How many solar panels do I actually need?

Updated 2026-11-069 min readSolar Math

Almost every answer you find to this question is wrong for you, because almost every answer quietly assumes a location. "A typical home needs 20 panels" is true in Birmingham and absurd in Abu Dhabi. The honest version of the answer has four steps, and you can do all of them on the back of an envelope.

Step one: find your real consumption

Panels are sized against kilowatt-hours, not against money. Your bill is money, and money is contaminated by fixed charges, taxes, meter rent and slab pricing. So find the units.

Look on your bill for a line reading units consumed, kWh, or a present-minus-previous meter reading. That number is what you want. Take twelve months of them and average, because sizing against a single July bill in a hot country will leave you with an array that is gigantic and idle for eight months of the year.

If you genuinely cannot find the units, divide your bill by your tariff — but be aware that in countries with steep progressive slabs, dividing the total bill by the average rate understates your consumption, because the top of your bill is being charged at a much higher rate than the bottom.

A household using 800 units a month is burning about 26 kWh a day. Hold on to that number.

Step two: find out how much sun you get

This is the step that everyone skips, and it is the one that moves the answer most. The quantity you need is peak sun hours — the number of hours per day for which the sun would have to shine at exactly 1,000 W/m² to deliver the same total energy your location actually receives. It is numerically identical to the daily solar irradiation in kWh/m²/day, which is why the two terms get used interchangeably.

Indicative ranges for horizontal surfaces. Tilting the panels towards the equator adds roughly 2–5% near the tropics and up to 25% at high latitudes — which is why tilt matters far more in Berlin than in Dubai.

The calculator pulls this number from NASA's satellite record for your exact coordinates instead of making you guess from a table.

Step three: apply the losses

A panel rated 585 W will never produce 585 W on your roof. Between the datasheet and your meter you lose energy to heat, dust, wiring resistance, inverter conversion, mismatch between panels and occasional shade. Bundled together these are called the performance ratio.

• 0.80 – 0.84 — a clean, well-ventilated, unshaded install in a temperate climate
• 0.72 – 0.78 — a typical hot, dusty rooftop with a sensible wash schedule
• 0.62 – 0.70 — flush-mounted on a hot roof with no airflow, washed twice a year, some shade

Heat is the biggest single component and the one people underestimate. Crystalline silicon loses about 0.4% of its output for every degree the cell sits above 25 °C, and on a still 40 °C day a roof-mounted panel runs at 65 °C or more. That is a 16% haircut before you have counted anything else. There is a full breakdown of every loss here.

Step four: do the division

Our 26 kWh/day household in Karachi, with 5.6 peak sun hours and a performance ratio of 0.75:

The same household in Manchester, with 2.7 peak sun hours and a better performance ratio of 0.82 because it is cooler:

Nearly double the array, for the same house and the same appliances. This is the whole reason a generic panel count is useless.

Turning kWp into panels

At 585 W per panel, 6.2 kWp means 6,200 ÷ 585 = 10.6, so eleven panels. At 450 W it is fourteen. Always round up: you cannot install two-thirds of a panel, and strings have to be built from whole modules anyway.

Will they fit?

A rough planning figure is 5 m² of roof per kWp for modern panels on a pitched roof. Flat roofs need roughly 1.7 times that, because you tilt the panels up on frames and then have to space the rows apart so the front row does not shade the one behind it in winter.

Measure the usable roof, not the roof. Water tanks, vents, stair housings, satellite dishes and the walkway you need to clean the things all come off the total first.

Should you size for 100% of your bill?

Not automatically. Three things push the right answer away from exactly 100%:

• Export is worth less than consumption. Wherever you are credited less for exported units than you pay for imported ones, every unit you generate beyond what you use yourself is worth a fraction of its face value. That argues for a smaller array.
• Slab tariffs. If your tariff steps up sharply above a threshold, the first units of solar you generate are displacing your most expensive units. An array sized to knock you down a slab can have a dramatically better return than one sized to zero the bill.
• Your consumption will grow. An electric vehicle, another air conditioner, or a child moving back in will all raise it. Adding capacity later means a second visit, possibly a second inverter, and a worse price per watt. Many installers suggest building in 10–20% of headroom.

A worked example, start to finish

A house in Lahore using 1,100 units a month, pitched roof facing roughly south, moderate dust, no significant shade:

• Daily consumption: 1,100 ÷ 30.4 = 36.2 kWh/day
• Peak sun hours: 5.2
• Performance ratio: 0.75
• Array: 36.2 ÷ (5.2 × 0.75) = 9.3 kWp
• Panels: 9,300 ÷ 585 = 16 panels (giving 9.4 kWp after rounding)
• Roof needed: about 47 m²
• Inverter: 9.4 ÷ 1.2 = 7.8 kW, so an 8 kW unit — unless backup matters, in which case see the inverter sizing guide

Checking an installer's quote with this

Run your own numbers before the sales visit, then ask two questions. First: what annual generation are you promising, in kWh? Second: which irradiance dataset is that based on? A quote promising a specific yield far above what this method gives you is either using optimistic losses or a sunnier location than yours. Neither is fraud, exactly, but you will be the one living with the shortfall.

Run these numbers for your own roof →