Doubling a recipe should be the easiest thing in the kitchen: take every quantity, multiply by two. Most of the time it works. But anyone who has tried to triple a brownie recipe or halve a sourdough has discovered that scaling is not quite that simple. Some ingredients scale linearly. Some scale partly. Some don\'t scale at all.
This is the reference for what to multiply, what to taste-test, and what to leave alone. We work through the main exceptions and explain why each one breaks the math.
What scales perfectly (just multiply)
The bulk of any recipe — flour, sugar, butter, milk, oil, water, fruit purée, chocolate, cream, broth, cheese, vegetables, meat — scales linearly. Want to make twice as many cookies? Use twice as much flour, twice as much butter, twice as much sugar.
These ingredients scale because their function in the recipe is bulk: they make up the substance of the dish, and doubling the substance gives you double the food. The chemistry between them does not change with quantity. A given ratio of flour to butter behaves the same whether you have a teaspoon of each or a kilogram of each.
For most cooking — soups, stews, roasts, sautés, casseroles, pasta dishes — almost everything is in this category, and straight multiplication works. The math is the same as the Recipe Scaler Calculator does automatically: scale factor = target servings ÷ original servings, multiply each ingredient by the factor.
What scales but needs adjustment: salt and spices
Salt is the most-discussed scaling exception. The chemistry is linear — twice the salt in twice the food gives the same concentration — but the perceived saltiness is not. Doubling the salt does not taste "twice as salty" to the eater; it tastes meaningfully saltier, because human taste perception is roughly logarithmic at high concentrations.
The practical rule: scale the salt with everything else as a starting point, then taste at the end and adjust. For larger batches in particular, you almost always end up using less salt than the linear-multiplied amount calls for. The same applies to chili, garlic, pepper, ginger, fish sauce, and any strong flavour.
For mild spices (oregano, basil, thyme, paprika) the linear math is closer to correct. For very strong ones (cumin, cayenne, smoked paprika) reduce to about 75 percent of the linearly-scaled amount and adjust from there.
What scales in whole numbers only: eggs
Eggs come in whole units, and you cannot use 1.5 of one. When a scale factor gives you a fractional egg, you have three options:
Round to the nearest whole egg. The simplest fix and works for most recipes — a cake that calls for 2.5 eggs at scale rounds to 3, and the cake is fine. For delicate cakes and pastries, this can be too much; for everyday baking, it is invisible.
Beat one extra and measure. Crack one egg into a bowl, beat lightly, and measure out the partial amount you need. A large egg is about 50 mL of beaten egg, so half an egg is 25 mL. This is the precise option and is what professional bakeries do.
Switch unit. Recipes for very large batches (a hundred portions, say) are usually written in mass rather than count: "500 g eggs" instead of "10 eggs." This avoids the fractional-egg problem at any scale.
What needs careful adjustment: leavening
Baking soda, baking powder, and yeast follow the linear math for most everyday scaling (1×, 2×, half). But for large multipliers (3× and above), they break in a specific way: too much leavening makes the rise overshoot and collapse.
A 2× or 3× cake usually works with linearly-multiplied leavening, but the rise tends to be slightly less reliable. At 4× and above, professional bakers reduce the leavening by about 15-20 percent of the linearly-scaled amount to account for the increased bulk-to-surface-area ratio of the larger batch — bigger batches retain more steam during baking and need less chemical lift.
For very small batches (half or quarter), increase the leavening slightly — small batches lose more leavening to surface evaporation during mixing.
For bread yeast specifically, the relationship is different: yeast is alive, and it reproduces during fermentation. Doubling a sourdough recipe does not require doubling the starter; you can use 1.5× the starter and let it ferment slightly longer. This is part of why professional bakeries can make 50-loaf batches from a small inoculation.
What does not scale: baking time
Baking time is set by heat transfer, which depends on the depth of the food, not its volume. A 1× cake in an 8-inch round pan and a 2× cake in two 8-inch pans bake in almost exactly the same time, because each pan has the same batter depth. The doubled recipe does not need 2× the time; the recipe-time figure is still correct.
If you 2× a recipe and bake it in one big pan instead of two — making the batter twice as deep — you do need significantly more time. The math is roughly linear with depth: 1.5× depth ≈ 1.5× time. Start checking 10 minutes early no matter what.
For roasts and other large food, the depth-not-volume rule also applies. A 5 lb roast does not take twice as long as a 2.5 lb roast; it takes about 1.3× as long, because the heat has to penetrate the centre and the centre is only slightly further from the surface in the larger piece.
What does not scale at all: oven temperature
The oven temperature is the same whether you make one batch or twenty. A 350 °F cake bakes at 350 °F regardless of pan count. Resist the temptation to crank the oven for a bigger batch; it doesn\'t help and usually browns the surface before the centre cooks.
The threshold where a recipe stops scaling
Around 4× to 5×, most home recipes stop behaving the same. A doubled cookie recipe is reliably twice as many cookies; a quadrupled cookie recipe rises differently because the chemistry is no longer in the comfortable range the original recipe was designed for. The leavening drift, the salt-perception issue, the slightly different mixing dynamics — they all compound at large multipliers.
Professional bakeries don\'t "scale up" home recipes. They use formulas designed for commercial-scale production, with the leavening ratios and mixing methods tuned for big batches. For one-off occasions (a wedding cake, a large dinner party), three options:
Find a recipe written for the size you want. Search "wedding cake recipe" rather than "cake recipe ×8." Recipes written for large batches are reliable at that size.
Make multiple smaller batches. Two 2× batches of brownies are more reliable than one 4× batch. The total time is about the same.
Switch to baker\'s percentages. Professional formulas express everything as a percentage of the flour weight. A bread formula at 100% flour, 65% water, 2% salt, 1% yeast scales perfectly to any size — just pick a flour weight and multiply through. Most home recipes can be converted to this format with a kitchen scale and the Ingredient Density Converter.
Pan substitution
When you scale a recipe, you usually need a different pan. The Pan Substitution Calculator handles this: it computes the area ratio between two pans (rectangular, square, or round) and tells you the scale factor.
The principle: pan capacity scales with floor area, not diameter. A 9-inch round (63.6 in²) holds 27 percent more batter than an 8-inch round (50.3 in²), not 12.5 percent. The same is true for square and rectangular pans. Always scale by area when substituting.
Calculators on this site
- Recipe Scaler — paste an ingredient list, get the scaled version.
- Pan Substitution Calculator — area-based scaling when you change pans.
- Ingredient Density Converter — volume to weight conversions for baker\'s percentages.
- Cocktail Batch Scaler — for batching cocktails, which scales differently because of dilution.
Sources
- King Arthur Baking Company Ingredient Weight Chart and scaling articles.
- On Food and Cooking by Harold McGee (taste perception, baking chemistry).
- The Bread Baker\'s Apprentice by Peter Reinhart (baker\'s percentages, large-batch yeast adjustments).
- USDA FoodData Central (ingredient densities).