My New Lavender Smells Stronger from the Bottle but Weaker in Wax — Should I Optimise It as a Completely New Fragrance Rather than Treating It as a Substitute?

Why the bottle and the candle disagree

When you open a bottle and sniff, you are smelling whatever leaves the liquid fastest at room temperature. In general perfumery terms that is the most volatile part of the oil — the bright, sharp top of a lavender. A candle asks something quite different of the same oil. For cold throw, the fragrance has to leave solid wax, where it is held in the wax structure. For hot throw, it has to leave a pool of melted wax that is much warmer than the room, over hours, and keep smelling like lavender the whole time rather than just at the start. An oil can be strong at the first job and ordinary at the other two. That is not a defect in the oil; it is a different oil being asked a question it was never compared on.

It is commonly described that a fragrance with a very lively top can seem powerful on a smelling strip and then fade during cure, because that same lively fraction is what the wax lets go of first. It is also commonly described that the way an oil sits in a particular wax — how well it is held and how readily it is released from the melt pool — varies from oil to oil. CSI publishes no measured throw, retention or composition data for any oil, including this one, so none of that can be predicted from a product page. It has to be measured in your candle, against your old candle.

The practical point is sharper than the chemistry. The only reason you could treat a replacement as a substitute — keep the same 8%, keep the same wick, run a short side-by-side — is the working assumption that it behaves like the old oil. Your own observation has now contradicted that assumption. The page on a lavender that smells almost identical covers the case where the assumption still holds. Once the bottle says stronger and the cured candle says weaker, you are outside that case.

What you give up by switching routes is time, not money. A substitute check is one rung and a handful of candles. A new-fragrance optimisation is a ladder, cure checkpoints and a second, confirming round. On a 200 g candle in Luxury Soy Wax Chunks the whole thing is 14 test candles including controls, and the new oil for it costs ₹1,185.02. What you gain is a load you chose on evidence, rather than an 8% you inherited from an oil that no longer exists in your formula.

The retest matrix for a lavender that failed the substitute test

Every page in this series uses the same four verdicts. KEEP means the old result is still valid. CONFIRM means a short side-by-side check against the control. REPEAT means the full test again. NEW means a test that your old formula never needed. For a replacement oil that behaves like the old one, most lines would be CONFIRM. For this lavender the verdicts move to the right, because the evidence says the new oil is different.

Two lines deserve a word. The wick moves to REPEAT only after the load is chosen, because testing the wick on a rung you later reject wastes a set of candles. And documentation is NEW, not optional: CSI can send MSDS and IFRA documents for an oil on WhatsApp request, and you should hold them for the oil you actually use. The general rule for which lines move when a fragrance changes is set out in the fragrance-change checklist; this page is the version for an oil that has already shown it is not a like-for-like swap.

The two-stage optimisation, candle by candle

The protocol below is a working practice, not a standard; change the numbers to suit your range. It runs in two stages so that most of the oil goes into the rungs that deserve it. Stage 1 is a scout: one candle at each of four loads — 6%, 8%, 10%, 12% — plus one control made with the old lavender at 8% (or taken from a retained batch if the old oil is gone). Stage 2 takes the two rungs that scored best and pours three candles of each, against three more controls, so every decision rests on three candles rather than one. All candles are 200 g in the same jar, wax, wick and fill line as production.

The scoring sheet against the control

Score every variant relative to the control candle, not on an absolute scale. A 1–5 throw score means little from one week to the next; "one point below the control, same room, same checkpoint" means a great deal. Measure what can be measured — melt-pool diameter and depth, weight before and after each burn, time to full melt pool — and score only what cannot, such as throw. Consumption in grams per hour is simply weight lost divided by burn hours.

CSI publishes no numeric flame-height or jar-temperature limit, and this page does not invent one. The safety lines pass when they are no worse than your approved control on every burn and stay within your jar supplier's guidance and the candle fire-safety standard you work to — ASTM F2417 is a commonly cited one. A stronger throw never buys a hotter jar or a taller flame. If a higher rung throws better but runs hotter, it fails.

The priced test kit

The kit covers both stages. The new oil is Stage 1's 72 g plus Stage 2's 132 g — priced at the heaviest likely pair, 10% and 12% — which is 204 g, plus a 5% allowance for weighing and pour loss (a working assumption), so 214.2 g. The wax covers all 14 candles including the controls. The controls' oil is your old lavender and is not priced. Wick and jar are CSI's defaults in these guides; if your approved wick or jar is different, use yours — the test must match production.

If you already hold jars and wicks from production stock, the cash outlay drops to the oil and wax: ₹2,982.02. The oil line is worth a second look. The cheapest way to cover 214.2 g from the listed sizes is 2 × 100 g + 1 × 15 g; three 100 g bottles would cost ₹1,620.00. Lavender is priced at ₹5.40 a gram at both 50 g and 100 g, so for this oil there is no penalty for buying the smaller bottles to fit the test. The 15 g size carries the usual premium — ₹7.00 a gram — but as a top-up for the last few grams it is still ₹164.98 cheaper than a third 50 g bottle.

The decision rule, and what each load costs

Write the rule down before Stage 1 is poured, so the result cannot talk you into it afterwards. Ours is simple: choose the lowest rung whose cured cold throw and hot throw reach the control, with every safety line no worse than the control through the full-life burn and no sweating in storage. If two rungs tie, take the lower one. If none of the four rungs reaches the control by the 14-day checkpoint, the oil fails — it is not a load problem you can push through, and the ceiling is 13% anyway.

Every two-point step up the ladder adds 4 g of oil and removes 4 g of wax. At the 1 kg rate that is ₹20.88 more oil and ₹2.40 less wax, so each rung costs about ₹18.49 more per candle. That is small next to a weak candle, but it is not nothing across a production run, and it is the reason the rule says lowest passing rung. The full comparison of oils at different loads and prices is in the ₹5,000 vs ₹3,500 per kg page; the case where the replacement needs less oil than the original is in the 6% vs 9% page.

Releasing the new lavender and keeping the record

Release is a written decision, not a feeling that the candles are fine. Once Stage 2 passes and the full-life candle has burned to the end of usable wax without failing a safety line, record the approved formula as a new fragrance record rather than an edit of the old one. The old lavender's record stays on file, marked as withdrawn and why. If a customer complaint arrives next year, you will want to know which lavender, at which load, was in that candle.

Keep a retained sample of the new oil — 15 to 20 g in a sealed amber bottle, labelled with the order date — and one unburned candle from the first production batch. CSI does not publish lot numbers, lot-to-lot consistency data or change notices for any oil, so your retained sample and your order record are the reference if a later bottle ever smells different. The reformulation page shows how to use that sample when a later batch seems to have changed.

Where the sibling pages pick up

This page owns one situation: a replacement that has already shown, in wax, that it is not a like-for-like swap. The neighbouring pages own the others. If the new oil is stronger in the candle rather than weaker, the sandalwood page works the ladder downwards. If the new oil is doing well at a lower load than the old one, the 6% vs 9% page covers redesigning around it. If a higher rung has started to sweat, the sweating page comes first. If the candle behaves at the top but struggles halfway, the halfway-down page explains why the full-life burn cannot be skipped. And for the whole list of what to retest on any fragrance change, the checklist is the hub.

Frequently asked questions