Your first cold process batch, step by step

Cold process soap is four ingredients and about ninety minutes of work: fats, sodium hydroxide, water, and time on a shelf. The chemistry looks after itself — if the weights are right, the batch in front of you will become soap. Almost everything that goes wrong on a first batch goes wrong before the blender switches on, in the arithmetic.

So this page is built the other way round from most tutorials. Instead of a recipe at one fixed weight that may or may not fit your mold, you get the recipe as percentages, the numbers for five common batch sizes, and the reason to pick one size over another. Then the method.

What you need

Nothing here is specialized except the lye, and the whole list can come out of a supermarket and a hardware store. The one item worth spending on is the scale, for reasons the next section makes arithmetic.

• A digital scale that reads to 0.1 g, and ideally a second one for large weights. A 1 g kitchen scale will make soap; it will just make soap at a superfat you didn't choose.
• A stick blender (immersion blender). The cheapest one works — it is a motor and a blade. Hand-stirring a cold process batch to trace can take an hour and often stalls.
• A heat-safe container for the lye solution — polypropylene (#5), HDPE (#2) or stainless steel. Never glass, never aluminum.
• A stainless or heavy plastic bowl or pot for the oils, big enough that the batter fills it no more than two-thirds.
• Silicone spatulas, and a thermometer — an infrared one is convenient, a $5 probe is fine.
• A mold. A silicone loaf mold, a lined wooden box, or a clean cardboard box lined with freezer paper. Measure its inside dimensions before you do anything else.
• Goggles, gloves, long sleeves, closed shoes, and a room you can ventilate.

Ingredients: olive oil, coconut oil, palm oil and castor oil from any grocery or soap supplier; 100% sodium hydroxide (the label must say sodium hydroxide and nothing else — drain cleaners with added aluminum shavings are not usable); and distilled water. Skip fragrance and color on batch one. You want to see what plain soap does before you add variables.

The recipe: the classic beginner bar

Almost every beginner recipe published in the last twenty years is a variation on the same four oils, and there's a good reason: coconut for lather and hardness, palm for a firm bar that isn't drying, olive for mildness, castor to make the lather last. Ours is the same:

• 35% olive oil — mild, slow-moving, the body of the bar
• 30% coconut oil (76°) — big bubbles and hardness; drying above about 30%, which is why it stops there
• 30% palm oil — hardness and a creamy, stable lather
• 5% castor oil — lather that keeps going; more than about 10% turns the bar sticky

At 5% superfat and a 33% lye solution that blend scores hardness 45, cleansing 20, conditioning 51 — squarely in the ordinary-bar range on every axis, which is exactly what you want from a batch whose job is to teach you the process. You can see how it compares with anything else in the oil properties chart.

No palm? It is the one ingredient here with a supply-chain argument attached, and it is also the easiest to replace: the recipe builder's palm-free starter swaps it for shea, sweet almond and cocoa butter and lands at 139.1 g of lye per kilo of oils instead of 143.8 g. Lard and tallow work at least as well and cost less. Whatever you swap, put the new blend through the calculator — every oil takes a different amount of lye, which is the whole subject of the saponification guide.

How big should a first batch be?

Here is the number every tutorial skips. The standard advice is to start with a 1 lb test batch so a mistake costs little, and as far as money goes that's right. But the smaller the batch, the more a gram of lye matters — and beginners are the people most likely to be weighing on a 1 g kitchen scale.

Olive 35 / coconut 30 / palm 30 / castor 5 at 5% superfat and a 33% lye solution. Computed by this site's engine — the lye calculator will do it for any weight or any blend. Water is distilled water for the lye solution; batter is what goes into the mold.

Now the same batches, viewed as tolerance. One gram of lye either way is the resolution of a cheap scale — not a mistake, just the smallest difference it can see:

The superfat a 1 g weighing difference actually produces, computed from the same recipe. The superfat adjuster does this for your own batch, in either direction.

Get your own numbers, not a blog's

The recipes you find elsewhere are frozen calculator output, and the settings that produced them are usually not on the page. Take a much-copied beginner recipe: 34 oz of oils (coconut 8, olive 15, palm 11), published as 4.8 oz of lye and 11.2 oz of water at 5% superfat. Put that blend through our engine and it returns 4.81 oz and 11.21 oz. It is the same math, and it is correct — but the water figure is a 30% lye concentration (2.33:1 water to lye) that the recipe never names, and it is fixed at 34 oz of oils whether or not that is what your mold holds.

Two minutes with the tools gives you the version that fits your bench:

1. Measure the inside of your mold and put the dimensions into mold volume. It returns the oil weight that fills it — mold volume is not oil weight, and the gap between them is where overflowed molds come from.
2. Take that oil weight to the lye calculator with the four percentages above. You get the lye and water for your batch, with the SAP value used for each oil shown.
3. Write the numbers down, or print the sheet at the bottom of this page. Then check them against a second calculator — a habit worth keeping forever, and the reason our numbers are published with their sources.

Leave superfat at 5% and lye concentration at 33% for a first batch. Both are defaults for good reasons, worked through in the superfat guide and the water discounting guide.

Making the lye solution

Weigh the water into the heat-safe container first. Weigh the lye separately, on the finest scale you have. Then add the lye to the water, in stages, stirring gently — never water onto lye, which can flash-boil and erupt.

It gets hot fast: a 33% solution reaches somewhere near 90 °C (about 190 °F) within a minute, and it gives off a sharp vapor for the first thirty seconds or so. Stir until the liquid goes from cloudy to clear, then move it somewhere safe and undisturbed to cool. On a 1 kg batch you are cooling 436 g of caustic solution, which takes 30–45 minutes in a sink of cold water and rather longer on a counter.

