How to Choose a Candle Fragrance Oil

The five gates, in one table

Most people choose a fragrance oil the wrong way round: they fall for a name, buy a bottle, pour it, and then discover the load they needed, the discolouration they did not want, or the fact that the page was describing a wax melt. The method below reverses that. It spends its first four gates on information that is free — the product page, your own arithmetic — and only reaches the expensive gate, a physical sample, once the catalogue has shrunk from 92 oils to three or four candidates. Nothing in it is specific to CSI; it works on any supplier’s catalogue, and it works better on a thin page than a thick one, because it tells you exactly what to go and ask for.

The order matters more than any individual gate. Sampling first is what beginners do and it is the single most expensive habit in Indian candle making, because a sample only answers "do I like this", which you could have guessed, while gates one to four answer "can I actually sell this", which you could not. The funnel below shows what the specification gate alone does to a 92-product catalogue.

Gate 1: decide the format before you look at a single scent

A fragrance oil page usually states loads for several products, and they are not interchangeable. Lavender Fragrance Oil states 6–10% of total wax weight for candles, 10–15% for wax melts, 10–30% of the diffuser base for reed diffusers and 8–15% of the spray base for room sprays. Citrus Lemon and Apple Cinnamon state the same four bands. Blue Ocean and Sandalwood state up to 12% for wax melts against up to 10% in soy candles.

The reason is mechanical, not commercial. A candle has to burn: the oil has to dissolve into the wax, travel up a wick and vaporise in a flame without flooding it or pooling on the surface. A wax melt never burns — it only warms — so it can carry more oil without a wick to drown. A reed diffuser has no heat at all and relies on capillary action and evaporation, so it needs several times the concentration to register in a room. Copy a diffuser percentage into a candle and you get a wick that drowns, a surface that weeps and a candle that smells weaker, not stronger.

72 of the 92 CSI oil pages state a candle range, and most of those state ranges for the other formats as well. British Rose prints all of them on one page: soy candles up to 10%, paraffin up to 6%, wax melts up to 12%, reed diffusers 10–30% of the base and room sprays 8–15% of the base. If you make more than one product, note that an oil which is excellent in one can be ordinary in another — one CSI reviewer reports exactly this, liking a fragrance but getting low throw from a reed diffuser formula built at 10% oil. That is a formulation question for the diffuser base, not a verdict on the oil.

Gate 2: choose the family from the room, the season and the buyer

Fifty-eight of the 92 CSI pages state a family, a note pyramid or both. Use them as a map, not as a recommendation — this guide deliberately does not tell you which family is best, because that depends entirely on who is buying. What it can give you is the three questions that turn a preference into a decision.

Which room, and how big is it?

A 10x12 ft Indian bedroom with the door shut is a very different acoustic for a scent than an open-plan living-and-dining space in a 2BHK with a fan running. Small closed rooms forgive a light fragrance and punish a heavy one; large open rooms do the reverse. If your product photographs show a bedside table, you are selling into the first case. The live CSI guides to small bedrooms and large living rooms go through the trade-off in detail.

Which months will it actually sell in?

India has a fragrance calendar and it is not the northern-hemisphere one. Fresh, citrus and aquatic families carry April to June, when a Delhi or Ahmedabad room sits above 35°C and anything heavy reads as cloying. Earthy and green families own the monsoon. Gourmands, spice and woods take over from Rakhi through Diwali and into the December wedding season. Buying a spiced gourmand in March is not wrong, but it is capital sitting still for six months.

Is the buyer keeping it or giving it?

A candle someone buys for themselves can be polarising — that is the point of it. A candle bought as a Diwali hamper, a wedding favour or a corporate gift cannot be, because the giver is guessing on the recipient’s behalf and will default to something safe. Gifting pushes you towards the broadly liked middle of a family and away from its edges, and it also pushes you towards oils that will not discolour a white candle before it is opened.

Gate 3: the four lines the page has to print

Run your two or three candidate families through this gate and most of them will fall out. The counts below are for CSI’s catalogue as pulled on 12 September 2026; the shape is similar across the industry, and thin pages are normal rather than suspicious. What matters is knowing which missing line you can live with.

The candle usage range — non-negotiable

72 of 92 pages print one. This is the supplier telling you how much of the bottle the formulation is built to carry in wax, expressed as a percentage of wax weight — so 8% of a 200 g pour is 16 g of oil, going into 200 g of wax, for 216 g in the jar. Two things follow. First, your wax has its own stated maximum, and when the two disagree the lower one governs: Luxury Soy Wax CSI 464 states up to 10%, Luxury Soy Wax Chunks and Premium Coconut Soy CSI 468 state up to 13%, and Eco Soy CSI 400 does not publish a figure at all. Second, if the oil page prints nothing, do not estimate. Ask on WhatsApp for the candle range before you buy — it is a one-line question and it takes a minute.

The family or note pyramid — near-essential if you sell online

58 of 92 give one. You need it less for pouring than for selling: the note pyramid is the raw material of your product description, your label copy and your Instagram caption. Without it you are describing a smell you have only met once, which tends to produce the same four adjectives on every SKU. Where a page gives no notes, a 15 g sample and your own nose will fill the gap — but that is a cost, and it belongs in gate five rather than being waved away.

