Self-consumption is the single biggest economic lever
The economics of a commercial solar PV installation come down to a remarkably simple equation. Every kWh the system generates is worth either your avoided grid retail price (currently 25-40p/kWh for most UK commercial customers) or the SEG export price (currently 5-12p/kWh). The fraction split between those two prices is what’s called the self-consumption ratio.
For a typical UK commercial customer, the difference matters enormously. At 30p/kWh retail and 9p/kWh SEG, every percentage point of self-consumption is worth 21p × annual generation. On a 300 kWp office system generating 276,000 kWh/year, every 1% of self-consumption is worth £580/year — about £14,500 over 25 years.
Across the typical 50-percentage-point range between low-self-consumption and high-self-consumption commercial installations, that’s the difference between a 4-year payback and an 11-year payback.
Why offices win
Office buildings hit higher self-consumption ratios than almost any other commercial building type. The reason is in the demand profile.
A typical UK office runs at 60-90% of peak demand from 8am to 6pm, Monday to Friday. Cooling load dominates summer demand and aligns almost perfectly with solar generation. IT and lighting baseload provides 24-hour minimum load that absorbs morning and afternoon generation. Hybrid working has slightly reduced peak demand but increased the daytime baseload across more days of the week (Tuesday-Thursday especially), strengthening self-consumption further.
The typical UK office without battery storage hits 70-85% self-consumption on a well-sized PV system. With a 200-400 kWh battery, that lifts to 90-95%.
By contrast, a typical UK warehouse runs at lower daytime demand (lighting and MHE only, no significant cooling load), often operates on 24-hour shift patterns where night-time consumption is the dominant load, and has minimal weekend demand. UK warehouses without battery typically hit 40-55% self-consumption — meaning roughly half of generation exports to SEG at low tariffs.
A retail building with no air-conditioning sits somewhere between — maybe 55-65% without battery. A factory with high process load and three-shift operation can hit 80%+ without battery but often has roof constraints that limit system size.
What this means for sizing
Higher self-consumption means a larger economically-optimal system. An office that can absorb 80% of generation can justify a system sized to 80% of annual demand. A warehouse that only absorbs 50% of generation should be sized to 50% of demand (or smaller) — anything larger pushes the additional generation into export at unfavourable tariffs.
This is why generic kWp-per-sqm-of-roof rules of thumb produce worse outcomes for offices than for warehouses. The office can typically support a larger system economically; the warehouse should typically be smaller.
Working out your specific self-consumption
The half-hourly meter data approach we use models self-consumption explicitly. For a given system size, we overlay 12 months of historic generation (from PVSyst model, building-specific) onto 12 months of historic consumption (from meter data, building-specific). The result is a half-hour-by-half-hour record of when the building self-consumes generation, when it exports, and when it still imports from grid.
That model is then used to project economics across the 25-year asset life — varying the SEG tariff assumption, the grid retail price escalation, and any future changes to building load (planned EV charging, heat pump retrofit, etc.).
The output is a specific self-consumption ratio for your building, your load shape, and your roof — not a generic industry assumption.
Tactics to lift self-consumption
For office buildings already on a good baseline, several measures lift self-consumption further:
Battery storage. 200-400 kWh covers the typical morning-peak / late-afternoon-peak shift. Lifts self-consumption by 10-20 percentage points.
Workplace EV charging. Daytime EV charging from solar generation creates additional Monday-Friday daytime load. 10-20 × 7-22kW chargers absorb meaningful generation in the spring/autumn shoulder season.
Building cooling pre-cooling. BMS strategies that run cooling slightly harder in summer afternoon (when generation is highest) can shift consumption into the generation window without affecting occupant comfort.
Process load timing. For offices with significant kitchen, server cooling, or other shiftable loads, scheduling these to align with generation peaks lifts self-consumption further.
We model these tactics in every proposal — the marginal economics are often surprisingly significant.
Request a free feasibility study modelled from your specific load profile.
Why office self-consumption is 75-88%: the load profile analysis
The 75-88% self-consumption figure for well-sized UK office solar systems is not an assumption — it follows directly from the overlap between the building’s electricity demand profile and the PV generation profile.
UK commercial offices have three distinct load components that drive daytime demand:
1. HVAC cooling load (30-50% of peak demand in summer). Office air conditioning systems run predominantly during occupied hours, typically 7am-7pm Monday to Friday. Cooling load correlates strongly with outdoor temperature — it peaks on hot summer afternoons precisely when PV generation is at its maximum. This near-perfect alignment is the most powerful factor in high office self-consumption.