While it cools, melt the hard oils. Coconut and palm are solid at room temperature; warm them gently in a pot or the microwave until just melted, then pour the olive and castor in cold. That usually lands the oil pot in the right range on its own, without any further heating. Stir the palm oil in its container before you weigh it — it separates into a soft and a hard fraction while it sits, and scooping from the top of an unstirred tub gives you a different oil than the one the calculator priced.

What temperature should I soap at?

The top-ranking tutorials say 100 °F, 100–110 °F, and "130 °F or below, ideally within 10 degrees of each other" — three different rules, each stated as the rule. They're all fine, because the temperature window is wider than any of them admits and it isn't the chemistry that sets it.

Saponification does not need the heat. It will run at room temperature; it produces its own heat as it goes. What temperature actually controls is how fast the batter thickens and whether the loaf goes through gel phase. The floor exists for a completely different reason:

So: aim for both somewhere between 38 °C and 49 °C (100–120 °F) and don't chase a match. Ten degrees apart is fine, twenty is usually fine. The one combination to avoid is cool oils with a cool solution on a butter-heavy recipe. Hotter batter moves faster and is more likely to gel, which matters when you start using fragrances that accelerate — but that's a problem for batch five, not batch one.

How do I know when it's at trace?

Pour the lye solution into the oils down the shaft of the stick blender, or over a spatula — anything that stops it splashing. Then, with the blender head submerged and tilted to let trapped air escape: pulse for three to five seconds, stir for twenty with the motor off, repeat. Short bursts are the whole technique. Running the blender continuously whips air into the loaf and can take a plain recipe to a thick trace faster than you can pour it.

You are watching for the batter to go from oily and translucent to opaque and uniform — that's emulsion, and it is the point past which the batch is safe to pour. A few seconds more and you get light trace: drizzle some off the blender and the drips sit on the surface for a moment before sinking, like thin custard. That's where a first batch should go into the mold. Thick trace — pudding that holds a spoon mark — is for layered and textured work, and if you get there by accident you have about a minute to scoop it in and flatten the top.

On this recipe, with short pulses, expect emulsion in under a minute. If two or three minutes of blending leave it thin and glossy with no thickening at all, stop and consider whether the lye actually dissolved — an undissolved layer at the bottom of the jug is a real and dangerous failure mode, and it is why you stir until the solution goes clear.

How much time you get is mostly a property of the recipe rather than of your technique, and it is worth knowing before you plan anything ambitious. The trace guide ranks the common blends by working time and covers what to do when a batch goes further than you wanted.

Into the mold, and the next 48 hours

• Pour in one go, scrape the bowl, and tap the mold on the counter a few times to bring air bubbles up. Flatten or texture the top.
• Insulate — or don't, on purpose. Covering the mold and wrapping it in a towel holds the heat in and pushes the loaf through gel phase: it goes translucent and dark from the center outward over a few hours, then cools. Gelled soap is slightly harder sooner and has deeper color; ungelled is paler and creamier. Neither is better. Leaving it uncovered in a cool room usually prevents gel; a partial gel — a dark circle in a pale bar — is only a cosmetic fault, and it comes from insulating just enough to start one.
• Leave it alone for 24–48 hours. Saponification is essentially finished in that window; there is no free lye left after it. Everything that follows is about water leaving.
• Unmold when it's firm and releases without distorting. A silicone loaf usually peels away at 24–48 hours; a high-olive recipe can need three or four days.
• Cut while it's still slightly soft, with any straight blade. Left too long the loaf goes crumbly at the edges and chips instead of slicing. If you're not sure how many bars to cut, the bar count calculator works out the thickness and what each will weigh once cured.
• A white powdery film on the top is soda ash — cosmetic, harmless, and rinses or steams off. It is not lye on the surface.

Then the bars go on a rack, standing on edge with air on as many faces as possible, for four to six weeks. That wait is not optional and it is not about safety — the soap is safe within days. It's about water leaving, and it is the single biggest difference between a bar that lasts a fortnight in the shower and one that lasts a month. The cure guide explains how to tell when it's actually done, with a scale, instead of counting weeks.

Cleaning up

Everything you used is coated in raw soap, which is caustic until it saponifies. Keep the gloves on. Scrape as much batter as you can into the mold or the bin rather than washing it down the sink, then leave the bowls, blender and spatulas somewhere safe overnight — by morning the residue has turned to soap and washes off with hot water and almost no effort. Washing up immediately means fighting greasy caustic slime with detergent, which is worse in every way.

Dedicate the equipment to soap if you can. It's not that trace amounts of cured soap are dangerous; it's that you don't want to be reasoning about which bowl had lye in it while cooking dinner.

What actually goes wrong on a first batch

• The mold overflowed, or came out half full. The commonest one, and it is entirely a sizing problem — mold volume first, then the recipe.
• It seized or turned to lumps. Almost always fragrance, and almost never on an unscented first batch. Another reason to skip it.
• Oil pooling on top after pouring. False trace: it was poured before it emulsified. Recoverable — scoop it back into the pot, blend it properly, re-pour.
• The bar is soft after a week. Usually just cure, sometimes a soft recipe, occasionally too much water. Give it the full six weeks before deciding.
• It stings or zaps on the tongue after cure. Stop using it. That is unreacted lye, which means the weights were wrong — the superfat adjuster's second mode will tell you what superfat your actual lye weight produced, and whether the batch is rebatchable or bin material.

Note what isn't on that list: the chemistry failing. It doesn't. If the weights are right and it emulsified, you have soap.

Print this batch sheet — no email, no signup