The IFRA claim — ask for the document, do not infer it

46 of 92 state IFRA certification. The right way to read that line is exactly as written: the page states it. It is not a safety guarantee you can pass on to a customer, and it is not a claim that the oil is safe on skin or around children. If you need it for a corporate buyer, a marketplace listing or an export enquiry, request the actual documentation — MSDS and IFRA paperwork is available from CSI on request on WhatsApp +91 7397976926, for the specific oil you are buying. CSI has a live page on buying IFRA certified oils in India.

The flash point — a handling number, not a quality score

Only 16 of 92 print a numeric flash point, and it is the line most often misread. It is the temperature at which the liquid gives off enough vapour to ignite near a flame: a transport, storage and handling property used to classify shipments. It does not rank hot throw and it is not the temperature at which fragrance "burns off" in a candle. It changes exactly one thing in your process — the add temperature. Most stated figures sit far above the usual 80–90°C window: Vetiver at 206°C, Sandalwood at 203°C, British Rose at 192°C. A few sit close to it or inside it: Palo Santo at 93°C, Gardenia Tuberose at 98°C and Neroli at 85°C. With those, add at the bottom of the window and stir straight away.

Gate 4: cost the oil in one candle, not in one bottle

Two oils sitting next to each other at ₹212.40 and ₹170.00 for 50 gm look almost identical. At 1 kg the same two are ₹3.07/g and ₹2.83/g, and across a hundred 200 g candles at 8% that gap is ₹377.60. Neither number is visible on the shelf. You get it by doing one multiplication: per-gram rate × wax weight × load. That is the whole of gate four, and it is the gate that decides whether a scent you like can survive your retail price.

Start with grams. The convention on CSI’s product pages is a percentage of wax weight, so a 200 g pour at 8% takes 16 g of oil into 200 g of wax — 216 g in the jar, not 200 g. Get that wrong and every cost below is understated by about 8%. The table uses Eucalyptus Spearmint at its 1 kg rate of ₹3.07/g as a worked example; substitute your own oil’s rate and the shape of the answer does not change.

Now widen it. The per-gram rate at 1 kg across CSI’s in-stock oils runs from about ₹2.83/g to ₹24.06/g — a factor of more than eight. That spread does not track hot throw, and nothing on the store publishes a comparison that would let anyone claim it does. It tracks the raw materials in the blend: a citrus and a mass-market floral are cheap to build, a perfume-inspired composition carrying oud, tonka and resins is not.

Read the last column against your selling price rather than against each other. A 200 g candle retailing at ₹899.00 carries the top of that table comfortably; the same candle at ₹499.00 does not. And the arithmetic scales brutally: at ₹3.07/g, 100 candles of 200 g at 8% is 1.6 kg of oil and ₹4,908.80 of fragrance. At the other end of the table the same run is ₹38,489.60. That is the difference between a fragrance line you can ignore and one you have to plan around.

Gate 5: the 15 g sample that ends the argument

By now the catalogue should be down to three or four oils that suit your format, sit in a family your buyer wants, publish the lines you need and cost what you can afford. The sample decides between them — and it only decides cleanly if the sample is controlled. Three testers poured in three different waxes, at three different loads, in three different jars, tell you nothing about the oils. They tell you about the differences you introduced. One variable changes: the oil.

A 15 gm bottle is sized for exactly this. At 7% it fills one 200 g candle with a gram spare for a blotter; at 8% it fills one 180 g candle. Three of them is ₹205.32, five is ₹342.20. A complete tester — 200 g of Luxury Soy Wax CSI 464 at the 1 kg rate, one Eco Candle Wick Thin (C1) at ₹5.90, and 14 g of oil at the example rate — comes to about ₹150.33 in materials. Three of those is under the price of one 500 gm bottle bought on a hunch.

One more control worth the fifteen minutes: pour a fourth jar with no fragrance at all. It costs about ₹101.48 of wax plus a ₹5.90 wick and it settles the argument you will otherwise have with yourself about whether the wick is wrong or the oil is weak. If the unscented control burns cleanly and the scented ones tunnel, the load is fighting the wick; if all four behave the same way, the wick is simply the wrong size for the jar. CSI’s live wick guide by diameter is the place to fix the second case.

The decision sheet: one page, three candidates, one answer

Write the five gates down the left of a page and your candidates across the top. Fill in every cell before you smell anything. The worked example below takes a real brief — a fresh, clean scent for a 200 g soy jar candle retailing around ₹899.00, poured in CSI 464 at 8% — and runs three in-stock CSI oils through it. Notice that two of the three have a blank cell, and that a blank is not a disqualification. It is a question for WhatsApp and a note in your file.

Three things fall out of a sheet like that. First, all three clear the hard gate, so this is a real choice rather than a rescue. Second, the fragrance line at 8% differs by ₹30.21 per candle between the cheapest and the dearest, which on 300 candles a month is ₹9,062.40 — worth knowing, not worth deciding on alone. Third, two columns print a family and two print a flash point, but only one prints both — and if you sell online the family is the line you will miss most, because it is the one that writes your product description.

The ceiling rule sits over the whole sheet. Your wax has a stated maximum and the oil has a stated range, and the lower of the two governs every time. CSI 464 states up to 10%, so an oil rated to 13% is still a 10% oil in that wax. Luxury Soy Wax Chunks and Premium Coconut Soy CSI 468 state up to 13%, so with an oil rated 6–10% the oil governs instead. Eco Soy CSI 400 does not publish a figure — with that wax, the oil page is your only ceiling, and it is worth a WhatsApp message before a large batch. Japanese Cherry Blossom is the widest range on the list at 6–13%, and even that becomes 10% the moment you pour it into CSI 464.

Frequently asked questions