My Fragrance Supplier Reformulated an Existing Oil but Kept the Same Product Name — Should I Repeat Candle Testing Before Using the New Batch Commercially?

Why a same-name oil can still be a different oil

A fragrance oil is a blend of many materials, and whoever makes it can change that blend — to replace an ingredient that is no longer available, to meet a revised standard, to cut cost or to fix a problem — and still sell it under the same name. When that happens, the oil in your next bottle may smell the same on a strip and behave differently in wax, or smell subtly different and behave the same, or both. In general candle-making practice, the composition of an oil is one of the things that can affect throw, colour stability, surface appearance and how the wick burns. So the question in the title has a practical shape: how do you notice a change, and how much testing does a change deserve?

Start with what you can rely on. CSI does not publish reformulation notices, change notifications or lot numbers for any of its fragrance oils, and does not publish lot-to-lot consistency data. We are not claiming that any CSI oil has been reformulated — and we are also not promising that you would hear from us if one were. That is the honest position of most small-pack suppliers, and it means the reference you need has to be one you keep yourself: a retained sample of every lot you approve, and a record of when you bought it.

If your supplier has told you an oil was reformulated, the answer is easy: it is a new oil with an old name, and it goes through the substitute route in the fragrance-supplier page at minimum. The harder and more common case is the one where nobody told you anything and a bottle simply seems different. That is what this page is built for.

The incoming-lot check: retained sample, record, triangle test

The method has three parts and none of them needs a laboratory. The first is a lot record: for every bottle you receive, write the product name, bottle size, order date, invoice number and a note of colour and clarity when you hold it against white paper. The second is a retained sample: from every lot you approve, decant enough oil into a small sealed amber glass bottle — filled close to the top, labelled with the order date — and store it cool and dark. The third is a blind comparison between the new lot and the retained sample of the last approved lot, before any of the new oil goes into production wax.

Why at least 50 g? Because a retained sample has two jobs. It has to supply a smelling strip or two, and it has to supply the oil for three control candles if the strip test is inconclusive. Three 170 g candles at 8% need 40.8 g of oil; add about 5 g for strips (a working allowance) and you need 45.8 g. Decanted from a 1 kg bottle of Bergamot Amber at ₹5.44 a gram, 50 g ties up ₹271.85 of oil per lot. That is the whole price of the insurance.

The triangle test, done properly

A triangle test is the simplest blind comparison there is. Dip three plain paper strips: two from one oil and one from the other, or the other way round, coded underneath so the person smelling cannot tell which is which. Ask them to pick the odd one out. If the two lots are indistinguishable, they will pick correctly only by chance — one time in three. Run three rounds with the arrangement shuffled each time. Getting it right in all three rounds by pure chance happens about 3.7% of the time, so an odd one out found three times is a strong signal of a real difference. Our working rule: odd one found in two or more of three rounds → escalate. It is a rule of thumb, not a statistical standard; tighten it for your most important SKUs.

The retest matrix depends on what the check finds

This is where the title's question gets its real answer. You do not decide in advance whether to repeat all candle testing for a new lot; the incoming check decides for you. If the new lot is indistinguishable from the retained sample on strips, a short CONFIRM in candles is enough. If it is distinguishable, it is a new oil with an old name, and the verdicts move to REPEAT.

The CONFIRM column is deliberately light: it exists to catch a change the strip test missed, not to re-prove a formula that has not moved. The REPEAT column is the same work you would do for any new fragrance — which is exactly what a distinguishable lot is. The fragrance-change checklist lists every line and why it moves; the halfway-down page explains why the full-life burn belongs in the REPEAT column rather than being skipped.

The candle check, step by step

The example candle is Bergamot Amber at your approved 8% in Premium Coconut Soy Wax CSI 468, poured into CSI's 200 ml Salsa jar. The jar's fill weight is not published, and a 200 ml brim volume is not a fill weight; this page assumes you weighed your own fill line at 170 g (an example — weigh yours). At 8% that is 13.6 g of oil and 156.4 g of wax per candle. No fragrance ceiling for CSI 468 is quoted in this series; your approved load is your own validated figure, so it stays as it is.

The scoring sheet for a lot comparison

Because the formula has not changed, the sheet compares only two things — new lot and retained lot — and asks one question per line: same, better or worse? Measure what can be measured. Throw is scored 1–5 relative to the control. The decision uses the worst line, not the average; one bad line is enough to escalate.

CSI publishes no numeric flame-height or jar-temperature limit and this sheet does not invent one. Safety lines pass when they are no worse than the control and within your jar supplier's guidance and the fire-safety standard you work to — ASTM F2417 is a commonly cited one. A new lot that throws more but runs hotter is not a better lot.

What the check costs, and what it saves

The oil for the check is already in your hands — new-lot oil from the new bottle, control oil from the retained sample — so the kit is wax, wicks and jars. Wax covers each candle with a 5% allowance (a working assumption). If you pour tests in production jars you already stock, the cash cost falls to the wax.

Set that against the cost of not checking. A 170 g Bergamot Amber candle in a Salsa jar has a material cost of about ₹221.62 at 1 kg oil and wax prices (oil ₹73.94, wax ₹101.66, wick ₹5.90, jar ₹40.12), before labour, packaging, freight and GST. An example production run of 300 candles from an unchecked lot is ₹66,486.96 of materials — about 60 times the CONFIRM kit — and that is before returns, refunds or a customer who notices their favourite candle has changed.

The decision rule and the release step

Written before you smell a strip: release the lot if the triangle test finds the odd one out in no more than one of three rounds and every candle line in the CONFIRM set is the same as or better than the controls. Escalate if either condition fails. Reject if the escalated set shows any safety line worse than the controls or a scent character your customers would notice. An escalated lot that passes the full REPEAT becomes, in effect, a newly approved oil: it gets its own record, and its retained sample becomes the reference for the lot after it.

If you release a lot that is detectably different but acceptable, record it as a revision of the SKU, not a silent continuation. The formula file should say which lot is in which production batch, so that a complaint months later can be traced. The formula-record page sets out the fields, and the discontinued-material page shows how to document a change so a future employee understands why the candle changed.

Related pages in this cluster

This page handles a change nobody announced. If your supplier openly offered a replacement and called it a direct swap, the direct-replacement vanilla page covers that claim. If you changed supplier on purpose, the fragrance-supplier page sets out the substitute check. If a changed lot turns out to need a different load, the lower-load page and the lavender ladder take over. And if you want a qualified alternative ready before a lot problem or a stock-out forces the question, the backup-supplier page builds it into your purchasing.

Frequently asked questions