I Created a Fragrance Blend That Smells Perfect Neat but Completely Different in Soy Wax — Should I Evaluate Blends in Wax Before Finalising Their Ratios?

Why the beaker lies

When you mix two oils in a beaker and sniff, you are smelling the blend at full strength, at room temperature, from an open surface. Nothing about that resembles a candle. In a 200 g candle at 8%, the same blend is 16 g spread through 184 g of wax. It reaches the air in two quite different ways. Cold throw is what escapes from the solid surface of a cured candle. Hot throw is what the melt pool releases while the candle burns. A blend can do well in one and poorly in the other, and neither is the beaker.

Several mechanisms are commonly described for why the balance changes. Dilution changes perception: at full strength one component can seem dominant, and at 8% a quieter one can come forward, or the reverse. Wax interaction matters: soy wax is described as holding some aroma materials more readily than others, and it crystallises as it cures, which can change what the cold surface gives off. Volatility matters in the burn: lighter, more volatile materials tend to lift first from a hot pool, heavier base materials later and more slowly. None of this is measured per oil in CSI's data, and none of it can be predicted from a sniff. It can only be observed, which is the point of testing in wax.

The single-oil version of this problem is covered in why bottle strength does not predict candle performance and why a scent changes once the candle is burning. Blends add one more variable: the ratio. A ratio that is balanced in the beaker is only balanced at 100% strength and room temperature. The job of the funnel below is to find the ratio that is balanced where it matters.

What each stage can tell you

Each stage of testing answers a different question and costs a different amount. The table sets them side by side, including the stage most makers skip: the unwicked cold-throw pot.

The cold pot is the stage that saves money. It uses a quarter of the oil of a full test candle, needs no wick and no burn time, and answers the question "what does this ratio smell like at 8% in cured wax?" For containers, the Heart Cutout Velvet Jar is listed with a 50 g wax capacity at ₹45.00 each or ₹225.00 for five. Any small, clean, heat-safe container works; weigh the fill rather than trusting a line on the glass. What the cold pot cannot tell you is how the blend behaves in a melt pool. A 50 g pot also should not be lit as a stand-in for your 200 g candle: a different diameter and wick give a different pool, so its hot throw would not be comparable.

The three-stage funnel, in grams

The example is a two-oil blend of Jasmine and Palo Santo, screened at five jasmine shares: 80, 65, 50, 35 and 20%. Every pot is 50 g at an 8% total load, so 4 g of blend and 46 g of wax. The load is fixed; only the ratio moves. That is the rule from the blending hub: one load for the whole blend, split by ratio.

Each component weight in a pot is small: 0.80 g of palo santo in P1, 0.80 g of jasmine in P5. A scale that reads to 0.01 g is ideal for this stage. If yours reads to 0.1 g, mix a larger master blend for each ratio (say 10 g) and weigh 4 g of it into the pot; the ratio is then set on bigger, more accurate weights. Pour all five pots from the same wax melt on the same day.

After the pots have cured to your checkpoint, pick the two that come closest to the brief. In the worked example these are the 65/35 and the 50/50. Pour each as a 200 g wicked candle at 8%: 16 g of blend and 184 g of wax.

Across both stages the example uses 28.4 g of jasmine and 23.6 g of palo santo. The last column costs each finalist at 100 g rates (Jasmine ₹4.30 a gram, Palo Santo ₹6.02): the two ratios differ by only ₹4.12 a candle, so cost should not decide between them. If the finalists were further apart in price, it might. Comparing oils by cost per finished candle covers that decision.

Scoring the pots and the burns

Makers tend to choose the strongest-smelling test. For a blend, strength is only half the question; the other half is whether the balance you designed survives. Score both, separately, on a 1–5 scale. A simple sheet looks like this:

The example scores are placeholders to show the shape of the sheet, not results; CSI publishes no throw data for any oil. The rule for reading the sheet is simple. If a component scores 1 or 2 for noticeability in the burn, the ratio has failed for that component, however strong the candle is overall. Rebalance by moving the ratio inside the same total load before touching the load. The ladders in lavender and vanilla and oud and rose show how. If a note vanishes only while burning, hot-throw balance is the page for it.

Is the funnel worth it?

The alternative to the funnel is to pour all five ratios as full 200 g candles and burn them. That works, but it costs more oil, more wax, more jars and five times the burn hours. Here is the comparison at 50 g oil rates, with the 50 g Heart Cutout Velvet Jar at ₹45.00 as the pot:

On materials the funnel is ₹350.03 cheaper for this two-oil blend. The bigger saving is time: two burn tests instead of five, and each burn test is several hours of supervised burning at the same cure age. With three oils, or more ratios, the gap widens. The funnel's weakness is that the pots cannot show hot throw, so a ratio that is dull cold but blooms when burning can be eliminated too early. If you suspect that, keep a third finalist. Weak cold throw, strong hot throw explains why that pattern happens.

Priced blend-test kit

CSI does not sell a blending kit, sample set or pre-mixed blend; you build the kit from single oils. Here is the one for the Jasmine + Palo Santo funnel, at CSI live pricing, pulled 24 September 2026.

If you already have wax, jars and wicks, the oils and pots together are ₹667.40. The wax left in the bag (402 g) covers a second round, such as a third finalist or a load test on the winning ratio; the hub shows that test. Wick size for your jar is not something this page can assign; confirm it with a burn test.

Once the funnel gives you a formula, the rest of the cluster takes it forward. Coffee-vanilla after cure covers balance that shifts during cure. Two good oils, one weak blend covers blends that throw poorly overall. Compatibility across suppliers and blend sweating cover physical problems. Formula first or wax and wick first? covers how the blend and the candle system are finalised together.

Frequently asked questions