Plant milks
Soy, oat, almond and the rest — and a measured finding that in at least one product they behaved far more like dairy milk than the usual advice suggests.
Also called soy milk (US); soya milk (UK); oat milk (INTERNATIONAL); almond milk (INTERNATIONAL); coconut milk drink (INTERNATIONAL); non dairy milk (US); dairy free milk (INTERNATIONAL); plant based milk (INTERNATIONAL); milk alternative (INTERNATIONAL); milk vs almond milk (INTERNATIONAL)
Plant milks are water extracts of a seed, nut, grain or legume, usually with added stabilisers and often with added sugar, salt and oil. They vary enormously between types and between brands, and BakeHQ models them as one category with stated differences rather than as a dozen entities, because the baking-relevant distinctions are fewer than the shelf suggests.
The headline evidence is more reassuring than most advice on this subject. A controlled comparison of eight plant milks against cow’s milk in muffins found no significant difference in baking loss, volume index or symmetry — milk substitution did not affect baking performance in that product. Batter consistency DID differ: coconut, oat and flaxseed milks gave thicker batters, almond and hazelnut thinner ones. All the plant formulations were lower in protein than the dairy control.
That finding is scoped to muffins, and BakeHQ does not extend it. A muffin is a forgiving, chemically leavened batter with a short bake. Where milk protein is doing structural work — an enriched bread, a custard, anything relying on protein coagulation — lower protein plausibly matters much more, and no comparable measurement was found.
The practical differences worth knowing are unsweetened versus sweetened, because sweetened varieties change the sugar in a formula; protein content, where soy is closest to dairy and oat and almond are well below it; and flavour, which coconut and almond carry noticeably and oat and soy carry mildly.
Composition and specification
Effect on baking performance in muffins
Across eight plant milks in muffins, baking loss, volume index and symmetry showed no significant structural differences from a cow’s milk control (p > 0.05).
ONE PRODUCT, ONE STUDY, AND STATED AS SUCH. It is good evidence that a muffin tolerates the swap and it is not evidence about bread, pastry or custard. BakeHQ does not generalise it, and the common claim that any plant milk works exactly like dairy milk remains unsupported outside this scope.
Batter consistency
Coconut, oat and flaxseed milks produced batters with higher shear stress and viscosity; almond and hazelnut produced lower.
Measured differences in the batter that did NOT translate into measured differences in the baked result. Worth stating precisely because it is the sort of finding usually reported as though the two were the same thing.
Protein
All plant-based muffin formulations were lower in protein than the dairy control — 5.40 to 5.92 g/100 g against 8.51.
These are figures for the finished muffin rather than for the milks themselves, and the paper draws no causal link between them and how the muffins baked.
What changes if it is left out
Dairy milk substitutes for plant milk straightforwardly in any formula where the plant milk was supplying liquid. Going the other way is well supported for a muffin and unmeasured elsewhere.
Follows from the measured comparison recorded above and from its stated scope.
RecordedFollows from evidence recorded elsewhere on BakeHQ
Why BakeHQ does not have a page per plant milk
Soy, oat, almond, cashew, hemp and the rest differ from each other in protein, fat, sugar and flavour, and a separate page for each would mostly repeat this one.
The distinctions that change a bake are few enough to state here: whether it is sweetened, roughly how much protein it carries, and whether its flavour survives baking. Everything else is brand variation, and brand variation within one type is often larger than the difference between types — two oat milks can differ more than an oat and a soy.
Where a specific plant milk turns out to have a distinct, evidenced baking behaviour worth its own record, it will get one. None has yet.
Related
What it does
Hydration
A water extract of a seed, grain or nut, and it is the liquid of a vegan bake. Measured against dairy milk in muffins it produced no significant difference in the baked result — but sufficiency still depends on quantity, which BakeHQ does not model.
Browning contribution
Sugars and protein are present and brown, in smaller quantities than dairy milk carries. Sweetened varieties bring more sugar than the recipe intended, which is the commonest surprise.
Can stand in for
Milk— for the liquid — measured as equivalent in a muffin, and not measured anywhere else
THE SUBSTITUTION IS BETTER EVIDENCED THAN USUAL AND NARROWER THAN CLAIMED. Eight plant milks in muffins gave no significant difference in baking loss, volume or symmetry. That is one product with a forgiving chemically leavened batter. Where milk protein does structural work — enriched bread, custard, anything relying on protein setting — no comparable measurement was found and BakeHQ does not extend the finding.
Compared in
One side of the comparison, so it is reachable from the record a reader is more likely to land on first.