How Many Wax × Fragrance Combinations Should I Test Before Launching?

The formula, and why the full matrix is the wrong answer

The number of test candles for a launch is a product of five factors:

N = W × O × L × K × R, where W is candidate waxes, O is oils (scents), L is loads per oil, K is wick sizes and R is repeats of each cell.

A cautious maker with six scents who wants to "test properly" writes W = 2 (CSI 464 and CSI 468), O = 6, L = 3 (low, middle, high), K = 2 (Eco C1 and C2) and R = 2. That is 2 × 6 × 3 × 2 × 2 = 144 candles, 1,728 g of oil and nearly 20 kg of wax. At CSI prices on 25 September 2026, the materials consumed come to ₹35,442.61: ₹9,118.90 of oil valued at each oil's 100 g price, ₹5,041.53 and ₹6,329.23 of wax at the 10 kg prices, ₹13,593.60 of jars and ₹1,359.36 of wicks. It would also take one very long fortnight to burn and score 144 candles fairly: at one candle per evening per room, it cannot be done in 14 days.

The full matrix is wasteful for three reasons. It tests every oil in a wax you may never use. It tests every load at a wick you already know is wrong. And it repeats cells that were never contenders. None of those candles changes the launch decision. Pruning is not cutting corners; it is refusing to pay for answers you do not need. How many candles to pour for a single oil before a bulk order is a narrower question, covered in How Many Test Candles Should I Make Before Buying Fragrance Oil in Bulk?.

Six pruning rules, and the count after each

The pass and fail counts (five pass, one fails) are assumptions for budgeting, not predictions about any particular oil. If two oils fail, rule 6 adds 3 candles each; if none fails, you stop at 29.

What pruning costs you

Honesty requires the other side. The pruned plan does not tell you how every oil would have done in the second wax, or whether a load you skipped would have beaten the two you tried. For a launch that is a fair trade: you need a set of scents that work in one system, not a map of every system. It also takes longer, because each round waits for the one before; the full matrix is one round of 144 candles, the pruned plan three short rounds of 12 to 12 to 8. If your launch date is fixed and close, see the calendar note below before choosing.

Which loads each oil gets: the two-ceiling rule

"Three loads" does not mean 6%, 8% and 10% for every oil. The lower of the wax page's maximum and the oil page's stated range governs, and the ladder should start near the bottom of the oil's stated window. Here is the ladder for the six scents in CSI 464 (stated "Up to 10%"). Ceilings are formulation limits, not regulatory limits, and the top of a window is rarely the best-performing load.

In a 150 g candle, 6% is 9 g of oil, 7% is 10.5 g, 8% is 12 g, 9% is 13.5 g and 10% is 15 g, all on total weight. Weigh on a scale; at these differences, a spoon cannot tell 8% from 9%. Every round-1 figure fits inside a 50 g bottle, but repeats and rule 6 do not, which is why the basket buys 100 g of each.

Rule 4: is it actually the oil? The kill rule

Rule 4 drops an oil that is weak at every load at its winning wick. It only applies when most of the other oils pass in the same round, same wax, same jar, same room. That condition is the whole point. If one oil of six is weak and five pass, the evidence says the oil does not suit this system, and rule 6 gives it one chance in the other wax. If three or more of six are weak, the evidence says something they share is wrong: a wick range that is too small for the jar, candles judged too young, a room too large or too draughty, or a process fault. In that case the right number of extra combinations is zero.

A kill is also not a verdict on the oil everywhere. It says "not in this wax, this jar, this wick range, at these loads". An oil that fails in CSI 464 may do well in a different system; rule 6 tests exactly one such alternative, and the matching method for oils and waxes is in How to Match Fragrance Oil With Soy Wax for Better Hot Throw.

The rounds, day by day

The pruned plan runs in three rounds. Candles within a round are poured together; each round is scored at its own day 14, because comparing candles of different ages is not a fair test. The day 3, 7 and 14 schedule is common practice, not a rule.

Your number, for four to ten scents

With the same rules (one wax first, two wicks, three loads per oil, repeats for finalists, a second-wax pair plus repeat for each failure), the pruned count is 5 × O + 2 × F, where O is the number of scents and F is the number that fail round 1. The full matrix is 24 × O. The budget column multiplies the pruned count by the worked example's ₹244.67 of materials per test candle; it is a rough guide, since your oils' prices will differ.

Past about ten scents, the question shifts from "how many combinations" to "how do we standardise testing", and one house wax with a wick map becomes essential; that programme is in My Candle Business Needs Consistent Hot Throw Across 10–20 Fragrances — How Do I Standardise Fragrance Oil, Wax and Wick Testing Before Buying in Bulk?. For a first collection, fewer scents tested properly beats more scents tested loosely; the collection-level view of how many of each family to test is in our guide to a 10-fragrance collection.

The basket, checked candle by candle

The oil use is highest for the oil that goes to rule 6, which the worked example prices as Coffee Vanilla Tobacco, the dearest per gram of the six: 88.5 g of its 100 g. Any of the six could be the one that fails, and each 100 g bottle covers its own round 1, a rule 6 pair and a repeat, whichever it is. If you reuse jars from earlier tests, the basket without jars is ₹6,328.20. Figures exclude labour, freight, packaging and GST.

What to buy for a six-scent launch

The house wax, the wick bracket, and four of the six launch scents, chosen because their stated windows differ (6–9%, 6–13%, 7–9% and 7–11%) and so show how the two-ceiling rule changes the ladder; Lavender (100 g ₹540.00) and Citrus Lemon (100 g ₹300.00) complete the range. Oils are chosen on stated load window, notes structure, price per gram and stock at every size on 25 September 2026, not on throw, which CSI has not measured. Oil candle counts assume 12 g per candle (8% of 150 g).

Calendar: which launch dates this plan can make

The pruned plan takes 34 days to a launch decision, then a production pour and its own cure (we use 14 days). From 25 September 2026: Diwali on 8 November 2026 is 44 days away, too close for a new range tested this way; Christmas on 25 December 2026 works if round 1 is poured by mid-October; Valentine's Day on 14 February 2027 works comfortably if you start by the end of November. Dispatch time depends on your location; confirm on WhatsApp before you promise a date.

Frequently asked questions