2. IT and server room load (20-35% of baseload). Office IT infrastructure — servers, network equipment, UPS systems — runs continuously at approximately constant load 24 hours a day, 7 days a week. This baseload absorbs morning generation from approximately 8am as soon as panels start producing meaningfully, and continues absorbing generation through the evening. A typical 5,000 sqm office might have 40-80 kW of continuous IT baseload — at this level, a 300 kWp system is effectively continuously consumed for the first 4-5 hours of daily generation.
3. Daytime occupancy-driven loads (lighting, lifts, kitchen, printing). These loads are highly correlated with occupancy — strong on Tuesday to Thursday (typically the highest office occupancy days in hybrid-working patterns), lighter on Monday and Friday. These loads add 20-40% of peak demand during core hours and fall away in evenings and weekends.
The combination of these three load types means an office building maintains meaningful electricity demand for 12-14 hours on weekdays and 6-8 hours on weekends — covering essentially the entire solar generation window.
Contrast with warehouses: A distribution warehouse without significant refrigeration or automation typically runs at 10-30% of peak demand during daylight hours (lighting only, forklift charging), with peak demand on night shifts when solar generates nothing. Self-consumption naturally falls to 40-55% without battery.
Battery storage impact on self-consumption: the numbers
Battery storage lifts office self-consumption by approximately 8-15 percentage points, depending on baseline self-consumption and battery size relative to generation.
Worked example: 320 kWp office, Birmingham Without battery: Annual generation 294,000 kWh, self-consumption 79%, export 61,740 kWh With 200 kWh battery: Self-consumption lifts to approximately 87% (additional 23,520 kWh self-consumed), export drops to 38,220 kWh With 400 kWh battery: Self-consumption lifts to approximately 93% (additional 41,160 kWh), export drops to 20,580 kWh
Financial impact of battery addition: Battery capex: 200 kWh = 72,000 (at 360/kWh); 400 kWh = 144,000 Annual benefit increase from 79% to 87% self-consumption (at 30p avoided vs 9p SEG): 23,520 kWh x 21p = 4,939/year Simple payback on 200 kWh battery: 72,000 / 4,939 = 14.6 years Annual benefit increase from 79% to 93% self-consumption: 41,160 kWh x 21p = 8,644/year Simple payback on 400 kWh battery: 144,000 / 8,644 = 16.7 years
Battery payback is materially longer than system payback (14-17 years vs 4-6 years). The battery is economically justified only where:
- Grid electricity retail rate is significantly above 30p/kWh (e.g., 40p+, which is common on some unhedged commercial contracts)
- Battery also provides demand charge reduction or time-of-use arbitrage
- Battery also provides resilience value (business continuity) that has explicit financial value to the occupier
For most office solar projects, the battery is an optional enhancement rather than a required component for strong economics.
Battery storage impact on warehouse self-consumption: The warehouse without battery (40-55% self-consumption) can lift to 65-78% with appropriately sized battery that shifts midday generation to evening charging/refrigeration/shift-change load. The lift is larger in percentage points for warehouses (20-25pp) than offices (8-15pp) because the baseline is lower and there is more room to improve. This is one reason battery storage is more commonly bundled with warehouse solar than with office solar.
PVSyst modelling: what a proper model captures
PVSyst (and equivalents including SAM, PVLib, Helioscope) produces the half-hour-by-half-hour simulation that drives accurate self-consumption modelling. Understanding what the model includes helps evaluate proposals.
Irradiance inputs. PVSyst uses European Commission PVGIS meteorological data — hourly irradiance (global horizontal, diffuse, direct normal), temperature, and wind speed for the specific site latitude/longitude. The PVGIS CM-SAF dataset covers 1979-2023 and is updated annually. This provides a statistically robust irradiance baseline that accounts for UK seasonal variability.
Building consumption profile. This is the step most installers get wrong. A proper model uses actual half-hourly consumption data from 12 months of smart meter records, not a generic office profile. Generic office profiles (from CIBSE TM54 or similar) have standard deviations of 25-40% around actual measured profiles — meaning self-consumption predictions can be 10-20 percentage points wrong without real meter data.
We request half-hourly meter data from the building’s electricity supplier as the first step in feasibility assessment. The request takes 5-10 working days to fulfil (data is held by the DNO and accessed via the supplier). This data is then uploaded to PVSyst alongside the yield model to produce a half-hour-by-half-hour generation versus consumption overlay.
