Bills & Tariffs
Chasing a self-consumption percentage can cost more than it saves
Self-consumption is a diagnostic ratio that got promoted to a goal, and households distort their lives to raise a number that is not the thing being optimised.
By Manish Trivedi3 min read
A ratio that became an objective
Monitoring apps display self-consumption prominently: the share of generated electricity that was used on site rather than exported. It is a useful diagnostic, because it tells you something real about how well an array matches a household. It is not a goal, and treating it as one produces some strange behaviour.
The confusion is understandable. A high percentage feels like efficiency and a low one feels like waste, and the language of the interface encourages that reading. But nothing is wasted when a unit is exported. It goes somewhere, it does work, and in most arrangements it is paid for, however modestly.
What the household is actually trying to reduce is the money leaving the account, and that quantity is set by units imported at the import price, less whatever export earns. Self-consumption is one input to that calculation and it can be raised in ways that make the calculation worse.
How the ratio can be improved by making things worse
The simplest way to raise self-consumption is to consume more. Run the tumble dryer at midday when the washing could have gone on a line. Heat water you were not going to use. Leave something on because the app shows a surplus. Every one of those raises the percentage and none of them saves anything.
A subtler version is downsizing the array. A smaller system exports less, so its self-consumption ratio is higher, and it also generates less and saves less. Any metric that improves when you shrink the thing producing the benefit should be treated with suspicion.
The version that costs most is buying hardware to raise the number. A battery bought principally to stop exporting, in a household where export is reasonably paid and the import-export gap is narrow, will raise self-consumption impressively and may not repay itself. That decision should be made against the gap, not against the ratio.
When self-consumption genuinely is the strategy
There is a class of arrangement where the ratio and the objective almost coincide. Where export earns nothing at all, or where the meter simply refuses to record it, every exported unit is genuinely given away, and moving consumption under the generation curve is the only mechanism available.
In that situation the priorities invert cleanly. Shifting flexible loads into daylight becomes valuable for its own sake. A diverter or a battery has an easier case to make. And sizing the array conservatively becomes defensible, because capacity that only ever exports is capacity that only ever donates.
The reverse case exists too. Where export is paid at a rate near the import price, self-consumption barely matters financially, and the sensible strategy is to generate as much as the roof allows and let the meter sort it out. Two households in different countries can be given directly opposite advice and both be right.
Load shifting is where the free money is
Between those extremes sits the measure that costs nothing and works nearly everywhere: moving a flexible load from a time when the house is importing to a time when it is generating. The dishwasher, the washing machine, the car, the pool pump, the immersion, anything with a timer.
The value is the full difference between import and export prices, per unit moved, and it involves no equipment and no compromise beyond remembering to set a timer. Where a load can be shifted, it should be, and that is true whether the household cares about the percentage or not.
The limits are practical rather than technical. Not every load is flexible, most people are not home to supervise appliances mid-afternoon, and there are sensible reasons not to run a dryer unattended. Automation helps; so does simply establishing a habit for the two or three appliances that account for most of it.
What to look at instead
The number worth watching is imported units, tracked over comparable periods. That is the quantity the household is paying for, it responds to everything that actually matters, and it cannot be improved by wasting energy.
Alongside it, exported units are worth knowing simply to understand what the array is doing. If exports are large and export income is negligible, there is a genuine opportunity to shift or store. If exports are small, there is no surplus to capture and any hardware sold to capture one is solving a problem the household does not have.
Self-consumption still has its place as a description. It answers the question of how well the system fits the house, and a very low figure is a signal worth investigating. Just don’t optimise it. Optimise the bill, and let the ratio land wherever it lands.
Common questions
What is a normal self-consumption figure?
It varies so widely with household pattern, array size and whether storage is fitted that a typical figure is not useful guidance. A house that is empty on weekdays with a generously sized array sits low; a house with people at home, a car charging and a battery sits far higher. Compare against your own history rather than against a benchmark.
Does a battery always raise self-consumption?
Yes, mechanically, because it absorbs surplus that would have been exported and returns it later. Whether the raise is worth the cost depends entirely on the gap between import and export prices and on how often the battery cycles. The percentage rising is not evidence that the purchase was sound.
Should I turn appliances on just to use free electricity?
Only if you were going to use them anyway and are moving the timing. Electricity from the roof is not free in the sense of being worthless; in most arrangements it could have been exported for something. Consuming it for no purpose converts a small payment into nothing.
Deputy editor, Power Your Roof
Manish has been reporting on solar basics, batteries, bills & tariffs since long before it was fashionable and would rather show the working than assert the conclusion.
