You unmold at three days and the bar dents under your thumb. Or you cut at a week and the edges smear. Or it's a month later, the calendar says cured, and the thing still feels like cheese.
Those are not the same problem, and the advice you find is a list of causes with no way to tell which one you have. There is a way, it takes a kitchen scale, and it costs nothing.
First: is it soft, or is it young?
Almost all hardening in the first six weeks is water leaving. So the question "will this bar ever firm up?" has a direct measurement behind it: put a bar on a scale and weigh it again in a week.
- Still losing weight? Your problem is time, and time is already fixing it. Nothing else needs doing.
- Weight has plateaued, and it's still soft? The water has gone and this is the bar. No amount of further curing will change it — the cause is one of the three below.
That single fork removes most of the guesswork, and it is the reason to weigh bars at all. The cure guide uses the same measurement to answer the opposite question — when a normal bar is finished — and gives the weight to expect.
Is your recipe even capable of a hard bar?
The standard advice is to use at least 40% hard oils. It is repeated everywhere, and it is a poor proxy — because "hard oil" is a category describing what a fat does at room temperature in a jar, not what it does to a finished bar. Our engine computes hardness directly from the fatty acids, against the customary range of 29–54. The two disagree more often than they agree:
| Blend | “Hard oils” | Hardness | INS | Verdict |
|---|---|---|---|---|
| Castile100% olive | 0% | 17 | 105 | Genuinely soft, and meant to be. Months, not weeks — this is the recipe, not a fault. |
| Bastileolive 80 / coconut 15 / castor 5 | 15% | 25 | 127 | Soft by design. Firms up usefully by 8–12 weeks; will never be a gym-bag bar. |
| Lard barlard 50 / olive 30 / coconut 15 / castor 5 | 15% | 38 | 144 | A hard, long-lasting bar — and the rule of thumb fails it badly. |
| Palm-freeolive 40 / coconut 25 / shea 15 / almond 10 / cocoa 5 / castor 5 | 25% | 37 | 146 | Perfectly firm. The butters do the work the rule refuses to count. |
| Classic barolive 35 / coconut 30 / palm 30 / castor 5 | 60% | 45 | 162 | Firm at four weeks. The rule and the measurement finally agree. |
"Hard oils" counts only the oils this site's table classifies as hard — coconut, palm and the like — which is how the rule of thumb is normally applied. Hardness and INS are computed from the fatty-acid profile of the whole blend; the customary range is 29–54, and green means inside it. Per-oil figures are in the oil properties chart. The 2 highlighted rows are where the rule and the measurement disagree — in both cases the rule rejects a recipe that makes a perfectly firm bar.
The lard bar has 15% "hard oils" — a quarter of what the rule demands — and a hardness of 38, sitting comfortably inside the 29–54 range, with an INS of 144. It makes a dense, long-lasting bar, and it has been doing so for centuries. The palm-free starter fails the rule the same way at 25% and lands at 37. Both are "wrong" only because lard is filed under animal fats and shea and cocoa under butters, so the rule doesn't count them. Look at the hardness number, not the category label — and if your recipe is below 29, no additive and no extra cure will rescue it, because the bar is doing exactly what the fats you chose make it do.
Too much water
The commonest cause of a bar that is soft early, and it fixes itself. Water is the largest thing that has to leave: the classic bar at a 25% lye solution carries 431 g of water per kilo of oils against 216 g at 40%, so the same recipe can be firm in ten days or still yielding at five weeks depending on a number you set at the mixing jug. Extra liquid on top of the lye water — milk, purée, beer added as well as rather than instead of — compounds it.
If a bar is soft, still losing weight, and you used full water, the answer is patience and airflow. If you want firmer bars sooner next time, a steeper concentration is the lever, and the water discounting guide covers what else it changes.
