Ask a candle maker to strengthen a weak scent and they will do the same thing: add more oil. It feels like the obvious lever. It is usually the wrong one. In three years of watching small-batch and commercial candle production, the fragrance load has caused more failed SKUs than any wick, any vessel, and any pour temperature combined — because it is the one number where “more” quietly becomes “worse” and nobody catches it until the candles are in a warehouse.
This is not a “how much fragrance do I use” post. The internet already answers that with a lazy 6–8% range. What this is: where the ceiling actually sits, why the math most shops do is wrong by a full percentage point, why 10% candles often smell weaker on burn than 8% ones, and a test protocol that tells you what your batch can actually hold instead of what a forum says it should.
The Math Most Shops Get Wrong: Wax Weight vs. Candle Weight
Before the chemistry, the arithmetic. Fragrance load is defined as fragrance oil weight divided by wax weight. Most calculators and most production sheets do this correctly:
For 1,000 g of soy wax:
- 6% load = 60 g oil → 1,060 g finished wax
- 8% load = 80 g oil → 1,080 g finished wax
- 10% load = 100 g oil → 1,100 g finished wax
Now the mistake. A number of production sheets quote the load as a percentage of the finished candle (wax + oil together). A “10% of the candle” batch made from 1,000 g of wax carries only about 91 g of oil — a 9% wax-weight load — yet it gets logged, labeled, and replicated as 10%. The inverse error is worse: someone who hits 10% of candle weight with oil is really at ~11% of wax weight, past the ceiling of most soy grades, and won’t know it until the candles sweat.
Pick one definition, print it on the batch sheet, and never mix them. If your supplier’s IFRA sheet and your wax mill’s recommendation use different bases, convert before you compare. Every “my 10% recipe works here but sweats there” thread we have ever seen starts here.
Where the Ceiling Actually Sits
Industry guidance is consistent enough to treat as a spec, not a suggestion:
- 6% is the standard development baseline. It is the load at which a fragrance formula’s true character is visible: if a scent needs more than this to register, the problem is usually the formula, not the quantity.
- 8% is the common commercial optimization point for soy container candles — the load where most balanced formulas stop leaving scent on the table.
- 10% is the maximum the major wax manufacturers will stand behind for soy. It is a ceiling. Nobody serious designs a recipe at the ceiling the way you’d design at a material’s proof load.
Two numbers matter as much as the load itself. First, the addition temperature: major suppliers specify adding fragrance at roughly 185°F / 85°C for soy — hot enough for the oil to disperse fully through the melt, cool enough that the most volatile top notes have not already started stripping out of the batch before it ever reaches the jar. Second, the cure: 48 hours of “resting” tells you almost nothing; the industry reference point is a two-week cure before a formula is judged. A 6% candle tested at day one and an 8% candle tested at day 14 are not comparable, yet they get compared every day.
The Paradox: Why 10% Often Smells Worse on Burn
This is the part that keeps non-chemists honest. Push a soy candle past its real binding limit and three things happen, and all three reduce the scent in the room:
- The flame starts working for the oil instead of the wax. Fragrance oil is far denser and heavier than wax. At 9–10% in a wick chosen for a 6–8% batch, the melt pool carries an excess of unburned organics. The tell-tale symptoms are not subtle: flame surging, repeated self-extinguishing, heavy mushrooming at the wick tip, visible soot on a still-air burn, and a melt pool that runs deeper and hotter than the vessel was sized for. You have not made a stronger candle; you have built a less stable one.
- The oil stops binding and starts separating. Sweat — a sheen of free oil that migrates up the side of the candle and pools at the surface — is the physical signature of a load the wax matrix cannot hold. It is not a cosmetic flaw you can sand out. It is the wax telling you the recipe is over. Bleeding through the container wall in thin-walled vessels is the same failure at higher severity.
