I'm Wasting Money Testing Random Wax, Wick and Fragrance Combinations — What Order Should I Test Raw Materials In?

Why random testing costs more: the arithmetic

Say you are choosing between two waxes and two wick sizes for six oils, and you want to try each oil at 6%, 8% and 10%. Every combination at once is 2 × 2 × 6 × 3 = 72 candles and, at an average 12 g of oil each, 864 g of oil, before a single repeat. As a worked example, valuing the oil at ₹4.60 a gram (Jasmine's 50 g per-gram price) with CSI 464, a clear 150 g jar and a C1 wick, that grid is worth ₹16,237.53 in materials. Nobody pours that, so random testers pour a handful of combinations instead, and learn nothing from them: when a lavender in wax A on a C1 smells weak, is it the lavender, the wax, the wick, the load, or the age of the candle?

The sequence below answers each question once, in the order that later questions depend on it. For the same six oils it uses 39 candles, including repeats and a stability check, and every candle has a job. The full isolation grid for diagnosing one weak candle is a different tool, set out in How to Test Whether Weak Hot Throw Comes From the Oil, Wax, Wick or Process; the system idea of choosing wax, then wick, then oils is introduced in How to Choose Candle Raw Materials That Work Together Instead of Testing Randomly. This page is the complete programme with gates, budgets and dates.

Is it actually the fragrance oil you have been testing?

In a random test log, the oil is usually the variable the maker thinks is being tested, and often the one that is not. An oil tested on the wrong wick is a wick test. An oil tested at 10% in one wax and 8% in another is a load test. An oil burned on day 2 is a cure test. That is why oils sit fourth in the sequence: by stage 4 the wax, wick and jar are fixed, so an oil that fails really has failed.

The eight stages, each with a gate

Stage 1: format, on paper

Choose the jar and the fill weight before you buy wax, because both decide how much wick you need. Measure the inside diameter yourself; product pages often give outer dimensions (the Clear Glass Jar with Golden Lid (150 gram) page states approx. 7 cm outer and "Safe for hot-pour up to 85°C"). If you plan two formats, stage 3 runs once per format; everything else runs in your main format.

Stage 2: wax, with the wick bracketed inside it

Pour the control oil at 8% in two candidate waxes, each on a C1 and a C2: four candles. Bracketing the wick inside the wax stage is deliberate; if you pour both waxes on one wick, you may choose the wax that happens to suit that wick. Judge surface, adhesion, frosting and whether either wick reaches the glass. For a container range we would compare Luxury Soy Wax CSI 464 (pour 80°C–90°C, up to 10%) with Premium Coconut Soy Wax CSI 468 (pour 58°C–60°C, up to 13%). Their stated pour temperatures are at least 20 degrees apart, which is the practical reason wax comes first: every later stage inherits the wax's temperatures and ceiling. Why the same oil behaves differently across CSI waxes is covered in Why Does the Same Fragrance Oil Perform Differently in Different Soy Waxes?.

Stage 3: wick family, in the chosen wax

Now bracket properly: C1 and C2 in the winning wax with the control oil at 8%, each poured from two separate melts, three burns each of up to four hours, pool width logged hourly. If neither size reaches the glass cleanly, try another family (for example wooden wicks) here, not later. The gate is strict because every oil after this is judged on this wick.

Stage 4: the oil screen, one candle each

Only now buy candidate oils, 50 g each. Pour each at 8% on the proven wick beside one control candle, judge character at day 3 and throw blind at day 7 and 14, and drop the ones that fail. One candle per oil is enough to screen, because everything else is fixed; it is not enough to approve. Survivors go to stage 5. A 50 g sample covers stage 4 (12 g) and stage 5 (36 g) with 2 g left; buy a second 50 g for survivors before stages 6 and 8.

Stage 5: load, per survivor

Pour each survivor at 6%, 8% and 10%, within the lower of the wax's and the oil page's stated maximum. Keep the lowest load whose day-14 score matches the next load up. That is a per-oil decision; loads for a whole collection are discussed in Should Every Fragrance in My Candle Collection Use the Same Fragrance Load?, and the descending method in How to Find the Minimum Fragrance Load That Still Gives Strong Hot Throw.

Stage 6: wick fine-tune, only where needed

Any survivor whose pool missed the glass at its chosen load gets a C1/C2 recheck at that load. Most will not need it; those that do become the second wick class on your wick map. The deep version of this step, with a control candle and a load candle, is in the post on one scent tunnelling where another does not.

Stages 7 and 8: cure, then stability

Cure is not a pour; it is the rule that every judgement in stages 4 to 6 happens at the same candle age, at day 3, 7 and 14. Several CSI oil pages state a 48–72 hour soy minimum before burn testing; day 14 is common practice, not a rule. Stability comes last because it tests the finished recipe: two candles per SKU, one on a warm shelf (Indian summer rooms get hot) and one in light, watched for 30 days for sweating, pooling and colour. The longer protocols are in How to Test Fragrance Oil Compatibility Before Buying 1 Kg or 5 Kg and How to Test Fragrance Oil for Discoloration Before Bulk Production.

When you can skip a stage

A sequence is a tool, not a ritual. You can skip a stage when you already hold its answer as evidence, not as a feeling.

Budgets at three scales

Three worked examples. Starter: one format, three candidate oils (Lavender, Jasmine, Sandalwood), assuming two survive. Small range: one format, six candidates (adding Citrus Lemon, Apple Cinnamon, Blue Ocean), four survive. Launch: two formats, ten candidates (adding Woody Oud, Roasted Coffee, Gardenia, Fresh Strawberries), seven survive. Which oils survive in your candles is unknowable in advance; the counts are for budgeting.

Three honest notes. The jar line is the biggest item and the most compressible: test jars can be cleaned and reused between stages, and a used jar is fine for a stage-2 surface check. The launch control oil is a 500 g because two formats and a longer programme use more than 115 g; at ₹3.19 a gram it is cheaper per gram than the 100 g (₹3.42). And none of this includes labour, freight, packaging or GST, or the scale and thermometer if you lack them. Compare the small-range total, ₹9,977.50, with the ₹16,237.53 worked-example value of the random grid: the sequence is cheaper and it produces answers. How many combinations to test before launch, as a formula, is in How Many Wax × Fragrance Combinations Should I Test Before Launching?.

The calendar: what a start this week can reach

Worked example, starting Monday 28 September 2026 (week 0). The festival dates are calendar facts; the rest is a plan.

If you need Diwali or Christmas stock this year, do not compress the sequence; split it. Sell only recipes you have already proven, and run the programme for the next launch. A shortened sequence that skips the wick stage to hit a festival date is how random testing starts again.

What to buy, in the order you need it

Stage 2 and 3 materials first, then one candidate oil as an example of stage 4. Waxes are chosen for stated ceilings, temperatures and container use; the control oil for a stated soy load inside both waxes' ceilings, a low per-gram price and full stock; not on throw, which CSI has not measured. Candle counts assume 150 g at 8%.

Frequently asked questions