One Fragrance in My Collection Has Weak Throw While Every Other Scent Performs Well in the Same Wax and Jar — Should I Formulate That Fragrance Separately?

Is it really an outlier?

Ranges grow one scent at a time. The weak one was usually poured on its own day, from its own bottle, judged at whatever age it had reached when you got round to it, and compared with your memory of the others. Any of those can make a scent look like an outlier. So the first stage is not a fix; it is a check that the problem belongs to the scent at all. Pour the weak scent beside one that throws well in your range, from one bag of wax, same wick, same jar, same load, same hour, and judge both at the same cure age in the same room. If you have not yet ruled out the basics (a pool that reaches the edge, cure, a fresh nose), the triage hub does that from the burns you already have.

The reference scent should resemble the outlier as closely as your range allows: a woody or oriental scent beside Sandalwood rather than a citrus, because like-with-like is fairer (our judgement). And both must be judged at the same cure age; unequal ages bias a comparison, which is why the same-cure-age post staggers pours for two oils. Here they share a pour day, so the ages match by design.

Why one scent lags its range

When one wax, one jar, one wick and one load suit most of a range, the scent that lags is usually different in a way the spec never allowed for. The table ranks the reasons we see most often, most common first. The ranking is editorial judgement, not measured data; CSI publishes no failure statistics or throw data.

Rows 2 and 3 are the reason the protocol exists. A house load and a house wick are compromises, and a compromise can miss a scent in either direction: a dense oil may want more heat, and a heavy load can foul a wick rather than lift throw (the pattern the wick autopsy reads). That is why the ladder goes one point down as well as up. Row 4 is the honest one: some oils are simply quieter in a candle, and no spec card makes them loud.

The outlier protocol in three stages

The worked example uses a house spec of Luxury Soy Wax CSI 464, the 150 g Clear Glass Jar with Golden Lid, Eco C1 and 8% of total weight, and Sandalwood as the outlier. Load is a percentage of total candle weight: 7% of a 150 g candle is 10.5 g of oil, 8% is 12 g, 9% is 13.5 g. Sandalwood's page states "up to 10% for soy" and CSI 464's page "up to 10%"; neither says whether that is total or wax weight, and 10% of total is 11.11% of wax, so the ladder stops at 9% (9.89% of wax), inside both readings.

Costs are worked examples and exclude the reference oil from your stock, labour, freight, packaging and GST. If 8% wins the ladder, stage 3 uses 12 g a candle and the protocol takes 96 g of Sandalwood rather than 99 g; the 100 g covers either.

Reading each stage

"Throws" is a score of 2 or 3; "faint" is 0 or 1; "range standard" is whatever your other scents score on your house spec, with no discount for the outlier. Decide only when both rounds agree.

Two cautions. A rung "reaching the standard" means it scores what the rest of your range scores, on both rounds, not that it is the best of a weak set. And the ladder and the wick step are not a 2×2: you change the wick only at the ladder's best load, so the protocol stays at ten candles. If you want the full load × wick interaction for a scent that is strong but struggles, the 2×2 test is built for that.

Spec card or drop: the business decision

If the protocol finds a fix, the scent can stay, on its own spec card. The card copies the house card and changes only what the evidence proved; the diagram above shows the worked example. Before you write it, count what each changed line costs you in money and in handling.

The money is small; the handling is where outlier scents cause trouble. Every scent-specific line is a chance for the wrong wick or the wrong load to go into a production run, and the error shows up in a customer's home, not your test room. Our rule of thumb (judgement, not data): one or two changed ingredients (a load, a wick), written on a card that travels with the pot, are manageable; if the scent also needs its own wax or jar, that is a sign to drop it or move it to a format where it works. The separate pot in the diagram is not a third change; it follows from the load. Policy for loads across a whole range belongs to the one-load-for-every-scent post, and whether one wick can serve many scents to the per-scent wick plan; this page decides one scent.

Dropping a scent is a legitimate outcome, and cheaper than it feels. If nothing reached the standard, the choices are to retire it, to offer it where it performs (a melt, if a melt tests well; see the melt-as-control post), or to replace it with a different oil for the same idea. Commercial success is more than a pleasant smell in the bottle, as the commercial-success criteria lays out.

Where the money shouldn't go

Equally, don't let sunk cost keep a scent. If you already hold a large bottle of an oil that no rung or wick step rescues, the options for the stock (other formats, blends) are in the bought-1-kg plan; they are cheaper than a permanent exception in your pour room.

What the protocol needs

Five items: the worked-example outlier oil, the one wick step the protocol allows, your house wax and jar, and a scale that can see 1.5 g. Chosen on page statements, price per unit and stock on 26 September 2026; CSI has no measured throw data, so nothing is ranked on throw. The oil table shows every size, 150 g candles at 8%, and which stages each bottle covers.

Frequently asked questions