Self-Consumption and Load Shifting: The Number That Decides Your Solar Economics

Annual production is the figure every quote leads with. It is also the figure that matters least. What decides the value of a rooftop array is the share of those kilowatt-hours consumed inside the building rather than pushed onto the network — the self-consumption ratio.

The arithmetic is blunt. If you buy electricity at 0.28 EUR/kWh and are paid 0.07 EUR/kWh for exports, a kilowatt-hour used at home is worth four times a kilowatt-hour exported. Two households with the same 5 kWp array, one at 30 percent self-consumption and one at 60 percent, do not differ by a few euros a year. They differ by roughly 40 percent of the total benefit.

Defining the two ratios people confuse

Two ratios sound alike and mean opposite things.

Self-consumption ratio = solar energy used on site ÷ total solar generation. It answers: how much of what I make do I keep?

Self-sufficiency ratio (also called autonomy) = solar energy used on site ÷ total household consumption. It answers: how much of what I need do I make?

A large array on a small load has high self-sufficiency and poor self-consumption. A small array on a large load has the reverse. Only the first ratio drives the value of each additional kilowatt-hour generated; the second describes exposure to the grid. A quote that quotes one and implies the other is being careless.

Figures are indicative for a central European load profile with daytime occupancy on two days a week. Your own profile can move each column by ten points or more.

Why the default is around a third

Residential demand is bimodal: a morning peak before people leave, a larger evening peak after they return. Solar generation is unimodal and centred on solar noon. The two curves barely overlap. Standing loads — refrigeration, standby, ventilation, a router, a heat pump in shoulder season — form a baseline of perhaps 150 to 400 W that solar covers easily. Everything above that baseline happens at the wrong time.

That is why an unmodified household with no storage and no controllable loads lands between 25 and 35 percent almost regardless of country. It is a shape mismatch, not a hardware defect.

Measuring your own profile before you spend anything

Do not accept a modelled ratio. Get the data.

1. Ask your supplier for half-hourly or quarter-hourly meter data. In most of the EU, smart meter data is yours to request under the Electricity Market Directive and is usually downloadable as CSV.
2. If no smart meter exists, clamp a logger on the incoming tails for two to four weeks covering a normal work pattern. A 50 EUR CT logger is enough.
3. Model the array separately with PVGIS or PVWatts at hourly resolution for your tilt and azimuth.
4. Overlay the two series hour by hour and sum min(generation, load) for each interval. Divide by total generation. That is your true self-consumption ratio with no battery.

The spreadsheet takes an hour and is the single most useful piece of analysis a prospective owner can do. Everything after this — battery size, array size, tariff choice — falls out of it.

Load shifting: the cheapest intervention

Before buying storage, move the load. Shifting a kilowatt-hour costs nothing but attention; storing one costs the battery.

• Water heating. An immersion or heat-pump cylinder is a thermal battery with a round-trip cost of zero capital if it already exists. A simple solar diverter that modulates surplus into the element typically adds 8 to 15 points of self-consumption in a household with an electric cylinder.
• Dishwashers and washing machines. Delay timers are standard. Two loads a week moved to midday is roughly 150 kWh a year of otherwise exported energy.
• Electric vehicles. The largest single lever available. A car charging at 3.7 kW absorbs almost any residential array's midday surplus. Solar-aware chargers modulate current to track available surplus rather than switching in fixed steps.
• Heat pumps. Pre-heating or pre-cooling the building fabric in the middle of the day exploits thermal mass. In a well-insulated dwelling a two-degree setpoint shift can absorb several kilowatt-hours without noticeable discomfort.
• Pool pumps, dehumidifiers, ventilation. Anything with a duty cycle rather than a fixed schedule should run when the sun is up.

Combined, these interventions routinely take a household from the low thirties to the high forties without any storage. That is the same improvement a small battery buys, at a fraction of the cost.

Where storage genuinely earns its place

Storage is not a way to raise self-consumption cheaply. It is a way to raise it further once the free options are exhausted, and to buy resilience and tariff arbitrage alongside.

The marginal value of each stored kilowatt-hour is the import price minus the export price, minus round-trip losses of roughly 10 to 15 percent, minus the amortised cost of the cycle. With a battery at 600 EUR/kWh usable, 6,000 cycles of warranted throughput and 90 percent round-trip efficiency, the storage cost per delivered kilowatt-hour is about 0.11 EUR before any financing. If your import-export spread is 0.21 EUR, storage clears the bar. If your spread is 0.09 EUR, it does not, and no amount of enthusiasm changes that.

Battery size follows the surplus curve, not the array size. Sum the daily exported energy across the year and look at the median, not the maximum. Sizing to the summer peak buys capacity that sits idle for eight months. Our battery sizing calculator works this through with a full year of daily surpluses.

Time-of-use tariffs change the answer

Where dynamic or time-of-use pricing exists, the value of a stored kilowatt-hour is no longer just the avoided import price. It becomes the difference between the cheapest and dearest hours, which in volatile markets can exceed the flat spread by a wide margin. Under those tariffs a battery may profitably charge from the grid overnight and discharge into the evening peak, quite apart from what the array does.

This also inverts one piece of conventional advice. Under a flat tariff, west-facing modules are penalised for lower annual yield. Under an evening-peaked tariff, their later generation profile may be worth more per kilowatt-hour than a south-facing array's midday surplus. Orientation should be chosen against the tariff, not against a yield table alone. The trade-off is set out in detail in orientation and tilt.

Common errors

• Sizing storage from array size. A "5 kWp needs 5 kWh" rule of thumb has no analytical basis.
• Counting the battery's rated capacity. Usable capacity after depth-of-discharge limits and reserve is typically 85 to 95 percent of nameplate.
• Ignoring standby draw. Hybrid inverters and battery management systems consume 20 to 60 W continuously. Over a year that is 175 to 525 kWh — a meaningful bite out of a small system's benefit.
• Assuming the ratio is static. Adding a heat pump or an EV can raise self-consumption by twenty points overnight and makes a previously marginal array look excellent.
• Modelling on monthly totals. Self-consumption is an hourly phenomenon. Monthly arithmetic hides the mismatch entirely and always flatters the result.

Frequently asked questions

What is a good self-consumption ratio? For a household without storage or controllable loads, 30 to 35 percent is normal and not a sign of a bad installation. With load shifting, 45 to 55 percent is achievable. With a well-sized battery, 60 to 75 percent. Above 80 percent usually means the array is undersized relative to the load rather than that the system is unusually good.

Does a bigger battery always raise self-consumption? It rises, but with sharply diminishing returns. Each additional kilowatt-hour of storage catches a smaller slice of surplus because the large surpluses occur on relatively few days. Beyond roughly one day's average household consumption, added capacity earns very little unless it is being used for tariff arbitrage or backup.

Is exporting to the grid a waste? No. Exported energy displaces generation elsewhere and still earns whatever the export rate is. It is simply worth less than energy used on site under most modern compensation schemes, which is why raising self-consumption is the priority when the two prices diverge.

Should I oversize the array if self-consumption will fall? Often yes. Fixed costs — scaffolding, design, connection, labour — are spread over more kilowatts as the array grows, so the cost per kilowatt-hour generated falls even as the average value of each kilowatt-hour falls. Run both effects through the payback calculator rather than reasoning about either in isolation.

References

• European Commission, PVGIS hourly radiation and PV performance data
• IEA PVPS Task 17, Trends in Photovoltaic Applications
• Fraunhofer ISE, Recent Facts about Photovoltaics in Germany
• Directive (EU) 2019/944 on common rules for the internal market for electricity, Article 23 on data management