

UK optimum is roughly 35° (which matches typical UK pitched roofs). Each 10° deviation from this costs about 2-3% annual generation. Flat (0°) loses about 9%, vertical (90°) loses about 30%. We use a linear approximation - more than accurate enough at this level of estimation.

Step 2: Self-consumption split

Solar generates in the middle of the day; most household demand is morning and evening. Whatever the overlap does not cover gets exported. That overlap is worth several hundred pounds a year, because a kWh you use yourself is worth the import rate you avoid — around 25p — while a kWh exported earns the SEG rate, which for most people is a fraction of that.

What this used to do, and why it was wrong

Until September 2026 this calculator assumed 30% self-consumption without a battery and 70% with one, for every household. Both figures are quoted widely across the industry. Neither can be right for everybody, because a fixed percentage cannot depend on the one thing that decides the answer: how much electricity the household actually uses.

The second figure is worse than imprecise. For a household on Ofgem’s medium consumption of 2,500 kWh with a 4 kWp array generating 3,600 kWh, 70% of generation is 2,520 kWh — which is 101% of everything they use in a year. Every unit they consume, at every hour, including three in the morning in January, would have to come off the roof. No battery does that. The number is not optimistic, it is unreachable, and it was ours.

What it does now

Self-consumption is an overlap between two curves, so we integrate it: min(generation, demand), across 24 hours and 12 months.

• Demand comes from the monthly hour-of-day curves in our open cohort data, derived from around a billion smart meter readings and scaled to the annual consumption you enter.
• Generation shape comes from solar geometry — declination, hour angle and elevation at 54.5°N.
• Generation total stays exactly as Step 1 calculated it, from the PVGIS-derived regional irradiance figures.

The split matters: geometry knows the shape of a solar day and nothing about British cloud; the regional irradiance figures know the annual total and nothing about the shape. Each is used only for what it is actually evidence for.

Why the weather had to go in

The first version of this modelled each month as one average day. It returned about half of generation self-consumed and made battery size irrelevant — 5 kWh and 20 kWh gave the same answer. Both failures were the same failure: the minimum of two averages is not the average of two minima. Smoothing away the bright days and the dull ones deletes precisely the surplus that gets exported, and precisely the surplus a battery exists to catch.

So each month is modelled as a spread of day types instead:

These weights redistribute generation between days. They never create or destroy any: the annual total still comes from Step 1, unchanged.

This is the one calibrated parameter in the model, and it is worth being blunt about that. The geometry is exact, the annual totals are PVGIS-derived and the demand curves are measured — but “how variable is a British day” is a spread we have tuned so that the typical configuration lands on the roughly 30% that MCS and Energy Saving Trust guidance reports. If you want to attack this model, attack that table.

What it produces

For that same 2,500 kWh household with a 4 kWp array: 33% self-consumed with no battery, rising to 45% with a 10 kWh one — not the 70% we used to claim, because the house cannot absorb that much.

The part worth reading before you size an array is what happens when it grows. Same household, same demand curve, bigger roof:

Doubling the panels does not double what you keep. Past a point you are building an export business at whatever SEG rate you can get, which is a different proposition from cutting your own bill — and a flat percentage cannot tell you where that point is.

The same ceiling applies to the battery. For this household extra capacity stops earning above about 4 kWh: a solar battery is limited by how much electricity the house uses after dark, not by how much it can hold. Our battery ROI calculator reaches the same conclusion from a completely different direction.

If you know your own figure — a smart meter export reading against generation is the way to get it — the “Advanced options” section still lets you override the model entirely.

What this still does not capture

• No link between weather and demand, though cold dull days really do arrive together.
• Whole hours rather than half hours, which smooths over kettle and cooking peaks and nudges self-consumption slightly upward.
• A battery that charges from surplus and discharges into unmet demand the same day. No forecasting, no grid charging, no carrying summer into winter.
• One representative latitude rather than yours. It changes the shape of the day, not the annual total.

