Why this decides your whole estimate
Coats multiply. Every other input in a paint estimate — area, coverage rate, waste — moves the total by a few percent. Coats move it by 50% or 100% in one step. Getting the area wrong by half a metre costs you nothing; assuming two coats when the wall needs three means a third of the job is unpainted when the last tin runs dry.
litres = net area × coats ÷ coverage per litre
Because coats sit in the numerator, the decision is worth more thought than any other line. It is also the one input people copy blindly from the tin.
What a coat is actually for
A coat of paint does two separate jobs, and confusing them is why the advice is contradictory. The first job is opacity: hiding whatever is underneath. The second is film build: laying down enough thickness that the surface is durable, washable and even in sheen. A wall can be fully opaque after one coat and still look patchy, because the film is too thin in places to reflect light consistently.
This is why professionals apply two coats to a wall that already looks covered. They are not hiding anything at that point — they are building a uniform film. Skip it and the wall will show roller laps under raking light for as long as it exists.
The number by situation
| Situation | Coats | Notes |
|---|---|---|
| Repaint, same colour, sound surface | 1 | Refresh only. Works if the existing finish is clean, matt and undamaged |
| Repaint, similar shade, one or two steps lighter or darker | 2 | The default. Assume this unless you have a reason not to |
| White or light over a mid colour | 2 | With a good opaque paint. Cheap trade white will need three |
| Light over a strong colour — red, teal, navy, black | 3 | Or primer plus two. Primer is usually cheaper than a third coat of finish |
| Any colour over new plaster | mist coat + 2 | The mist coat is diluted paint, not primer, and soaks in almost entirely |
| Strong colour over white | 2-3 | Reds, yellows and oranges have the weakest pigment opacity of all |
| Over a stain — water, nicotine, marker, soot | stain block + 2 | No number of finish coats will hold these back permanently |
| Ceiling, white on white | 2 | One coat on a ceiling always shows, because ceilings get the worst raking light in the house |
| Bare wood, trim and doors | primer + 2 | Three thin coats beat two thick ones on trim |
| Bare metal radiator or railing | primer + 2 | Use a primer matched to the metal or it will lift |
The colour change rule
Opacity is a property of the pigment, not the price. Some pigments simply cannot hide. Ranked from most to least opaque: black and dark grey, white with titanium dioxide, earth tones, blues and greens, then reds, yellows and oranges at the bottom. A bright yellow over white can genuinely need four coats, and no amount of paying more will fix it — the fix is a grey or tinted primer underneath.
- Going darker
- Usually easy. Two coats will cover almost anything if the new colour is meaningfully darker. This is the cheap direction.
- Going lighter
- The expensive direction. The old colour bleeds through the wet film and keeps bleeding until the film is thick enough. Budget three coats, or one tinted primer plus two.
- Same value, different hue
- Deceptive. A mid green over a mid terracotta looks like it should cover in one, and does not, because the two colours cancel unevenly. Two coats minimum, and check after the first under artificial light.
- Anything to bright red or yellow
- Buy a tinted primer in a similar tone. Three finish coats of an expensive red costs far more than one cheap primer plus two.
What the surface does
- New plaster drinks. The first application is a mist coat — paint diluted with 10-20% water — and it does not count as a coat, because most of it disappears into the wall. Estimate it separately as roughly one extra coat of material.
- Old chalky emulsion drinks too, and unevenly. If a dry cloth comes away white, the wall needs sealing before anything else.
- Bare filler and patch repairs absorb far more than the wall around them and will flash as dull spots through two coats. Spot-prime the repairs.
- Textured surfaces — woodchip, artex, heavy render — have far more actual surface area than their measured area. Add a coat, or 30% material, whichever is easier to think about.
- Glossy or previously oil-painted surfaces do not need more coats; they need degreasing and abrading, or the paint will peel off in sheets regardless of how many layers you apply.
Applying more coats never fixes an adhesion problem. If the surface is greasy, chalky or glossy, extra layers just make a thicker thing to peel off later.
A worked example
A living room, 5 × 4 m, 2.7 m high, going from a dark navy to an off-white. One door and two windows, 5.4 m² of openings in total. The paint claims 12 m² per litre.
walls = 2 × (5 + 4) × 2.7 = 48.6 m²
net area = 48.6 − 5.4 = 43.2 m²
two coats: 43.2 × 2 ÷ 12 = 7.2 L
three coats: 43.2 × 3 ÷ 12 = 10.8 L
The difference is 3.6 litres — a whole extra 5 L tin, and in this case almost certainly needed, because light over navy is the hardest direction there is. The cheaper route is a tinted grey primer at 43.2 ÷ 12 = 3.6 L plus two finish coats at 7.2 L. Same total litres, but primer costs less per litre than a good finish paint, and the result is more even.
Note what did not matter: whether the room was 5 × 4 or 5.1 × 4. That is a 1 m² difference, under 0.2 litres. The coats decision was twenty times more consequential.
Deciding without guessing
- Buy one tester pot and paint a square metre of the actual wall, in the actual room. Let it dry fully — paint changes opacity as it dries — and look at it at night under the lights you use.
- If the patch is solid after two test coats, two is your number. If the old colour still ghosts, add a primer rather than a third finish coat.
- Paint the least visible wall first. You will get an honest answer about coverage before you commit to the wall behind the sofa.
- Buy the paint for all coats in one purchase and one batch. Tinted colours vary between mixes, and a third tin bought a week later can be visibly different.
- Round up to whole tins at the end, not per coat. Rounding each coat separately buys you tins you do not need.
Coverage figures printed on tins are measured on smooth, sealed, primed surfaces under laboratory conditions. Real walls return 10-20% less. The coats decision and the coverage figure both push the same way, so do not optimise both.