Insulation usually beats generation on cost per unit saved

The comparison that ought to be made first

Any energy measure can be reduced to one figure: what it costs, divided by the units of energy it saves or generates over its working life. Doing that arithmetic consistently is the only way to compare a loft insulation job against an array of panels, and it is very rarely done because the two are sold by different trades to different parts of the mind.

When it is done, fabric measures usually come out ahead, and often by a wide margin. Insulation has no moving parts, no electronics, no inverter to replace and no degradation curve. It works for the life of the building, it requires nothing, and its cost is entirely up front.

That is a general finding rather than a universal one, and it depends heavily on how much insulation the building already has and on the climate it sits in. A well-insulated modern house in a mild climate has very little left to gain. An older one in a cold climate has a great deal.

A saved unit is worth more than a generated one

There is a second reason to look at demand before supply, and it is commercial rather than physical. A unit generated on the roof is worth the import rate if the house happens to be using it and the export rate if it is not. A unit never consumed is always worth the full import rate, because it was an import that did not happen.

Timing sharpens this further. Heating demand is concentrated in winter and in the evening, which is precisely when a rooftop array produces least, so insulation is reducing consumption in the hours when generation cannot help. The two measures act on different parts of the year and the fabric one acts on the expensive part.

This is why the sequencing advice — reduce demand, then generate, then consider storing — is not merely conventional wisdom. It follows from the arithmetic, and it holds under most tariff structures, though the details differ by country and by what the local export arrangement pays.

Where the heat actually leaves

Heat escapes a dwelling by conduction through the fabric and by air moving out of it, and the proportions differ enormously between buildings. In an older house with solid walls, single glazing and an uninsulated roof, conduction dominates. In a newer, well-insulated one, ventilation and air leakage become the larger share simply because the fabric losses have been reduced so much.

Within the fabric, the roof is generally the first target and the cheapest, because heat rises, loft space is usually accessible, and laying insulation is unskilled work. Walls hold a larger area and are harder to treat. Floors matter more than people expect in older properties with suspended timber over a ventilated void. Windows are conspicuous, emotionally significant and rarely the best value per unit saved.

The useful order is therefore roughly: find out what the building already has, treat the accessible and cheap first, and resist starting with the measure that is most visible.

Diminishing returns are steeper than they look

Heat flow through a layer is inversely proportional to its thermal resistance, and that reciprocal relationship produces a curve that surprises people. Going from no insulation to a modest layer removes a large fraction of the loss. Doubling that layer removes a large fraction of what remains, which is a much smaller quantity. Doubling again removes a fraction of a fraction.

Take an illustrative case with round numbers. If an uninsulated element loses ten units and a first layer of insulation cuts that to two, then adding a second identical layer might bring it to one and a half. The first layer saved eight units; the second saved half a unit for the same material cost.

The consequence is that topping up existing insulation is a far weaker measure than installing the first layer, and that a proposal to add depth to a loft that already has a reasonable amount deserves the arithmetic rather than the assumption. More is better remains true. Much better stops being true quite early.

The measures that do not pay, and saying so

Some fabric work simply does not recover its cost in energy terms, and it is worth being direct about which. Internal wall insulation in an occupied house is disruptive, expensive, reduces room dimensions and carries real risk of creating condensation problems if the detailing is wrong. It can be the right answer where there is no alternative, and it is not a value measure.

Replacing functional double glazing with a higher-specification unit rarely repays the cost in saved energy alone within any sensible horizon. Replacing single glazing frequently does. The distinction is between a first improvement and an incremental one, and it is the same reciprocal curve at work.

There is also the case for doing something for reasons that are not energy. Comfort, draught elimination, reduced noise and the elimination of cold surfaces that attract condensation are all real benefits, and people are entitled to buy them. The mistake is buying them under the impression that they are an investment, and finding out afterwards that the arithmetic never supported it.