How to scale a recipe without ruining it

Doubling a recipe looks like arithmetic. For a stew it very nearly is. For a cake it is the reason the cake came out dense, flat and slightly bitter.

The scaling factor is the easy part. What follows is everything the multiplication does not cover.

The factor itself

Divide the servings you want by the servings the recipe makes:

factor = desired servings ÷ original servings

6 people from a 4-person recipe  →  6 ÷ 4 = 1.5

Multiply each quantity by that. The recipe scaler does this across a whole ingredient list at once, understanding fractions, mixed numbers, unicode fractions like ½ and ranges like 2–3, and leaving the ingredient names and notes untouched.

Presenting the result usefully matters more than it sounds. Scaling ¾ cup by 1.5 gives 1.125 cups, and nobody owns a 0.125-cup measure. A useful tool rounds to fractions a kitchen actually has — halves, thirds, quarters, sixths and eighths — and falls back to a decimal when nothing close exists, rather than inventing a measurement like ⅐ of a cup.

What scales linearly

Most savoury cooking. Soups, stews, braises, sauces, curries, stir-fries, roasted vegetables: doubling the ingredients doubles the dish, and small errors are correctable by tasting. This is the case where arithmetic really is enough.

Bulk ingredients in baking too — flour, sugar, butter, milk — scale cleanly. It is everything else in a baked recipe that does not.

What needs judgement

Eggs come in whole units. A recipe scaled by 1.5 asking for 4.5 eggs forces a decision. Weighing is the accurate answer: a large egg is roughly 50 g out of the shell, about 30 g white and 20 g yolk. Beat one and use half by weight. The easier answer is to adjust the factor slightly so it lands on whole eggs and accept a marginally different yield — usually the better trade for home baking.

Salt and strong spices rarely scale linearly to taste. Chilli, cayenne, saffron, cloves and nutmeg all intensify more than proportionally in a larger batch, partly because a bigger dish cooks longer and extracts more. Scale these a little under — around 75–80% of the arithmetic answer — and correct at the end, which is the one direction you can always correct in.

Yeast does not double with the dough. A larger mass ferments faster because it retains its own heat, so doubling yeast in a doubled dough tends to over-prove. Keeping the yeast closer to the original amount and judging by the dough rather than the clock is the reliable approach.

Chemical leavening — baking powder and bicarbonate — is the least forgiving. Excess produces a soapy or metallic taste and a crumb that rises fast and collapses. Many bakers scale leavening at slightly under the factor for this reason.

Gelatine and setting agents depend on the ratio to liquid, which does scale — but also on surface area and depth for cooling, which does not.

The scaler flags these ingredients rather than silently multiplying and hoping. The arithmetic is still shown, because you need it as a starting point; the flag is there so you decide rather than assume.

Pan size: scale by area, not by length

This is the single most common scaling mistake in baking.

Doubling a recipe does not mean doubling the tin’s dimensions. Batter depth is what determines baking time, and depth depends on area:

Tin	Area	Relative to 20 cm
20 cm round	~314 cm²	1.0×
23 cm round	~415 cm²	1.3×
25 cm round	~491 cm²	1.6×
20 × 20 cm square	400 cm²	1.3×
23 × 33 cm rectangle	759 cm²	2.4×

So a doubled cake batter goes into a tin with roughly twice the area — a 23 × 33 cm rectangle, not a tin twice as wide. A 25 cm round holds about 1.6×, not 2×.

For a round tin the area is πr², so going from 20 cm to 28 cm diameter roughly doubles it. If you keep the same tin and simply pour in more batter, the result is deeper, which changes the bake fundamentally.

Time and temperature

Temperature usually stays the same. Very large roasts are the exception, where a slightly lower temperature helps the centre reach doneness before the outside overcooks.

Time does not scale with quantity. Two principles cover most cases:

• A larger volume takes longer to come to temperature but nowhere near proportionally. A double batch of stew might need 20–30% longer, not 100%.
• A larger surface area at the same depth bakes in roughly the same time. A tray of cookies twice as wide bakes in the same time; the same batter twice as deep takes considerably longer.

The practical rule: check for doneness early and often rather than trusting the original time. Internal temperature, a skewer, or the visual cues in the original recipe are all more reliable than the clock once you have scaled.

When not to scale at all

Some things resist scaling and are better made in batches:

• Anything relying on evaporation — reductions, caramel, jam. A wider pan reduces faster, a deeper one slower, and the ratio you are aiming for is a concentration rather than a quantity.
• Pan-frying and searing — crowding a pan drops the temperature and steams instead of browning. Cook in batches at the original quantity.
• Emulsions — mayonnaise, hollandaise. They can be scaled, but the margin for error narrows and a split batch wastes far more.
• Very small scaling down. Quartering a recipe often lands on quantities too small to measure or to behave properly in a pan.

A workable procedure

1. Work out the factor and scale the bulk ingredients.
2. Hold back salt and strong spices to around 75–80%, to correct later.
3. Decide the eggs deliberately — weigh, or adjust the factor.
4. Keep leavening at or slightly below the factor.
5. Choose a tin by area, not by width.
6. Keep the temperature; check doneness early.
7. Taste and adjust at the end, which is the step the arithmetic cannot do.