Too much superfat
Superfat is oil that never becomes soap, and oil is soft. The arithmetic is unforgiving:
| Superfat | Free oil per kg of oils | Share of the batter |
|---|---|---|
| 5% (default) | 50 g | 3.5% |
| 8% | 80 g | 5.6% |
| 12% | 120 g | 8.5% |
| 15% | 150 g | 10.8% |
| 20% | 200 g | 14.6% |
Classic bar at 1,000 g of oils, 33% lye solution. Computed by this site's engine — the superfat adjuster recomputes the lye for a different superfat without rebuilding the recipe.
At 5% there is 50 g of free oil in a kilo of oils; at 20% there is 200 g, which is 14.6% of the batter — a bar that is soft and slightly greasy by construction rather than by accident. High superfats have their uses, mostly in very high-coconut recipes, but they are not free, and a long cure at a high superfat trades softness for rancidity risk instead. The superfat guide makes the case for 5% as a floor rather than a preference.
The other suspects, honestly ranked
- Sugars — honey, milk, beer, fruit. Genuinely soften a bar, and the effect is real at the amounts people use. They also heat the batch, so they push it toward gel, which is a separate decision.
- Heavy fragrance loads. Fragrance oil is oil. At 6% of oils you have added 60 g per kilo that will never saponify — comparable to raising your superfat by six points. The fragrance calculator keeps that visible.
- Ungelled soap unmolds later than gelled soap, and can be noticeably softer in the first few days. It catches up.
- Silicone molds hold moisture against every face except the top, so a loaf in silicone genuinely does firm up more slowly than one in a lined wooden box. Real, and minor.
- Sodium lactate. The universal recommendation, and the wrong first answer. It makes bars release from the mold cleanly a day earlier. It does not turn a soft formula into a hard bar, and reaching for it when the hardness number is below range is treating a symptom you diagnosed wrongly.
- Salt and stearic acid will genuinely harden a bar, and both change it in other ways — salt cuts lather, stearic accelerates trace hard enough to need hot process above a few percent. Reformulating is usually cleaner.
What to do about the batch in front of you
- Weigh it and wait. If the weight is still falling, that is the whole answer. Give it two more weeks on a rack with air on every face.
- Check the recipe's hardness. Put the blend into the recipe builder. Below 29 and you have your answer, and it isn't fixable in this batch.
- Check the superfat you actually made, not the one you wrote. A mis-weighed lye can leave you far above the number in the recipe — the superfat adjuster reconstructs it from the weight you used.
- Rebatch, if it's worth it. Grating and cooking a soft batch drives off water fast and produces a firmer, rustic bar. It cannot make a soft recipe hard, only a wet one dry.
- Use it anyway. A soft bar is still soap. It will dissolve faster in the dish — keep it out of standing water and it will wash exactly as well as anything else.
Check the obvious thing first — a bar that is soft and zaps is not a hardness problem, it's a lye problem in the other direction, and worth ruling out in ten seconds. Then check whether "too soft" means "softer than the bar I bought", because commercial bars are made from different fats by a different process and are the wrong benchmark. A handmade bar at hardness 45 is a good bar; it is not a supermarket bar and it never will be.
Soaps that are supposed to be soft
Some of the best recipes fail every rule on this page on purpose. Castile at hardness 17 is far below the customary range and is not a mistake — it is soft out of the mold, needs cutting early and curing for months, and becomes one of the mildest, hardest-wearing bars there is after six. High-olive recipes generally behave the same way, proportionally. Salt bars want unmolding within hours and then months on the rack.
So before you add anything to fix a soft bar, check that you did not choose it. The fix for a castile that is soft at four weeks is week twelve.
The short version
- Weigh it. Still losing weight means time. Plateaued and soft means the recipe.
- Ignore "40% hard oils". It fails the lard bar at 15% and hardness 38.
- Below hardness 29, nothing fixes this batch. Reformulate the next one.
- Water is the usual answer, and it solves itself.
- Sodium lactate is for unmolding, not for softness. Don't let it stand in for a diagnosis.