You pour two identical batches on the same day. Same soy wax, same 8% fragrance load, same wick, same 14-day cure. One batch fills the room for the customer. The other one burns fine, looks fine, and smells like a faint suggestion of lavender. The batch log shows nothing wrong, because nothing in the batch was wrong.
When a candle line loses its scent batch to batch, the first place to look is almost always the ingredient you cannot inspect with your eyes: the fragrance oil. This is the sourcing problem behind most of the complaints that read like quality problems — weak throw, scent that dies after a few burns, one container in a dozen that underperforms.
The root cause is a trust error. Most candle makers — and a surprising number of manufacturers — treat the supplier’s spec sheet as a lab report. It is not. It is a sales document. And the three numbers that actually determine whether your candle smells — true dilution, IFRA version, lot-to-lot variance — are exactly the three most likely to be smoothed over on that sheet.
Why the spec sheet lies by omission
A fragrance oil spec sheet typically lists flash point, color, weight, and a perfumery description. Read a dozen of them and you will notice what is missing: carrier composition, the IFRA certificate version and date, and any per-lot analytical data. Each omission maps to a different failure mode in your finished candle.
Dilution is the first. “100% pure fragrance” on a price point that looks too good is almost never 100% pure. Common practice is to cut blends with isopropyl myristate, mineral oil, or other inert carriers to hit a price and a pour weight. The supplier is not cheating on the invoice — the weight is right. But your effective fragrance concentration is 60–80% of what you think it is. At an 8% load of a 70% blend, you are really loading 5.6% of fragrance. Budget retail candles run 3–5% loads as a cost measure; that is their honest spec. What you have done is accidentally land in the same band while paying premium prices, and your melt pool carries less volatile material than the recipe assumes.
IFRA version is the second. IFRA reformulations happen on a schedule, and a blend legal in one version can be over limit in the current one. The certificate matters more than the claim. “IFRA compliant” with no version number, date, or certificate ID is a claim, not a compliance document. If a retail buyer in the EU asks for your IFRA documentation and you hand them a one-line email reply, that conversation ends there.
Lot variance is the third, and the quietest. A fragrance house blends from natural and synthetic raw materials whose proportions shift season to season. The same FO name from the same supplier can carry a different top-note ratio in January than in July. Your recipe was tuned once; it does not re-tune itself. Customers do not file complaints under “seasonal raw material variance.” They file them under “candle stopped smelling.”
Test 1: Measure the flash point, do not trust the one on the sheet
Flash point is the single easiest number on a spec sheet to verify, and the easiest to fudge. A clean, undiluted fragrance oil typically tests between 150°C and 190°C (302–374°F). Run a closed-cup test on a sample of 50–100 ml, take the measured value, and compare it to the sheet.
Three scenarios, all readable from the gap:
- Measured ≥ sheet value. Normal. The supplier tested with margin. Move on.
- Measured 10–15°C below sheet. Mild carrier dilution, or a different blend cut than the reference lot. Ask for per-lot data.
- Measured 20°C+ below sheet. The product is not what the sheet describes. The weight is real; the fragrance is not. Re-price it as a diluted blend and re-tune your load — or re-source.
A closed-cup apparatus costs less than one bad batch. That is the entire business case for Test 1.
Test 2: The two-load burn differential
This is the test that exposes dilution in finished-candle form, and it doubles as your baseline for every new FO.
Pour two identical 8 oz jars with your reference wax. Jar A gets a conservative 6% load of the candidate FO. Jar B gets 10%. Cure both 14 days, same wick, same trim. Burn each for a full melt-pool cycle (the rule of thumb is roughly one hour per inch of diameter — for an 8 oz jar, 30–45 minutes, and the pool should reach the glass). Then walk into the room from the next space and rate the hit on a 1–5 scale.
A genuinely full-strength fragrance will show a clear step between the two loads — you should be able to rank them blind. If the 10% jar cannot separate from the 6% jar, the volatility content is lower than the sheet implies, and no wick change in the world will recover the difference. Note the failure mode: a diluted blend also shifts the top/mid note balance, so the candle does not just smell weaker, it smells different from the sample your buyer approved.
Test 3: The cold-shock stability check
This one protects your shipping lane, not your burn test. Put a sealed sample through a 24-hour freeze at −5°C (23°F) — a winter truck, a February port — then bring it to room temperature and inspect.
What you are looking for: cloudiness that does not clear within 24 hours, visible phase separation (a layer, a film, droplets suspended), or a pour that has turned stiff. Clean fragrance oils re-equilibrate quickly. A blend leaning on cheap inert carriers shows its work at low temperature, and the separation you see in the sample is the same separation happening inside finished wax after a cold shipment — uneven throw, sweating on the surface, a top layer that smells nothing.
Waxes and carriers behave differently below 10°C (50°F). If your product line ships into northern markets or cold-chain logistics, this 24-hour test is the cheapest insurance you will buy all year. And if you sell flameless lines for exactly these cold environments, the customer doing the comparison already knows the difference: our battery LED bullet candle line exists because the real thing has a cold-weather failure mode that a flame does not.
Test 4: Audit the IFRA certificate, line by line
Ask for the actual certificate, not a screenshot of a statement. Then check four fields against your purchase order:
- Certificate number — unique per formulation. If two of your FOs carry the same certificate number, they are either the same blend (in which case the names are marketing) or the paperwork is wrong.
- IFRA version and date — confirm it is the current version for your market. A certificate issued three years ago may still name a blend that is now over limit for candle use in the EU.
- Use category — the certificate must cover candle use. “Aroma – candle” and “aroma – general” are different rows with different maximums.
- Name match — the blend name on the certificate should match the supplier’s product name, or there should be a documented cross-reference. One word off, and the document may describe a different formulation.
This takes ten minutes with the paperwork in front of you. It is the difference between a compliance file a distributor will file in a drawer and one they will forward to their QC lab.
Put the spec on the purchase order
After the four tests, write the verified numbers back into the relationship. The purchase order should carry: measured flash point with a tolerance (say ±5°C from measured, not from sheet), IFRA certificate ID and version, carrier composition if any, and a per-lot data requirement — at minimum flash point and refractive index per lot, delivered with the invoice. Refractive index is the lot-consistency number a serious fragrance house will happily report, because it is cheap for them to measure; if they will not give it, that is an answer about where they source.
Run the same four tests on every lot change, not just the first. A fragrance that passed in March can fail in September, and the failure will land in your return rate, not in your supplier’s.
The practical bottom line
Weak throw, scent that fades after a few burns, one-in-a-dozen duds — these read as recipe problems because the recipe is the only thing you can see. But the recipe is downstream of the ingredient, and the ingredient is the only variable in the chain that is specified on a piece of paper instead of in your own lab. Measure the flash point. Run the two-load differential. Freeze the sample. Read the certificate line by line. None of it costs more than one failed batch, and together they convert the supplier’s spec sheet from a promise into a number you can hold them to.
YS Candle manufactures wax and flameless candles for wholesale, and runs its own incoming FO audit on every new fragrance before a recipe is locked. If your line has a throw problem you cannot explain in the batch log, that is usually where the answer lives.