Why kitchen recipes don't scale linearly (and the salt floor that breaks everything)
Doubling a recipe doesn't double the cooking time, and doubling the salt makes it too salty. The five things that don't scale the way you expect.
# The trap
You found a recipe for 6 and you need 4 servings. Or your dinner party went from 8 to 12. The instinct is to multiply every ingredient by 4/6 (= 0.667) or 12/8 (= 1.5) and call it done. For most ingredients that works fine. For three things it doesn't — and getting any of them wrong is what makes a scaled recipe taste off.
# The 30-second version
1. Bulk ingredients scale linearly. Flour, sugar, butter, eggs, milk, oil — multiply by the scale factor and stop thinking about it.
2. Salt and strong spices don't. Scale them at 80-85 % of the factor and adjust at the end.
3. Cooking time doesn't either. A doubled cake bakes 30-40 % longer, not 100 %. Use a thermometer, not a stopwatch.
4. Pan size needs the square root. Doubling the batter wants ~1.4× the linear dimension — or two pans.
5. There's a floor at ~½ recipe. Below that, switch to a recipe written for the smaller size.
# What scales linearly (the easy 80 %)
Almost everything you'd weigh. Flour, sugar, butter, eggs, milk, cream, oil, water, fruit, chocolate, vegetables, meat. The chemistry of the dough or batter cares about ratios — flour to liquid, fat to flour, egg to flour — so as long as you preserve those ratios, the recipe still works at any size.
Use the Recipe Scaler to do the arithmetic. It handles fractional eggs (round to whole, or use yolks) and unit conversions (cups → grams) along the way.
# Salt and spice are sub-linear
Salt perception in food is non-linear. The reason is biology — your taste receptors respond logarithmically to concentration, not linearly. When you double both the salt and the food, the perceived saltiness goes up by ~10-15 %. Same for chilli heat, black pepper, and most savoury aromatics.
<div class="callout callout-tip" role="note"><div class="callout-title">Tip</div><div class="callout-body"><p>Scale salt at <strong>80-85 %</strong> of the scale factor. So if you're doubling the recipe, multiply salt by ~1.7×, not 2×. Taste at the end and adjust.</p></div></div>
This applies most strongly to:
- Salt (the classic case)
- Black pepper, white pepper, chilli, paprika
- Strong herbs — rosemary, oregano, bay, sage
- Acid — vinegar, lemon juice, when used as a finishing seasoning
It applies less to subtle baking spices (cinnamon, nutmeg, vanilla, cardamom in small amounts), which can scale linearly without anyone noticing.
# Cooking time scales by geometry, not weight
The biggest trap. Doubling a cake recipe doesn't double the baking time.
Heat enters food from the outside. A 9-inch cake at 30 minutes has heat travelling roughly an inch and a half to the centre. Scale the cake to 13 inches with the same recipe ratio (same batter depth), and heat still only has to travel an inch and a half to the centre — the cake is wider, but it's not deeper. Baking time goes up maybe 5-10 minutes.
If instead you keep the same pan and pour twice the batter into it, the batter is now twice as deep. Heat has to travel twice as far. Bake time goes up substantially — but still not 2×, because the outside layers are already cooking while the inside catches up.
For roasts and large pieces of meat, cooking time scales roughly with weight to the 2/3 power. A 5 kg chicken doesn't take 2.5× the time of a 2 kg chicken (it takes ~1.8×). The 2/3 exponent is the geometric reality of heat-by-conduction through a 3D shape.
<div class="callout callout-warning" role="note"><div class="callout-title">Warning</div><div class="callout-body"><p>Always use a thermometer or visual cues for scaled bakes, never the timer. Internal temperature is what's actually telling you it's done.</p></div></div>
# Pan size matters more than people think
Cake batter in a too-large pan spreads thin and dries out. In a too-small pan it overflows or stays raw in the middle.
The right pan size for a scaled recipe is proportional by surface area, not volume. If you've doubled the batter, you want twice the surface area — which is the square root of 2 (≈ 1.4×) the linear dimension. A 9-inch round at 1× scale should become a 12.7-inch round at 2× scale, or — more practically — two 9-inch rounds.
For loaf pans, the same √2 rule holds: doubling means roughly 1.4× longer (if you keep the same width and depth) or two pans.
Using two of the original pan is almost always easier than hunting down the perfect-size single pan, and it preserves the bake-time geometry — your original recipe's time is right for each pan.
# A worked scale
Source recipe: 6-serving chocolate cake.
- Flour 300 g
- Sugar 200 g
- Butter 200 g
- Eggs 3
- Salt 1 tsp
- Bakes in a 9-inch round at 180 °C for 35 minutes
Scaling down to 4 servings (factor 0.667):
- Flour 300 × 0.667 = 200 g (linear)
- Sugar 200 × 0.667 = 133 g (linear)
- Butter 200 × 0.667 = 133 g (linear)
- Eggs 3 × 0.667 = 2.0 → 2 eggs (round naturally)
- Salt 1 × 0.55 (sub-linear) = ~0.55 tsp, adjust at end
- Pan: still 9-inch (it'll be a flatter cake, that's fine) OR 7-inch for the same depth
- Bake time: start checking at 25-28 minutes with a thermometer (90 °C internal for sponge cake)
The salt math is 0.667 × 0.85 = 0.55. If you'd scaled the salt linearly to 0.67 tsp, the cake would taste slightly salty — not enough to call out, but enough to feel "off".
# There's a floor at half-size
Below half a recipe, the math is still right but the practicality breaks. A single beaten egg weighs about 50 g. A recipe scaled to 1/4 size that originally called for one egg now asks for "0.25 of an egg" — which you can approximate (1 yolk, or 12 g of whisked egg) but it's fiddly.
Salt and spice quantities fall below the resolution of standard measuring spoons. Baking dimensions shrink to non-standard pans (a 4.5-inch cake tin doesn't exist in most home kitchens). Cooking times become weirdly short and prone to surface overcooking before the centre is done.
For small portions, the better answer is to find a recipe written for that size — most published cookbooks include single-serving or two-person variants of classics — rather than scaling down a six-portion recipe.
# What this means in practice
- For 0.75× to 1.5× scaling, multiply linearly and don't worry.
- Above 1.5× or below 0.75×, ease back on salt and spices by 15-20 %.
- For any baked good, check doneness with internal temperature and visual cues, not the timer.
- Switch pans (or use two) instead of cramming bigger batter into the original.
- Below ½ scale, find a different recipe.
# Related tools
Common questions
Frequently asked.
›Why doesn't doubling a cake recipe double the bake time?
Heat enters from the outside of the pan. A bigger volume of batter has more interior mass relative to its surface area, so the centre takes longer to reach temperature — but only ~30 % longer, not 100 %. Geometry, not the recipe, controls baking time.
›Should I double the salt when I double a recipe?
Almost never. Salt perception is non-linear — twice the salt in twice the food tastes slightly saltier, not the same. Most experienced cooks scale salt at 80-85 % of the recipe's scale factor and adjust at the end.
›What's the smallest fraction of a recipe I can reliably make?
Roughly half. Below that, ingredient measurements run into floor problems — a single beaten egg weighs ~50 g, salt and spice quantities drop below a teaspoon's resolution, and the pan size for baking gets weird. Find a recipe written at your target size rather than scaling.