Step 3: Self-consumption savings

Self-consumption savings (£) = Self-consumed kWh × Import unit rate (p/kWh) ÷ 100

We use your current import unit rate as you’ve entered it. The default is the current Ofgem price cap rate, but you can adjust it if you’re on a different tariff.

Notice this number doesn’t change between tariffs - it’s based on what you save by not buying grid electricity. Tariff choice only affects the export portion.

Step 4: SEG export comparison - the three tiers

Smart Export Guarantee (SEG) is the scheme that replaced Feed-in Tariffs. Suppliers with 150,000+ customers must offer one. Rates vary dramatically - from 1p/kWh (legal minimum) to 25p+ for top tariffs.

Almost every major supplier offers multiple tariffs at different eligibility levels. Choosing “Octopus” doesn’t mean a single rate - they have rates from 4.1p (open to anyone) to 32p (with battery, on closed Flux tariff). Most other UK calculators ignore this and pick one rate per supplier, often misleadingly.

We group the 26 tariffs we track into three tiers based on what you actually need to do:

Tier 1: Open (11 tariffs)

Anyone with an MCS-certified solar install and a smart meter qualifies. No need to switch electricity supplier. Top rate currently around 13p (Fuse Energy).

These are the only rates a user can definitely access without changing anything else about their setup. If you’re happy with your current supplier, this is your realistic best.

Tier 2: Switch electricity supplier (9 tariffs)

You must take the supplier’s electricity import tariff too. This is where the bulk of value sits - a willing user can typically jump from 4p to 12-16p. Top rate currently around 16p (Ecotricity).

The key caveat: check the import tariff isn’t terrible before switching. A great export rate paired with an expensive import rate can be worse overall than a lower export rate with a cheaper import. We don’t model that combined optimisation in V1, but we flag it loudly.

Tier 3: Installer-locked (6 tariffs)

You must use the supplier as your solar installer too. Highest headline rates (15-25p) but you give up choice of installer, which is a real trade-off - these aren’t always the best installers in the market.

Many installer-locked tariffs are 12-month promos that revert to lower rates. We expose this with a separate “Effective/yr” column. See “Effective rates” below.

Effective rates: the honest comparison

Some installer-tier tariffs market themselves with a high headline number that only applies for the first 12 months, then drops sharply. Two examples in our current dataset:

• Good Energy Solar Savings Exclusive: 25p year 1, then 15p for years 2-25. The 25p headline is technically true but misleading.
• So Energy So Bright: 20p year 1, then 4.5p for years 2-25. The 20p headline collapses to 4.5p almost immediately.

Our calculator shows two annualised numbers per tariff:

• Year 1: what you actually receive in the first year. The number most other calculators show.
• Effective/yr: the rate-weighted average across the whole 25-year lifetime. For non-promo tariffs this equals year 1 exactly. For promo tariffs it’s the honest blend.

On a typical 4kWp south-facing system without battery, So Energy So Bright shows £639/year in year 1 but only £333/year averaged across 25 years - which makes it worse than several open-tier tariffs that don’t require any commitment at all. The marketing rate is misleading. We surface this directly.

Step 5: Lifetime modelling

Solar systems lose efficiency slowly over time - typically 0.5% per year (the standard manufacturer warranty assumption). Over a 25-year lifetime that compounds to about a 7% reduction by year 25.

We use a simplified linear model:

Effective lifetime years = Lifetime - (Degradation × Lifetime² ÷ 2)

For 25 years at 0.5% degradation, this gives an effective ~23.4 years of full-output generation. The lifetime benefit calculation factors in both the rate schedule (including any promo period drop-off) AND degradation.

We use 25 years as the lifetime because that matches typical inverter warranty periods. Panels themselves often last 30+ years, but most installs need an inverter replacement once during their working life (typically £800-£1,500).

