How Many Test Candles Should I Make Before Buying Fragrance Oil in Bulk?

Why one candle proves nothing

One candle is where most bulk mistakes start. You pour it, burn it a few days later, it smells good or it smells weak, and you decide. Three things are missing, and each one can flip the verdict.

1. Nothing to compare it with

A candle at 8% on a C1 wick throws a certain amount. Would 10% have been clearly better, or no different and more expensive? Would a C2 have built the full pool this oil needed? A single candle answers none of that. Without at least two loads and two wicks you cannot tell whether you have found the right setting or just a setting.

2. No way to tell the oil from the pour

Two candles poured from the same jug can differ: a wick a few millimetres off-centre, a jar that cooled faster by an open window, an air pocket near the wick. A single candle bakes any of those into the result. A repeat poured separately is the simplest way to show that the result belongs to the oil and the setting, not to that one pour.

3. One age, one evening

Soy commonly throws more at day 14 than at day 3, and a nose that has spent the afternoon pouring is not the nose that will smell your customer's candle. One candle judged once is one reading of a moving target. Judging each candle at day 3, 7 and 14 costs no extra candles, only diary space.

The minimum matrix: loads × wicks × repeats

A test matrix is just a grid of candles in which each dimension answers one question. For a bulk decision on one oil in one wax, three questions matter, so there are three dimensions.

Three loads × two wicks gives six cells, one candle each. Once day 14 picks a winner, two repeats of that single cell make eight. Repeating every cell instead would make twelve, and most of those extra candles would repeat settings you have already rejected. That is why eight is our minimum and twelve is the ceiling: the extra four only earn their place when a whole launch rides on one oil and you want every cell confirmed before you choose.

Loads inside both ceilings

The top of the ladder is the lower of the wax's stated maximum and the oil page's stated maximum. Luxury Soy Wax CSI 464 states up to 10%; Premium Coconut Soy Wax CSI 468 up to 13%. Lavender and Apple Cinnamon state 6% to 10%, so 6/8/10 works in either wax. Fresh Strawberries states "6-9% for candles", so its ladder is 6/8/9 whatever the wax allows. Ceilings are formulation limits, not regulatory limits, and the top step is there to be tested, not assumed best.

Running the eight, in order

The decision rule for the repeats

Score every candle blind at the same check-in times on a 0 to 5 scale. If both repeats land within 1 point of the original cell's day-14 score, the result stands. If either is more than a point away, the cell is not settled: pour a third from yet another melt before you buy. The 1-point band is our judgement; the method for running the burns themselves is in How to Test Fragrance Oil Hot Throw Before Buying It in Bulk, and a head-to-head against a rival oil is in How to Compare Two Fragrance Oils for Actual Candle Performance.

Grams, bottles and what is left over

In a 150 g candle with the load as a percentage of total weight, 6% is 9 g of oil and 141 g of wax, 8% is 12 g and 138 g, 10% is 15 g and 135 g. So one row of the ladder takes 36 g of oil, two rows take 72 g, and the two repeats take 18 g, 24 g or 30 g depending on which load won.

The 10% row is the honest wrinkle. After the six-cell grid a 100 g bottle has 28 g left; one repeat at 10% takes 15 g, leaving 13 g, 2 g short of the second. Three ways round it: buy a 50 g alongside the 100 g at the start (for Lavender and Apple Cinnamon that costs nothing extra per gram, because their 50 g and 100 g are priced the same per gram); pour the second repeat at 10% from a 15 g bottle; or, if 10% only narrowly beat 8%, repeat 8% instead, since it is cheaper to produce and the repeats will show whether you lost anything. For Fresh Strawberries, whose page caps at 9%, the worst case is 96 g and the 100 g always covers it.

What one, three, eight and twelve candles cost

Costed for Lavender, the worked example: oil at the 50 g or 100 g price (both ₹5.40 per gram), wax valued as CSI 464 at ₹0.51 per gram (10 kg price), 150 g clear jars at ₹94.40 and Eco wicks at ₹5.90 (C1) and ₹9.44 (C2). Repeats are costed at 8% on C2. "New spend" is oil and wick packs only, assuming you own wax and jars.

Eight candles consume ₹1,898.67 of materials, ₹237.33 per candle, and the jars are the largest line; washed and reused, the real cost falls sharply. Going from eight to twelve adds ₹944.02 of materials. For the other two oils, the eight-candle new spend is ₹627.76 (Apple Cinnamon) and ₹531.00 (Fresh Strawberries). Against that, Lavender's 1 kg is ₹5,221.00, 7.5 times the eight-candle new spend. All figures exclude labour, freight, packaging and GST.

When more candles will not help

Count only matters once the system can throw. If every cell of your six-candle grid is weak at day 14, including the C2 row at 10%, twelve candles will be twelve weak candles. That pattern points away from the oil: a wick family too small for your jar, add temperature, cure, or a test room that is too large. Burn your reference candle beside the grid; if it is weak too, stop counting and start diagnosing, beginning with My Candle Smells Weak When Burning — Do I Need a Stronger Fragrance Oil?.

The opposite trap is a matrix that answers the wrong question. If your range uses two waxes, testing one oil in eight candles of one wax says nothing about the other; multiplying across waxes is a different count with its own pruning rules, in How Many Wax × Fragrance Combinations Should I Test Before Launching?. And if the question is "this oil or that one", eight candles of each is overkill once both have passed their screens; a four-candle pairing between the two finalists does that job (the paired comparison).

Testing several oils: screen three, promote to eight

If you are choosing two oils for bulk from a shortlist of five, eight candles each is 40 candles. A two-stage count does the same job with fewer. Stage 1: screen every oil with three candles on your usual wick (6/8/10), from a 50 g bottle each; judge at day 14. Stage 2: promote the best two, and for each pour the missing C2 row (three candles) plus two repeats of the winning cell. That is 15 screen candles plus 5 for each of two finalists: 25 candles instead of 40, 15 fewer, and ₹1,416.00 of jars you did not fill at the pack-of-10 price.

A finalist's stage-2 grams are 36 g for the C2 row plus 18 g to 30 g for the repeats, against the 64 g available from the 14 g left over plus a new 50 g. That covers every case except repeats at 10%, where you are 2 g short again; the same three workarounds apply. As a worked example, five 50 g screens of Lavender, Apple Cinnamon, Fresh Strawberries, British Rose and Jasmine cost ₹1,161.98 before wicks. Ask on WhatsApp to confirm stock of a longer list in one message.

What to buy for the eight

Three oils whose page ranges and pack prices illustrate the count (a 6–10% oil with equal 50 g and 100 g per-gram prices, a 6–10% oil flat per gram from 50 g, and a 6–9% oil that moves the ladder), plus the wicks and jars the matrix consumes. Chosen on stated loads, notes, price per gram and stock on 25 September 2026; CSI has not measured the throw of any of them. Candle counts assume 150 g at 8%.

Planning a summer 2027 range

Fresh and fruity oils such as Lavender and Fresh Strawberries are natural candidates for a range sold through the Indian summer, roughly April to June. Working back from an April launch: the eight-candle test takes about five weeks from the first pour to the repeats' day-14 reading (six candles on day 0, the day-14 scoring round ending about day 19, repeats poured then and judged from about day 33). Start in late January 2027: the winning cell is known on 13 February, just before the Valentine's Day rush of 14 February if that is your busy week, and the repeats decide by 1 March. Test in the season's conditions if you can; a room with fans running, as it will be in May, is a harder test of hot throw than a closed winter room.

Frequently asked questions