Fragrance Oil Safety and Documentation Guide for Indian Candle Brands

The five-document file, and which two you can simply ask for

Most Indian candle brands start the same way: a kilo of wax, three fragrance oils, a hundred jars, and a phone. Documentation feels like something that belongs to factories. Then one of three things happens. A company asks for 300 candles as Diwali hampers and sends a vendor form with a line marked MSDS of all fragrance inputs. A marketplace onboarding page asks for product safety information before it will list you. Or a customer writes to say the candle made their head hurt, and asks what is in it. All three arrive with a deadline, and none of them can be answered by remembering.

The file that answers all three is small. Five documents per fragrance oil, kept in one folder per oil — a physical folder, a phone folder, a shared drive, it genuinely does not matter which, as long as one person knows where it is. Two of the five already exist and are yours for the asking. Two more are five minutes of copying. The fifth is the one you generate yourself, and it is the only one that describes your candle rather than somebody else’s ingredient.

The fourth row is the one experienced makers add after being burned once. Product pages are living documents: sizes change, stock changes, wording is improved. If you formulated a candle at 8% against a page that stated a 6–10% soy range, and you want to know two years later what you were working to, the only reliable record is the one you made at the time. A screenshot with the date typed under it is enough. It takes less time than reading this paragraph, and it is the difference between "we formulated to the published range" and "I think it said ten".

A word on what CSI does and does not publish, because the file should record reality rather than hope. Across the 92 fragrance oil pages, 72 state a candle usage range, 58 state a family or note pyramid, 46 carry an IFRA certification claim, 53 carry a discolouration note, 17 print a vanillin figure and 16 print a numeric flash point. Where a page prints a number, copy it. Where it does not, the honest entry in your file is not published — requested on WhatsApp, followed by whatever the reply says. Never write an estimate into a document that somebody else may rely on.

What the product page already gives you, and what to request

There is a useful division of labour between a product page and a document. The page is written to help you choose and formulate: what it smells like, how much to use, whether it will warm the colour of white wax. The MSDS and the IFRA documentation are written for handling, transport and safety assessment. Makers routinely go looking for one in the other, get frustrated, and conclude that information is being withheld. It is not; it is filed elsewhere.

The last row is the one that decides most corporate conversations. A buyer comparing three candle vendors has the same supplier documents from all three, because a good many Indian brands buy from the same few raw-material houses. What separates the quotes is whether you can state, for your own product, that a 200 g jar with an Eco Candle Wick Thin (C1) at 8% of British Rose reaches a full melt pool in a stated time, burns for a stated number of hours, and was tested on a stated date. Nobody can send you that. It is the one asset in the file you own outright.

What belongs on a candle label — and what quietly does not

A label is a marketing surface and a safety surface at the same time, which is why so many of them read beautifully and say nothing useful. The test to apply to every line: if a customer read this aloud in a complaint, would you be comfortable defending it? Anything that fails is decoration; anything that passes earns its space.

The lines that earn their space

Scent name, and family if the page prints one. "Sandalwood — woody, creamy, warm" is fair, because the Sandalwood page states that family. Inventing a family for an oil whose page does not print one is the first small lie in a chain of them. Net wax weight in grams, not jar volume — a 200 ml jar does not hold 200 g of soy, and buyers who compare candles by price per gram will notice. Burn and care instructions, which is where a printed care sticker pays for itself at ₹0.83 each on the 500 roll or ₹0.70 on the 1000 roll. Your brand name and a contact route, so a complaint reaches you rather than a review page. And the line almost nobody prints: a batch code.

Standard care wording, and why it is standard

The care instructions used across the trade are not decoration; each line prevents a specific failure. Burn until the melt pool reaches the edge on the first burn — this is what prevents tunnelling in soy, and for a 200 g container candle it usually means two to three hours. Trim the wick to about 5 mm before each burn — this is what prevents mushrooming and soot. Do not burn for more than four hours at a stretch — this is what keeps the jar from overheating. Stop when about 10 mm of wax remains. Keep away from draughts, children, pets and anything flammable, and never leave a burning candle unattended. In Indian homes add one more that Western labels omit: do not leave the jar in direct afternoon sun, because a closed car or a west-facing windowsill in May will soften soy long before anybody lights it.

The batch code, which is the whole point

A batch code is any short string that points at one row in your notebook: BR-0826-03 is British Rose, August 2026, third batch. Print it, write it, or stamp it on the base sticker. When a customer says the candle smells weak, the code tells you the wax lot, the load, the wick size, the cure days and whether that batch was poured on the day the workshop hit 38°C. Without it you are guessing about a product you made yourself. With it, you can answer a complaint in one message, replace one batch rather than apologising to everybody, and tell a buyer with a straight face that you can trace any candle you have ever sold.

