How much of a coating a surface drinks in determines how much pigment ends up sitting on top of it. A porous surface — new plaster, bare timber, untreated masonry — pulls binder and pigment into itself, so the first coat looks patchy, weaker and often flatter than expected. A sealed or previously painted surface holds the coating on the surface, where its full colour and sheen develop. The same tin of paint can therefore produce two visibly different results in the same room if one wall has been made good and another has not.
A porous substrate draws the liquid phase of a coating into itself by capillary action, taking binder and some pigment with it. Less binder is left at the surface to form a continuous film, so the coating that remains is thinner, less glossy and less opaque than intended. A sealed substrate holds the coating where it was applied, so the film develops fully and shows the colour and sheen the product was formulated for.
Matters most on new plaster, filler, bare timber and anywhere the surface has been patched — which is to say, on almost every real decorating job. Matters least on a wall that has been previously painted throughout in the same product and needs no making good.
Sampling a colour on a patch of fresh filler or new plaster. What that patch shows is the absorbency of the substrate, not the colour of the paint, and the finished wall will look nothing like it.
Practical advice
This is the main reason a sample patch on bare filler or new plaster misleads: it is not showing you the colour, it is showing you the absorbency. Test on a primed board or a sealed area, and expect any porous area to need the preparation the product specifies before it will match the rest. Variations in absorbency across one wall are also why patchy areas can still show through after a second coat.
Always confirm with a physical sample in the actual finish before committing to a large surface — see methodology.