Candle-Safe Fragrance Oil vs General Fragrance Oil

"Candle-safe" is a claim, not a category

Start with what the phrase can and cannot be. A category is defined by someone other than the seller — a specification, a standard, a register. "Candle-safe" has none of those behind it, so it belongs to the same family of words as "premium" and "professional grade": it tells you what the seller wants you to conclude, not what they have committed to. A claim is still useful, though, because a claim can be interrogated. Three questions do it.

Who is making the claim, about which exact product, and what have they published under it? The first question rules out category pages and marketing headlines — a claim on a collection banner is not a claim about the bottle in your hand. The second rules out inference from siblings: a document for one oil says nothing about another, and one page in a range mentioning IFRA does not put IFRA behind the rest. The third is the whole of this guide, and it has six parts.

The six-point evidence scorecard

These six are not arbitrary. Each one changes a decision you have to make before you pour, and each one is either printed or it is not — there is no judgement involved in scoring it. Use it on any supplier's listing, ours included.

What the scorecard says about 107 real pages

A scorecard nobody has run is a rhetorical device. So here it is run across every fragrance-oil listing CSI published in September 2026, including the ones that score badly. The distribution is in the diagram above; the pattern underneath it is more useful than the headline.

Fifty-one pages score five of six, and 42 of those are missing only a measured flash point. That is one gap, repeated — and it is the gap most likely to be filled by a guess. Thirty-two pages state only "Flash point > 93°C (suitable for alcohol-based products)", which is a threshold for alcohol-based products and not a measurement of the oil; 56 state nothing about flash point at all; and one, Zesty Lemon, gives the most honest version of the answer by stating it as "Available on request / SDS". Only 18 pages publish a measured value, and those values run from "Flashpoint: 85°C" on Neroli to "Flashpoint 206°C" on Vetiver.

At the other end, 19 pages score nothing at all. Almost all of them sit in the fine-perfume side of the range — listings written for perfumery, solid perfume and diffuser formulators, where the copy is about the accord rather than about wax — plus one diffusion concentrate base that is not a candle product in the first place. A zero is a statement about the listing's audience, not about the oil's suitability. If you want one of those oils in a candle, that is a message to send, not a conclusion to draw.

Candle-first listing vs general listing, side by side

In practice the two kinds of page you will meet are a candle-first listing, written for someone pouring wax, and a general or perfume-grade listing, written for someone formulating across many categories. Both can be perfectly usable. They simply hand you different things, and knowing which you are holding stops you misreading a gap as a defect.

What to ask for, and what each answer actually proves

Two documents, requested by exact product name and size, and one confirmation. That is the whole list, and it is short because most of what you need is already printed. What matters is knowing what each item does — because a maker who thinks documentation proves performance will skip the burn test, and a maker who thinks a burn test replaces documentation will have nothing to say to a customer.

The four-jar test that settles it

Four jars, because you need three data points and a baseline. Jar A at the load you intend. Jar B at the same load on a different wick size. Jar C at two points lower, which is your escape route if A and B are close. Jar D unscented, from the same melt, which is the only honest way to judge colour a fortnight later. Same wax, same vessel, same add temperature, same stir time, all written down before you start.

Pour them, look at day one for an oily film, beads under the lid or a wet ring at the glass — those mean the load or the temperature is wrong, so re-pour rather than wait. If the tops are clean, lid them, box them out of sunlight and leave them. Judge nothing until day 14; the oil pages' 48–72 hour windows are the floor for a first burn test, and standard practice for soy containers is 14–21 days. Then burn A, B and C for up to three hours each in one closed room with nothing else scented, and score seven things: minutes to a full melt pool reaching the glass, pool depth, flame height and stability, soot or mushrooming, throw at 30 minutes, throw at two hours, and the wax colour against jar D. Copy the sheet in how to test hot throw before launching a candle fragrance.

Three listings that score six of six

Eight of CSI's 107 pages carry all six pieces of evidence. Three of them are below — chosen because they publish different kinds of answer rather than because they are the strongest smelling. Each card shows every size, the price per 100 g, and how many 200 g candles at 8% each pack covers: 16 g of oil and 184 g of wax per candle. The 15 gm trial reads ≈0 because it is a smelling size, not a full candle.

Frequently asked questions