The pack a corporate or retail buyer actually asks for

Corporate gifting is where Indian candle brands make the jump from hobby income to a business, and it is also where documentation stops being optional. The pattern is always the same: an enquiry in late September or early October for Diwali — 8 November 2026 this year — a sample request, then a vendor form. The brands that win those orders are rarely the ones with the best candle. They are the ones that answer the form the same day.

Assemble the pack once, as one PDF, and keep it current. It takes an afternoon and it is reusable for every enquiry for a year.

Two honest notes on this pack. First, quote lead time you can hold in the week before Diwali, not the one you can hold in July — couriers slow, dispatches queue, and a corporate buyer remembers a missed date longer than a beautiful candle. Second, if you are quoting a fragrance you have not cured and burned in the exact jar you are quoting, say so and send a sample first. A 300-candle order at ₹156.47 of materials each is ₹46,940.40 of wax, oil and wick before a single label, and it is a bad week to discover that a 250 ml jar needs a different wick from your 200 ml.

Claims to keep off your label, your listing and your Instagram

This is the section that costs a supplier money to publish, because every claim below sells candles. They are still the wrong claims. The common thread is that each one asserts something nobody has tested for your product, and each one is easy for a customer, a competitor or a buyer to challenge. The replacement wording in the right-hand column sells almost as well and is defensible in every case.

One nuance worth stating plainly, because it comes up in every IFRA conversation. When a CSI page states that an oil is IFRA certified — 46 of the 92 pages do, in various wordings — that is the supplier stating something about the fragrance material. It is not a statement about your candle, it does not travel to your label as a badge, and it is not a guarantee that the oil suits skin, children or pets. The accurate sentence on your own page or listing is the long one: the fragrance oil used is stated by the supplier to be IFRA certified, and documentation is available. Long sentences are cheap. Arguments are not.

The same discipline applies to the numbers you publish about your own candle. If you print "burns for 40 hours", somebody will time it. Burn hours depend on wick, jar geometry, fragrance load, draughts and how the customer burns it, so the defensible version is a range with a condition attached: approximately 35–40 hours in four-hour burns with the wick trimmed to 5 mm. That sentence is the output of your burn-test record, which is exactly why the record is worth keeping.

The batch record that makes the other four documents usable

Documents describe inputs. A batch record describes what you did with them, on a particular day, in a particular workshop, in a particular month of the Indian year — and that context explains most of the problems a candle brand ever has. Two identical formulations poured in Ahmedabad in May and in Bengaluru in December are not identical candles, and the record is what lets you see that instead of guessing at it.

Keep it to one line per batch in a notebook or a spreadsheet, with these columns. Batch code. Date poured. Wax grade and the pack it came from. Fragrance oil and pack. Load as a percentage of wax weight, and the grams that produced it. Wick type and size. Jar. Number of candles poured. Workshop temperature at the pour, which takes five seconds with a Pen Thermometer at ₹354.00 and explains half of all surface faults. Cure end date. Cold and hot throw notes at the scent gate. Anything unusual.

Then keep one candle from each batch on a shelf, unlit, labelled with the batch code — the retention sample. It is the only way to answer "has this batch changed?" a month later, because the comparison has to be against a candle from the same pour rather than against a memory. On Luxury Soy Wax CSI 464 with British Rose at 8%, one retention candle is ₹156.47 of materials and twelve of them — a year of monthly batches — is ₹1,877.62. That is the entire cost of being able to prove what you made.

The honest limitation: a retention sample from a monsoon pour in Kochi tells you about that pour, not about the oil. Vanillin-carrying oils warm the colour of pale wax over weeks whether they are lit or not — Vanilla Bean publishes 2.5% vanillin and Cinnamon Vanilla the same, both with discolouration expected — so a retention candle of a vanilla will look different from a fresh one by design. Note the expected colour shift in the record when you pour, and the sample stops being evidence of a fault and becomes evidence of normality.

Setting the whole thing up in one week

Nothing here needs software, a consultant or a budget. A five-oil range can be fully documented in about four hours of work spread across a week, and most of that is waiting for replies. This is the order that wastes the least time.

What you will notice, doing this, is how much of it is already half-done. You have the invoices. You know your loads. You have poured test candles and thrown the results away, which is the only genuine loss in the list. The work is not creating information; it is writing down information you already act on, in a form somebody else can read.

And the part CSI has no commercial interest in saying: none of this requires you to buy anything. The two supplier documents are free, the notebook is a notebook, the batch code can be a marker pen on the base of the jar, and a buyer pack is a PDF. If you are choosing between another kilo of wax and a printed sticker roll this month, buy the wax — but send the WhatsApp message for the documents anyway, because it costs nothing and you will need it in October.

Frequently asked questions