How to Tell If a Candle Fragrance Oil Is Good Quality

Why the bottle cannot tell you

Hold an uncapped bottle under your nose and what reaches you is concentrated liquid evaporating from a small surface in cool, still air — overwhelmingly the lightest materials in the blend. That is the top of the structure. In a candle the same fragrance has been diluted to 6–10% of the mass, bound into a solid, cured for weeks, then released from a warm pool a few centimetres across in a room with air moving through it. Almost nothing about the first situation predicts the second.

So the two most confident statements in candle-making conversation are usually both wrong. "It smells incredible, it will be a great candle" ignores that a top-heavy oil can vanish by minute forty. "It smells weak, it is a bad oil" ignores that many base-weighted oils are quiet in the bottle and still filling a room at minute ninety — and that an oil is diluted roughly ten-fold on the way into a candle, so the two intensities are not on the same scale.

One thing a bottle sniff is genuinely good for: elimination. If a direction is wrong for your brief — too sweet, too sharp, too medicinal — you will know in two seconds and you have saved a jar. Use the bottle to say no. Never use it to say yes.

The ten-point check: six free, four earned

Here is the whole assessment. The first six are free and done on the page before any money moves. The last four require candles, and no amount of reading replaces them. The column that matters most is the last — what each check does not prove — because almost every bad fragrance decision is a check asked to carry a conclusion it cannot support.

Notice how the checks change character halfway down. The first six are audit — what the supplier has committed to in writing, which is also what you can quote to a stockist. The last four are experiment — what happens in your workshop, which is what decides whether a customer reorders. Do only the first six and you buy well and pour blind; do only the last four and you eventually pour a candle you cannot document. The free half exists to tell you which candidates deserve the expensive half.

Six things that look like quality and are not

This is the uncomfortable part of an honest buying guide, because a supplier writing it has to point at its own copy. CSI's pages are written to sell, and several phrases on them are not evidence of anything. Here they are, with what to ask for instead.

Two more traps. "Phthalate-free" and "skin-safe" are stated on some pages and not others — 35 of 107 for each — so neither is a property of the range; request it for the specific oil if a stockist needs it in writing. And "skin safety tested", on 32 pages, names no test, standard, laboratory or date. Not a lie; simply not evidence you can hand to anyone.

Does a dearer kilo buy more evidence?

Here is the test almost nobody runs on a fragrance range: score every oil on how many of the six free checks its page passes, then sort by price. If price signalled quality, the expensive end would be better documented. On live CSI stock in September 2026, across the 87 oils with a 1 kg pack available, it is the other way round.

The reason is not that expensive oils are worse. The dearest listings are the perfume-grade and designer-inspired directions, built on a different page template: rich note pyramids, a broad usage table across seven product categories, and the >93°C threshold line instead of a measured flash point. They are also newer, which is why nobody has reviewed them yet. The long-standing single-note candle oils have simply accumulated both measurements and buyers. Price tells you about perfumery ambition and raw-material cost. It does not tell you about documentation, and it says nothing about throw in your jar.

Three oils that answer all six free checks

Seven oils pass all six. These three are the most useful to assess first, because each one demonstrates a different kind of published evidence — behaviour in a candle from a real buyer, a measured flash point at the low end of the range, and a stated vanillin figure that changes your format decision. Passing the audit is not a performance verdict: the burn test still has to happen.

The blind audit, costed

Every bias in fragrance assessment runs one way: you want the new oil to be good, you know which jar is which, and you know what each cost. So the protocol has one governing rule — the person scoring must not know which candle is which — and one structural feature: the candidate is judged against a benchmark you already sell, never against a memory. Six jars settle checks seven to ten.

Writing the verdict down

A quality assessment that lives in your memory has to be repeated every time someone asks. Write it as four short lines and it answers a customer, a stockist and your future self without any of you having to smell anything. Line one — what the page states, verbatim, with blanks left blank. Line two — what you tested: wax, load in grams, wick, vessel, cure days. Line three — what happened: pool time, cold and hot throw scores against the benchmark, surface at the ceiling, colour against control. Line four — the decision, with a date and a price.

The last line keeps you honest, because it forces you to say what the verdict is for. An oil can be approved for amber 200g jars at 8% and refused for white 150g jars — a normal outcome, invisible in any system that reduces an oil to "good" or "bad". Keep a sealed, dated 15 g retained sample from the bottle you tested, and request the SDS and IFRA documentation for that oil on WhatsApp before a stockist asks.

Frequently asked questions