I'm Blending Three Fragrance Oils for a Candle and Want an 8% Total Fragrance Load — Is 8% Applied to Each Oil or to the Combined Fragrance Blend?

The rule: one load, one blend

Fragrance load is a property of the finished candle. At CSI, and throughout this series, load means the share of the finished candle's weight that is fragrance oil: oil = finished weight × load, and wax = finished weight − oil. A 200 g candle at 8% is 16 g of oil and 184 g of wax. That definition does not change when the 16 g comes from three bottles instead of one. The candle still has 16 g of fragrance to melt, carry and release, and the wax still has to hold 16 g of it.

So the order of decisions is always the same. First choose the total load for the candle, the same way you would for a single oil, and keep it inside your wax's ceiling. Luxury Soy Wax Chunks are rated up to 13% and Luxury Soy Wax CSI 464 up to 10%. These are the wax makers' stated formulation maximums, not targets and not regulatory limits. Then choose the ratio, the recipe of the fragrance itself. The ratio is written in shares that add up to 100%, such as 50/30/20 or 60/40. Multiply the blend grams by each share and you have the weights to put on the scale.

The second and third rows are the same misunderstanding in two costumes. Both give each oil its own 8%. The first keeps the candle at 200 g and starves the wax, leaving 152 g of wax to hold 48 g of oil. The second keeps all 200 g of wax and ends up with a 248 g candle at 19.35%. Either way the load is roughly double or triple what you intended, and above every ceiling in CSI's range. Neither is a stronger candle. Oil beyond what the wax can hold commonly shows up as sweating, soft tops, a wick that drowns or soots, and a scent that is harsher rather than louder. Blend sweating covers what to do if you have already poured a batch like that.

Worked grams, candle to batch

The formula scales without change. Here it is for the 50/30/20 example at every size a small brand is likely to pour, using the finished-weight convention. Each component is the blend grams multiplied by its share, rounded to two decimals for the scale.

If your batch starts from a fixed amount of wax rather than a target finished weight, use the other form of the same rule. To reach a true 8% on 1,000 g of wax, the blend is 1,000 × 0.08 ÷ 0.92 = 86.96 g, giving a 1086.96 g batch. Split at 50/30/20 that is 43.48 g, 26.09 g and 17.39 g. If you simply add 80 g to 1,000 g of wax, the true load is 80 ÷ 1,080 = 7.41%. Neither method is wrong as long as you know which one you are using and write it on the formula card. The 80 g vs 86.96 g question works through the difference in full, and the vanilla-coffee 1 kg wax batch shows both conventions side by side for a gourmand blend.

Three ways a blend ends up overloaded

Makers who are careful with a single fragrance often overload blends, and the reasons are procedural rather than chemical. The recipe is written one way and weighed another, or a correction is made in the jar rather than on paper. Here are the three patterns worth checking for whenever a blended candle behaves worse than its components did alone.

If a blend has already sweated or tunnelled, audit the grams against these three patterns before you blame the oils. The arithmetic is usually quicker to check than a new round of test candles, and it is the first step in diagnosing a sweating blend. If the grams are right, the question becomes whether the oils are happy together, which is the compatibility check.

What each oil sees inside the blend

Getting the total right has a consequence that surprises people. If the whole blend is 8%, each component is present at less than 8%. In the 50/30/20 example, lavender is 4.0% of the candle, sandalwood 2.4% and bergamot amber 1.6%. Each oil is now working at a fraction of the load you might have tested it at alone.

This is the arithmetic behind a common experience: two oils that each smell fine alone produce a blend that smells thinner than either. Nothing has gone missing. Each oil is simply at a lower concentration, and fragrances commonly overlap or mask one another in the blend, so the total impression is not the sum of the parts. It is also why a small accent component, like the 0.80% slot in a four-oil blend, is easy to lose entirely once the candle is burning. Two good oils, one weak blend takes the dilution argument further, and how many components to start with covers when adding a fourth oil stops helping.

It does not mean you should raise the total load to compensate. Raising the total from 8% to 10% raises every component by a quarter, including the one that was already dominant, and brings the candle closer to the wax's ceiling. The better first lever is usually the ratio: give the shy component a bigger share inside the same total. The oud-rose ladder shows why that is almost always the cleaner test.

What the blend costs per candle

A blend's price per gram is the weighted average of its components' prices. At 100 g pack prices, Lavender is ₹5.40 a gram, Sandalwood ₹5.03 and Bergamot Amber ₹5.60. At 50/30/20 the blend costs ₹5.33 a gram, so the 16 g in one candle costs ₹85.25.

Material cost here is oil, wax, wick and jar only. It excludes labour, packaging, freight and GST. Two things follow from the table. First, blending does not make fragrance cheaper or dearer by itself; the blend costs exactly the weighted average of what goes in, so a blend led by a premium oil costs close to that oil. Second, the ratio is a cost lever as well as a scent lever. Swapping the 50% and 30% slots (sandalwood-led instead of lavender-led) changes the blend rate to ₹5.25 a gram. The difference is small here because the three oils are close in price, but with a premium oil in the mix a ratio change can move the candle cost by ₹10–₹20.

Testing the blend's total load

The load that suits each oil alone is a starting point for the blend, not an answer. The blend has a different mix of materials, it may melt and release differently, and it may need a different wick. The cleanest first test holds the ratio fixed and moves only the total load, in three candles poured on the same day from the same wax, wick and jar:

The three candles need 24.00 g of lavender, 14.40 g of sandalwood and 9.60 g of bergamot amber. A single 15 g of lavender will not cover it; a 50 g will, with 26.00 g to spare for a ratio test later. Sandalwood and bergamot amber fit inside their 15 g testers with 0.60 g and 5.40 g left over. Cure all three the same way and burn them at the same age. A labelled checkpoint such as 7 days is a reasonable first look, but it is a working checkpoint, not a specification; burn-testing at the same cure age explains why the age must match.

Blend-test kit and shopping list

CSI does not sell a blending kit, sample set or pre-mixed blend; custom blends are your formula, and CSI sells the single oils. You build a test kit from the smallest sizes. Here is the kit for the three-candle load test above, with every component priced at CSI live pricing, pulled 24 September 2026.

Wick size for your jar is not something this page can assign; the Thin C1 is listed as the default in our costing example, and you should confirm it with a burn test in your own jar. If you already own wax, jars and wicks, the oils alone are ₹468.24. The spare wax (448.00 g), two jars and the leftover oil are enough for a second round once you know the load, usually a ratio test.

When the blend is approved, cost production at 100 g or larger. Fifty 200 g candles at 8% need 800 g of blend: 400 g of lavender, 240 g of sandalwood and 160 g of bergamot amber, plus the 3–5% extra we suggest as a working assumption for pour loss and test burns. Lavender's 500 g at ₹2,620.00 covers the first comfortably, and 100 g packs of the other two cover theirs.

From here the cluster splits by the problem you are solving. For a fixed 1 kg wax batch with a named gourmand ratio, see the 50/30/20 vanilla-coffee-caramel conversion. To judge a blend in wax rather than in the bottle, see evaluating blends in wax. To pull a dominant note back into line, see lavender and vanilla or oud and rose. For hot-throw balance, cure drift and blend failures, see citrus-woody balance, coffee-vanilla after cure and two good oils, one weak blend.

Frequently asked questions