I Scaled My Candle Formula From 1 Kg to 10 Kg Using Exactly the Same Percentages — Why Does the Finished Candle Perform Differently?

First prove the difference is real

Here is the situation. Your 1 kg test in Premium Coconut Soy Wax CSI 468 used 920 g of wax and 80 g of Eucalyptus Spearmint, which is 8% of total weight. The 10 kg version used 9.2 kg of wax and 800 g of oil, filling fifty 200 g jars (48 after the 5% planning wastage). Every percentage matches. Yet the production jars seem to perform differently — softer throw, a slightly different melt pool, or a top that looks less even.

Before you diagnose anything, establish that the difference exists. Makers who have just spent money on a larger batch are primed to notice faults, and the original test candle has usually been admired for weeks. Memory of a scent is poor evidence. A difference you cannot detect blind is not a production problem; it is a worry.

The triangle test

The odd-one-out comparison is the cleanest way to answer “is there a difference at all?” without asking anyone which candle is better.

The log above is an illustration of the format, not a result. If your picks are right about as often as chance, you do not have a scaling problem — you have a comparison problem, and the fix is to judge candles at matched ages from now on. If the picks are right nearly every time, the difference is real, and you go looking for its cause.

The three hidden variables that most often move

When the materials and percentages match and the difference is still real, the cause is usually one of three things the recipe never recorded, because at 1 kg they did not need recording.

1. Cure age and storage

Fifty jars need space. They often end up in a stack of cartons in a warmer corner, while the test candle sat alone on an open shelf. Where and how the jars cured is part of the product. Check both the age at which each candle was judged and where it waited, and write both on the sheet.

2. Thermal history of the scented wax

Your 1 kg of CSI 468 probably went from melt to jars in a few minutes. The 10 kg charge sat in a melter for longer before oil went in, and longer again after, while fifty jars were filled. The product page gives a melt point of 50–53°C and a pour of 58–60°C. If you only know the melter's display setting, you do not know the wax's history. The questions are: what did a separate thermometer read when the oil went in, and how long was the scented wax held hot before the last jar was filled?

3. How fast the jars cooled

Five jars spaced out on a counter shed heat quickly. Fifty jars packed shoulder to shoulder on a tray warm each other and cool more slowly, particularly in the middle of the group. Cooling rate affects the surface and the way soy crystallises, and some makers find it affects how the melt pool forms on the first burns. Compare a jar from the centre of the tray with one from the edge before blaming the oil.

Notice what is not on this list: the percentage itself. Because every percentage matched, a load or basis error is already excluded, and so is any argument that the oil “does not scale”. Fragrance oil has no idea how big the batch is. What it experiences is temperature, time and the wax around it, and those are exactly the things the three hidden variables describe. That is why a maker who responds to this problem by adjusting the formula usually ends up with a new formula that has the same process problem.

Isolating the variable without remaking 10 kg

Each suspect can be tested at 1 kg, which keeps the cost of learning small. Make pairs of controls that differ in one thing only.

Keep a spare of the oil for this. Buying Eucalyptus Spearmint as 1 kg at ₹3,068.00 rather than 500 g at ₹1,652.00 plus three 100 g bottles at ₹354.00 each (₹2,714.00) costs ₹354.00 more but leaves 200 g after the 800 g batch — enough for two 80 g controls from the same lot.

What to do with the fifty jars meanwhile

If the triangle test showed no reliable difference, release the batch once it has passed a full burn at cure. If a real difference was found and the production jars still burn cleanly, perform safely and meet the throw you describe to customers, they may be sellable while you fix the process for the next run — a difference is not automatically a defect. If the difference is a fault, such as a poor melt pool or visible soot, hold the jars until the cause is known; a wick or cooling issue will not be solved by remelting with more oil.

The material in this batch is worth roughly ₹8,574.40 before jars and wicks: 9.2 kg of CSI 468 at ₹637.00 per kg (₹6,370.00 ÷ 10 kg) = ₹5,860.40, plus ₹2,714.00 of oil bought as above. That is the sum at risk if you guess.

The standing control for every future batch

Add a production line to the recipe card. Record thermometer reading at oil addition, time from oil addition to last pour, how the jars were arranged while cooling, where they cured, and the age at which they were judged. None of these appears in a percentage formula, which is why the formula alone could not predict the 10 kg result.

Then keep making a 1 kg reference beside each production run until three or four batches in a row pass a blind comparison. That record, not the percentages, is what lets you call the process validated. If you want help setting up the sheet, CSI's WhatsApp line is +91 7397976926.

Where this page fits

CSI already has guides that cover parts of this question. This page covers the failure investigation for the question above; these go deeper on the rest:

Three things to buy for this job

Frequently asked questions