- The hot throw thins out. This is the counterintuitive one that ruins launches. A formula can smell loud straight from the bottle — volatile top notes hit the nose instantly — yet produce a thin, fading hot throw because it lacks the persistent heart and base materials that survive combustion and diffuse through a room. At 9–10% some fresh formulas (citrus, tea, marine, green) don’t get richer; they get sharp and chemical, and the perceived scent in the room is lower than the 8% version. The room does not grade percentages. It grades airborne molecules.
So the practical rule is not “use 8%.” It is: the load is the lowest percentage that meets your throw, appearance, and burn-safety targets in your specific wax, wick, and vessel — and above about 8%, every additional percent has to earn its place with documented evidence, because the ceiling is a spec, not a target.
A Test Protocol That Actually Answers the Question
The single biggest gap between “we burn-tested it” and “we know what this candle does” is that most shops change two variables at once. A proper load test isolates the load and holds everything else fixed — same wax lot, same fragrance batch, same vessel, same room, different days.
Run this matrix:
| Cell | Load | Wick | What it answers |
|---|---|---|---|
| A | 6% | Baseline wick | The fragrance’s true character and the reference throw |
| B | 8% | Baseline wick | Whether extra oil buys a measurable gain or just risk |
| B2 | 8% | One size larger | Whether the 8% scent is wick-limited or oil-limited |
| C | 10% | Baseline wick | Drowning behavior: flame stability, soot, extinguishing |
| C2 | 10% | One size larger | Whether a heavier wick makes 10% viable — or just hotter |
Scoring checkpoints, in order:
- Cold throw at 24h, 48h, day 7, and day 14 — recorded in the same room, same time of day. This is where cure time shows up as data instead of folklore.
- Hot throw by burn cycle: first burn (full melt pool), second burn, and a burn at the end of the candle’s life. Note when the scent peaks and, just as importantly, when it collapses into waxy/ashy territory.
- Flame and combustion: height, surging, self-extinguishing events per burn, soot on the wick and the rim, mushrooming, and whether the melt pool stays inside the vessel’s thermal limit.
- Surface and wall condition after day 7: sweat lines, free oil at the surface, discoloration, frosting severity. Log it on day 14 too — separation is a time-delayed failure.
Record the variables that shops habitually skip: wax lot and production date, fragrance batch number, addition temperature, mix time, pour temperature, vessel dimensions, wick series and size. Two cells that look identical and smell different are almost always one of those, not “the fragrance changed.”
Then apply the decision rule: pick the lowest cell that hits your throw target with clean combustion. If that is 6%, shipping a 10% version is not stronger product — it is a more expensive, less stable one.
If You Must Run 9–10%: Treat It as a New Product
Some formulas and some markets genuinely need it. When the 8% candle produces a documented performance gap — say, a heavy gourmand that genuinely under-performs in a large room — moving to 9–10% is defensible. What is not defensible is pouring more oil into the same recipe with the same wick and calling it a tuning pass.
At 9–10% the wax-to-oil ratio has changed enough that the combustion chemistry is a different system. Reopen wick selection from scratch (expect one size up, and sometimes a different series, not just a larger tab), re-run full-burn and end-of-life observations, re-check vessel wall temperature, and re-confirm the wax supplier’s stated maximum in writing for that specific grade. A supplier who says “use 10% for strong scent” without asking which wax, which wick, which jar, and which room is selling oil, not solving your candle.
One adjacent trap worth naming: IFRA compliance is not a load spec. A fragrance can carry an IFRA Category 12 allowance comfortably above 10% while your soy wax physically stops holding scent well at 8%. The two ceilings are independent, and the lower one is your product spec. IFRA tells you what you are allowed to do; the wax matrix tells you what will actually work.
The Bottom Line
Fragrance load is the one variable in candle production where the naive direction of the lever is wrong. Six percent is where a formula shows its real character, eight is where most commercial soy candles should live, and ten is the edge of the cliff — past it, sweat, soot, and a thinner hot throw arrive as a package. The shops that win on scent are not the ones loading the hardest. They are the ones whose test data can show, cell by cell, exactly how much oil their specific batch can carry — and stopping one percentage short of the point where it stops carrying cleanly.