North-facing roofs: when solar is actually worth it

The figures

Output as a percentage of the same location’s south-facing 35° figure, averaged across all 25 reference locations.

The pattern is the reverse of a south roof, where steeper helps up to a point. On the north side every degree of tilt points the panel further from where the sun actually is, so the penalty compounds.

The full matrix for every orientation is on roof orientation and pitch.

Why flat recovers so much

A flat panel sees the whole sky dome. Roughly half of the sunlight reaching the UK over a year arrives as diffuse radiation from clouds and haze rather than as a direct beam, and diffuse light arrives from all directions. A horizontal surface collects it evenly regardless of which way the building faces.

Tilt the panel north and you trade sky for ground: less of the dome, and the part you lose includes the southern sector where the direct beam always is.

That is a genuine option on some buildings: a north-facing flat roof, a large outbuilding, a garage, a shallow-pitched extension. It is not an option on a steep north-facing slate roof, where the array follows the roof plane and 54% is what you get.

Where a north roof is genuinely worth using

When there is no alternative plane and the alternative is nothing. 54% of something beats 100% of a roof you cannot use. This is the strongest case and the one the blanket rule ignores entirely.

When it extends an array you are already installing. Scaffolding, inverter, isolators, cabling, notification and labour are largely paid once. Adding panels to a second plane during the same visit does not repeat all of that. So the marginal cost per extra kWh from a north plane is lower than the average cost of the system suggests — and it is the marginal comparison that decides whether the extra panels are worth it.

When the surface is shallow or flat. At 83% a near-horizontal north-facing surface is barely distinguishable from an east or west roof at 79%, and nobody disputes those.

When your consumption is high and steady. If you use nearly everything you generate, output is worth roughly your import rate rather than roughly your export rate — six times more per unit. That makes a lower-yielding plane much easier to justify. If you export most of it at 4.1p, it is much harder.

When the roof is large and cheap to work on. Area partly substitutes for orientation. A big north-facing barn or garage roof at 54% can out-generate a small south-facing dormer at 100%.

Where it is not worth it

Instead of a usable south plane. Fill the south roof first. Always.

On a steep pitch, when you are capacity-constrained anyway. If your budget or your DNO limit caps you below what the south and east–west planes can carry, north panels displace better-placed ones.

When the array would be shaded as well. A shaded north roof is the worst case available and the arithmetic gets bleak quickly.

Where snow matters. A north-facing pitch gets less direct sun to clear snow and can stay covered longer after a fall — a marginal effect in most of the UK, a real one in upland and northern areas, and one we have not quantified.

On the basis of a payback figure from a lead-generation site. These systems are outside the range most calculators are built for, and a tool that assumes a south-facing array will not tell you honestly what a north one does.

Why the received wisdom has not moved

“Never north” is not stupid. It was correct advice, and it dates from a period when two things were true that are no longer true.

Panels dominated the cost. When modules were the largest line in a quote, output per panel was close to output per pound, and a 46% penalty per panel was close to a 46% penalty on the investment. As module prices fell, the fixed costs — scaffolding, labour, inverter, electrical work, certification — became a larger share, and the per-panel penalty became a weaker guide to the per-pound one.

The Feed-in Tariff paid generously per unit generated. Under a per-kWh subsidy, maximising generation was the objective and lower-yielding surfaces were plainly inferior. Under the SEG , exported units earn a fraction of what used units save, so the objective shifted from maximising generation to matching generation to consumption. That change does not make north roofs good, but it does weaken the specific argument that used to rule them out.

Practical points if you go ahead

Get the modelled figure in writing, with location, pitch, orientation and shading factor stated. A north-facing array is exactly the case where a generic estimate is most likely to be wrong.

Check the orientation properly. “North-facing” is often 135° rather than 180°, and 62% is a materially different proposition from 54%.

Ask about string layout. Panels on planes of different orientation should generally not share a string, because the array follows the weakest contributor. Separate strings, a multi-string inverter, Power optimiser devices or Microinverter units all solve it, and which one is appropriate is a design question for the installer.

Expect to be talked out of it. Some installers decline north-facing work as a matter of policy, partly to avoid complaints about output. Ask for the modelled number and decide for yourself, rather than treating a refusal as a physical limit.

Planning is unchanged. Solar on a house is normally permitted development subject to conditions on how far the panels project and where they sit, with tighter rules in conservation areas and on listed buildings. Orientation is not a planning consideration, so a north installation faces no additional hurdle.