What is an area-weighted U-value?

A U-value measures how quickly heat passes through part of a building: the watts lost through each square metre for every degree of temperature difference between inside and out (W/m²K). The lower the U-value, the better the insulation.

Real building elements are rarely uniform. A wall might have a thinner build-up where it meets the existing house; an extension has walls, a floor, a roof and glazing, all with different U-values. An area-weighted U-value combines several U-values into one figure, giving each one the weight of the area it covers. In England, Wales and Northern Ireland, it is the basis of the main way to show that an extension complies when it doesn't meet the standard approach, usually because it has more glazing than the standard approach allows.

The calculation

Multiply each U-value by its area, add the results together, then divide by the total area:

Area-weighted U-value = (U₁ × A₁ + U₂ × A₂ + U₃ × A₃ + …) ÷ (A₁ + A₂ + A₃ + …)

England, Wales and Northern Ireland all set out this formula in their guidance (Approved Document L Volume 1, paragraph 10.9; the Welsh Approved Document L Volume 1, paragraph 13.4; and Technical Booklet F1, paragraph 3.14).

As an example, take an extension wall that is mostly built to 0.15 W/m²K, with a 5 m² section at 0.25 W/m²K:

4.25 ÷ 25.0 = 0.17 W/m²K. The thinner section is well above that, but the wall as a whole averages 0.17 W/m²K.

This matters because limiting U-values for walls, floors and roofs are area-weighted averages. In England, the limits for new elements in an extension are in Table 4.2 of Approved Document L Volume 1 (paragraphs 4.7 and 10.7), and its note 1 says they are "area-weighted average values, except for windows, doors, roof windows and rooflights". So one part of a wall can perform worse than the limit, as long as the wall as a whole meets it. The example wall's 0.17 W/m²K is within England's 0.18 W/m²K limit for walls.

Comparing a whole extension with a benchmark

When an extension doesn't meet the standard approach, the guidance lets you show that it performs no worse, as a whole, than a version of itself that does. In England this is the area-weighted U-value approach (paragraph 10.9), in Wales the U-value trade-off approach (Section 13), and in Northern Ireland the Calculated Trade-off Approach (paragraph 3.14). They work the same way:

1. Set up a benchmark extension. It has the same size and shape as the one you are proposing, and meets the standard approach: the limiting U-values, and windows, doors and rooflights no larger than 25% of the extension's floor area plus the area of any openings lost to the extension.
2. Work out the benchmark's area-weighted U-value across all its elements: floor, walls, roof and openings.
3. Work out the proposed extension's area-weighted U-value in the same way, using the U-values and areas you actually propose.
4. Compare. The proposed extension's area-weighted U-value must be no greater than the benchmark's.

Because the benchmark is the same size and shape, both extensions have the same total area. So comparing their area-weighted U-values gives the same answer as comparing their total heat loss (the U × A totals).

A worked example

This example uses England's Table 4.2 values for the benchmark. A single-storey rear extension has an 18.0 m² floor, 30.0 m² of external wall (including openings) and an 18.0 m² roof. It removes an existing 1.5 m² door. The design has 8.0 m² of windows and doors, which is more than the allowance of 25% of 18.0 m² (4.5 m²) plus the 1.5 m² lost, or 6.0 m².

Benchmark extension (Table 4.2 values, openings capped at 6.0 m²):

Area-weighted U-value: 18.66 ÷ 66.0 = 0.283 W/m²K

Proposed extension:

Area-weighted U-value: 17.22 ÷ 66.0 = 0.261 W/m²K

At 0.261 W/m²K against the benchmark's 0.283 W/m²K, the proposed extension complies. Better insulation to the floor, walls and roof, and better-performing glazing, make up for the extra 2.0 m² of glazing.

Mistakes to avoid

• Benchmarking against the existing house. The benchmark is a compliant version of the new work, not the existing dwelling. The Welsh guidance says outright that using the existing dwelling's average U-value as the benchmark "does not demonstrate compliance" (paragraph 13.4).
• Pushing one element too far. Averaging lets part of the extension perform worse, not any part be as poor as you like. In Wales, for example, no wall, floor or roof may be worse than the limiting values in column (b) of Table 11.1 (paragraph 13.9).
• Forgetting the openings you lose. The allowance includes the area of existing windows and doors that no longer exist, or are no longer exposed, because of the extension. Leaving them out makes the benchmark harder to meet than it should be.
• Giving the benchmark the wrong area of openings. In Wales, if your design is within the allowance, the benchmark has the same area of openings as your design; it's only capped at the allowance when your design exceeds it (paragraph 13.8).

If you're working in Scotland, the method is different: see What is a compensatory heat loss calculation, and when do you need one?

Sources

This guide refers to the editions in force in September 2026:

• England: Approved Document L, Volume 1: Dwellings (2021 edition incorporating 2023 amendments), paragraphs 4.7, 10.7 and 10.9, and Table 4.2. A 2026 edition applies from 24 March 2027.
• Wales: Approved Document L, Volume 1: Dwellings (2022 edition), paragraphs 10.11 and 13.3 to 13.9. A 2026 edition applies from 4 March 2027.
• Northern Ireland: Technical Booklet F1 (2022), paragraphs 3.11 and 3.14.