Self-consumption calculation. For each half-hour in the year, PVSyst calculates: if generation > consumption, excess is exported; if generation < consumption, deficit is imported. Self-consumption ratio is the fraction of total annual generation that is consumed on-site. The model explicitly accounts for overnight periods (zero generation), weekends (reduced demand), holiday periods (if specified), and seasonal demand variation.
Output metrics. A complete PVSyst report provides: annual yield (kWh), specific yield (kWh/kWp), self-consumption ratio (%), export (kWh), self-consumed (kWh), performance ratio, and a monthly breakdown of each metric. Any proposal lacking these specific outputs is not using a proper yield model.
The specific reasons offices outperform other building types
Across our installed base, the self-consumption ratio distribution by building type (all without battery, all systems 100-500 kWp):
| Building type | Typical self-consumption range | Key driver |
|---|---|---|
| Office (Mon-Fri HQ) | 75-88% | HVAC + IT baseload |
| Data centre | 88-97% | Continuous high load |
| Manufacturing (3-shift) | 78-88% | Process load overlap |
| Retail (shopping centre) | 65-78% | Extended hours, weekend load |
| School/college | 72-84% | Term-time daytime occupancy |
| Hotel | 68-78% | Variable occupancy, 24hr baseload |
| Warehouse (logistics) | 42-58% | Night shift dominance |
| Church | 22-38% | Very limited weekday load |
Offices sit in the second-best category (behind data centres and ahead of manufacturers), reflecting the Mon-Fri daytime load profile that most closely mirrors PV generation.
The specific characteristics that lift offices above the 75% typical lower bound:
- Chilled-beam or fan-coil cooling systems: always run during occupancy, aligning precisely with peak generation
- Dense IT tenancies (law firms, financial services, media): higher IT baseload, higher server room cooling load
- Retail-facing buildings with customer-facing zones: weekend opening hours lift weekend self-consumption
- Buildings with electric vehicle charging facilities: EV chargers absorb 10-50 kW of additional daytime load
Worked PVSyst modelling example
System: 240 kWp on 3,200 sqm flat roof, London (UKPN network). 4,000 sqm net floor area professional services office. 12-month half-hourly meter data provided.
Input parameters:
- Location: 51.5 N, PVGIS CM-SAF irradiance dataset
- Panel: 480 x LONGi Hi-MO 6 500W, east-west ballasted, 10 degree tilt
- Inverter: 2 x Huawei SUN2000-100KTL, DC-to-AC ratio 120%
- Shading: near-shading from 2m parapet modelled explicitly, 3.5% annual shading loss
PVSyst output:
- Gross annual yield (pre-shading): 216,480 kWh (902 kWh/kWp)
- Net annual yield (post all losses including shading, temperature, cabling): 200,880 kWh (837 kWh/kWp)
- Performance ratio: 80.4%
- Modelled self-consumption (from actual meter data overlay): 81.5% (163,717 kWh/year self-consumed)
- Modelled export: 37,163 kWh/year
Financial output (at 32p/kWh avoided, 10p SEG):
- Annual electricity saving: 163,717 x 32p = 52,389/year
- Annual SEG income: 37,163 x 10p = 3,716/year
- Total annual benefit: 56,105/year
- System capex: 204,000 (at 850/kWp)
- Simple payback: 3.6 years
- 25-year NPV at 8% WACC: 1.04 million
This is a representative London office result — shorter payback than UK-average due to higher London electricity rates, slightly lower yield than South West England due to latitude (London at 51.5N vs Bristol at 51.4N, similar but London has slightly more shading typically).
Key takeaways
- Office self-consumption is 75-88% without battery because HVAC cooling load, IT baseload, and occupancy-driven loads collectively cover the full PV generation window on weekdays
- Battery storage lifts office self-consumption by 8-15 percentage points but has a 14-17 year standalone payback — economically justified primarily where tariff rates are high or resilience value is explicit
- Proper self-consumption modelling requires actual half-hourly meter data from 12 months of consumption — generic profiles introduce 10-20pp error into self-consumption predictions
- Data centres are the only building type consistently outperforming offices on self-consumption (88-97%); warehouses are dramatically lower (42-58%)
- A complete PVSyst report includes annual yield, specific yield, self-consumption ratio, export, performance ratio, and monthly breakdown of all metrics