Things we deliberately do not model in V1:

• Inverter replacement cost - keeps the calculation simple but slightly optimistic
• Electricity price inflation - if import prices rise faster than SEG rates, your savings grow over time. We don’t assume this either way.
• SEG rate changes - rates can change. We assume the current rate stays available, with the explicit exception of promo periods which we do model.
• Membership fees - some tariffs (e.g. OVO SEG Beyond Exclusive) require paid memberships. We don’t deduct these from the headline. Factor them in yourself.

Step 6: CO₂ savings

Every kWh of solar you generate displaces a kWh that would otherwise have come from the grid. The grid emits about 124 gCO₂/kWh on average (NESO 2025 figure).

CO₂ saved per year (kg) = Annual generation × 124 ÷ 1000

This is a conservative figure - the UK grid is decarbonising rapidly, but solar offsets at the time of use, which means it offsets daytime grid mix. As the grid gets cleaner, the relative CO₂ benefit of solar shrinks slightly. But solar is always net-positive for emissions.

What the calculator doesn’t do (honest list)

V1 does not handle:

• Time-of-use tariffs end-to-end. Tariffs like Octopus Flux pay 30p+ during peak windows and ~5p off-peak. Modelling these honestly needs hour-by-hour generation and consumption profiles. We’ve excluded the pure time-of-use tariffs from our comparison rather than show misleading averages.
• Octopus Outgoing Agile. We don’t include Agile in the comparison table at all. Octopus themselves disclaim that any single “average” rate is predictive of what users earn - the rate changes every 30 minutes and earnings depend entirely on when your solar exports happen relative to demand spikes. We surface Agile as a separate callout linking to the Octopus dashboard rather than risk a misleading single-figure comparison.
• Battery time-of-use arbitrage. If you have a Go-style off-peak tariff, you can charge the battery cheap overnight and use it during peak rates. This can be worth £100-300/year on top of solar self-consumption. Our V2 “Battery-only ROI” calculator will model this properly.
• Combined import-export optimisation. The best export tariff isn’t always the best overall - if it locks you into an expensive import tariff, you can lose more on imports than you gain on exports. V1.1 will start modelling this.
• Split-orientation roofs (e.g. east-west arrays). Calculate each side separately and add the results.
• Shading. A neighbour’s tree or chimney can dramatically reduce generation. We can’t see your roof.
• EV charging interactions. The demand curve behind the self-consumption model excludes households with an EV, so a car that charges during daylight would raise your self-consumption above what we show.
• VAT on installation. Currently 0% VAT on UK domestic solar+battery installs through 31 March 2027. This is included in typical quotes.
• Local planning constraints. Some properties (listed buildings, conservation areas) have restrictions.

How we keep this data fresh

SEG rates change, and product codes change too. During an April 2026 audit we caught two of our Octopus tariff entries pointing at outdated product codes (one was a closed-to-new-customers product; one was a code that didn’t exist in the API at all). The user-facing rates were coincidentally correct, but the underlying source-of-truth was wrong. We fixed this by verifying every Octopus entry directly against the Octopus public products API. Our commitment going forward:

• Every tariff in our dataset has a `verifiedAt` date stamped against it. If you see a rate older than 90 days, treat with caution and check the supplier’s site directly.
• Cadence: Quarterly manual review, plus daily change-monitoring (V1.1+).
• V1.1 will add automated daily change-monitoring for each supplier’s SEG page, with email alerts when a rate appears to have changed. This catches drift between scheduled audits.
• If you spot something out of date, email hello@energycosting.co.uk. We treat this as a top-priority bug, not a feature request.

Sources

The data underlying this calculator comes from:

• PVGIS (EU JRC) - regional irradiance values, orientation factors, performance ratio
• MCS - install standards and panel degradation typical values
• Ofgem SEG scheme - export tariff regulation, mandated supplier list
• NESO Carbon Intensity - grid emissions factor
• Each supplier’s own SEG page (linked from each tariff in the comparison table)

Found a mistake or have a methodology improvement to suggest? Email hello@energycosting.co.uk. We genuinely want to know - calculator accuracy is the entire point.

See also: solar calculator, heat pump methodology.