I'm Making 100 × 200 g Candles at 8% Fragrance — How Much Fragrance Oil Will I Need Including Enough Extra for Testing?

The 1,600 g production figure

CSI's convention is that fragrance load is a share of the finished candle weight. A 200 g candle at 8% is 16 g of oil and 184 g of wax. A hundred of them hold 1,600 g of oil and 18,400 g of wax, 20,000 g in total. That is the net production figure. It includes no pour loss, no test candles and no spare.

The pour loss covers oil that is mixed into wax but never reaches a jar: the film on the pot, the skin in the jug, the drips on the stirrer, and the top-ups for candles that sink. Makers commonly plan 3–5% extra for pour loss and test burns. This page uses 3% for production, because testing gets its own separate line. If your setup loses more, use 5%, which would be 80 g. Weighing a 50-candle batch pour by pour shows how to measure your real figure after one run.

What this figure deliberately leaves out is the question in the title: how much extra for testing. That depends on what you still need to find out, and it is the subject of the rest of the page.

Three test rounds, sized in grams

Test candles answer specific questions: does this oil behave in this wax, does this wick give a full melt pool in this jar, does 8% smell right or does it need to move? Each question needs a certain number of candles, and each candle needs oil at its own load. That is why a flat "add 10% for testing" is the wrong model. It is too much for a repeat recipe and too little for a new one. Here are three standard rounds:

Round 1, confirmation. You have made this candle before, with this oil, wax, wick and jar. Three candles from the first production melt let you burn one at your usual curing checkpoint, keep one for a full-length burn and keep one as the retained reference for the batch. At 16 g each, they cost 48 g of oil.

Round 2, load × wick matrix. Something has changed: a new jar, a new wick, or a load you have not settled. Pour two candles at each of 6%, 8% and 10%, with each of two wick options. That is twelve candles. The oil is not 12 × 16 g, because each load carries a different amount of oil: 6% is 12 g, 8% is 16 g and 10% is 20 g. Four candles at each load gives 4 × (12 + 16 + 20) = 192 g. The wax also differs by load (188 g, 184 g and 180 g), so it totals 2,208 g. Luxury Soy Wax Chunks are rated up to 13% and CSI 464 up to 10%, so 10% is the top of the range in CSI 464. Wick size per jar diameter is not published; this round is how you find it.

Round 3, new-oil screen. You have not chosen the oil yet. Do not buy a hundred candles' worth of anything. Buy a 15 g tester of each candidate and pour three small test pots of each. The Heart Cutout Velvet Jar is listed with a 50 g wax capacity, so a 50 g pot at 8% (4 g oil + 46 g wax) fits inside its stated capacity. Three pots per oil use 12 g, leaving 3 g of each tester spare. For three Indian florals, Jasmine, Raat Ki Rani and Gardenia, the three testers cost ₹338.02.

Stacking tests onto production

The order of operations is the same whichever round you choose. Start from the net, add the pour loss as a percentage of the production oil, add the test round as whole candles, and round to real bottles last. Do not add the pour percentage to the tests as well. Test candles are poured from small melts, and their residue is part of what the round teaches you.

The full matrix costs ₹860.00 more oil than no testing at all, or ₹8.60 per saleable candle. Set that against the alternative: a hundred candles with a tunnelling wick or a load the customer finds too faint, all poured before anyone lit one. The testing allowance is the cheapest insurance in the order.

The test round also changes the wax and wick counts. The matrix adds 2,208 g of wax, twelve wicks (six of each wick option) and twelve jars. Where the round tests two wicks, you need both wick types for the tests, and then a full production quantity of the winner only. How many wicks to buy for a wick-size test goes into that in detail, and development and QC samples for 1,000 candles scales the idea up.

One order or two

Buying everything at once is simpler, and for a proven recipe it is the right call. For a new oil or a new jar, buying in two stages protects you: the test oil first, then the production oil once the round has passed. The worry is that small bottles cost more. For Jasmine they do not. Its 100 g is ₹4.30 a gram and its 500 g ₹3.92, and the two-stage route below uses exactly the same bottles as the one-order route:

On both oils the two routes cost the same, because the test round fits exactly into two 100 g bottles and the production quantity falls on the same larger sizes either way. Stage 1 leaves 8 g over, which is taken off the stage 2 need. The only cost of staging is time: the second order arrives after your test candles have cured and burned. So the real rule is timing. If the delivery date allows a test round and a second order, stage it. If it does not, you are buying production oil for a recipe you have not tested, and the matrix becomes a check you run in parallel with production rather than before it.

The priced order

Here is the full order for the hundred candles with the Round 2 matrix, bought together (for a staged purchase, split the oil line as shown above). Wax is Luxury Soy Wax Chunks, rated up to 13%, so the 10% test candles sit comfortably inside the ceiling. Production wax is 18,400 g net, 18,952 g with 3%, and the matrix adds 2,208 g, so the wax need is 21,160 g.

That is ₹291.31 per saleable candle with the whole test round absorbed. Labour, packaging, freight and GST are extra. The wick line assumes both test wick options are Thin C1 at different setups; if your second option is Thick C2 (₹94.40 for 10, ₹188.80 for 20), buy six of those separately. Wick sizing for your jar is exactly what the round is there to find out.

Two related calculations sit either side of this one. Whether one kilo covers the same hundred candles at 6% shows how quickly a lighter load still outruns a single bottle, and how much to buy of a new fragrance with unknown demand applies the test-first logic to launch quantities.

Once the round is done, keep one candle from each winning cell as a retained reference, labelled with the oil and the date. It costs 16 g, and it is the only way to compare a future lot of the same oil against the one you approved. Lot-to-lot consistency is not published in CSI's data, so if a repeat order must match, ask on WhatsApp and burn a new candle beside the reference before pouring the next run.

Frequently asked questions