I Have 20 Fragrances Using One Wax — How Can I Efficiently Qualify a Replacement Wax Without Blindly Testing Every Possible Combination?

Why "every combination" is the wrong target

A combinatorial test plan grows by multiplication. Every fragrance multiplies by every wick option, every load you might try and every repeat candle. For a twenty-fragrance range the numbers get silly quickly, and they also answer questions nobody asked — you are not trying to change your loads, you are trying to find out whether your existing formulas survive a new wax. Most of that matrix re-proves the same wick decision twenty times over.

The burn-time column is where a big matrix really hurts. Candles are not the bottleneck; burn positions are. With twelve positions (an example — count your own shelf space, with controls burning beside the tests) and four 3–4 hour cycles per candle, the full matrix needs 40 sessions. The screening plan needs 19. That difference is the one you feel in September, with a festive order waiting behind the test shelf. For the ten-fragrance version of this problem, where a simpler anchor-and-confirm split is enough, see the discontinued-wax post.

Step 1 — put the range on a family × tier grid

Two groupings are available to every maker without any chemistry data. The first is family, as the catalogue names the oil — floral, gourmand, fresh or citrus, woody, fruit or spice. The second is price tier, from the oil's own price per gram at 100 g. Neither says how an oil will burn; CSI publishes no measured burn, throw or retention data for any oil, and this page does not guess. What the grid does is make sure that every kind of fragrance in your range is represented in the test, and that a failure in one corner is not silently assumed away in another.

On this example range, the grid has fifteen cells and twenty fragrances. Some cells are empty, some hold three. The black entries in the diagram above are the anchors the next step picks. The tier boundaries — under ₹4.50 a gram, ₹4.50 to ₹6.00, over ₹6.00 — are simply one sensible way to split CSI's catalogue into thirds for this range; draw your own lines.

Step 2 — score each fragrance for risk and pick the anchors

Within each family row, the anchor is the fragrance most likely to reveal a problem. Score each one on facts you hold: its load (higher loads first, and the highest in the range always), its tier (an upper-tier failure is the dearest one to find in production), your own records (a fragrance that needed the bigger wick, or threw weakly, in the old wax), and commercial weight (your top seller). The points below are a working scheme, not a standard: the top scorer in each family is an anchor, and so is anything scoring 4 or more.

The scoring picks 6 anchors. Coffee Vanilla Tobacco runs at 12%, the highest load in the range — still inside the 13% that Luxury Soy Wax Chunks are rated for, but the formula with the least headroom. Vanilla Bean is the top seller. White Royal Oud needed the larger wick in the old wax, which your notes recorded. Eucalyptus Spearmint is the other end of the range: the lowest load and a record of weak throw, so it tests whether a wick chosen for heavy gourmands still delivers at the light end. Raat Ki Rani and Pumpkin Spice are the highest-risk members of their families. The other 14 become confirmers, each following its family's anchor.

Step 3 — run the two stages

Everything below is a working practice you may tighten: three candles per anchor variant, two per confirmer, 3–4 hour burn cycles with full cool-down, cure checkpoints at 48 hours, 7 days and 14 days as examples. The jar, fill weight, wick-centring method and room stay exactly as they were for the old candles. Wick A is your current approved wick, Wick B the one alternative you would move to — for CSI wicks, for example, Eco Thin C1 and Thick C2. CSI publishes no wick size per jar diameter; the test decides.

The scoring sheet and the range tracker

Each test candle is scored against its control candle burning beside it: time to full melt pool, melt-pool depth, flame height at 1 hour and at the end of the cycle, mushrooming, soot on a white card, jar-side reading at the same spot with the same thermometer, consumption in grams per hour (weight loss ÷ burn hours), blind hot-throw and cold-throw scores in a fixed room, and appearance and sweating after storage. The safety lines — flame, soot, mushrooming, jar heat — must be no worse than control and within your jar supplier's guidance and the fire-safety standard you work to (ASTM F2417 is commonly cited in the US). CSI publishes no numeric limit for any of them.

With twenty fragrances, the individual sheets are not the hard part — keeping track of the range is. Use one tracker row per fragrance so the whole programme is visible on one page. The example row below is filled in only with the plan, not with results; the results are yours to record.

The priced qualification kit

The kit uses only listed, in-stock CSI sizes, and oil per candle follows the CSI convention: load is a share of the finished 200 g candle, so 8% is 16 g of oil and 184 g of wax, and Coffee Vanilla Tobacco at 12% is 24 g of oil. Each oil need carries a 5% pour-loss allowance (a working assumption). CSI does not sell a sample pack or discovery set; this is built from single oils in their smallest sensible sizes, and we can help assemble it on WhatsApp.

For comparison, the full matrix — every fragrance with both wicks, three candles each — costs ₹38,558.68 in the same materials, ₹15,851.62 more, with 56 more candles to burn. The jar line is the largest in both plans; if your production jar is not from CSI, use your own, but always the production jar. Coffee Vanilla Tobacco has no 15 g size, so its anchor oil is bought as larger bottles; that is the one line where the kit carries more oil than the test strictly needs.

Decision rule, release and records

Each fragrance ends in one of three states. Pass: no safety line worse than control, throw within one point on the blind score, appearance and storage acceptable, and at least one candle burnt to the end. Fixable: the other wick or a lower load within 13% brings it into line — re-test that fragrance only. Escalate: both wicks fail an anchor, or a promoted confirmer fails again — the family is not qualified in this wax. Release is by family, not all at once: as soon as an anchor and its confirmers pass, that row of the grid can go into a small first production batch, with two candles retained as the new controls.

Record each outcome on the formula card: the new wax name and order date, the load, the chosen wick, the test candle numbers, the tracker row and the sign-off date. The formula-record post shows how to lay that out so the next change starts from evidence, and the change-control form turns this whole programme into one page you can file. If the replacement wax also changes your cost per candle, the total-economics post compares waxes on the finished candle rather than the price per kilo.

Where efficient qualification becomes careless

The efficiency in this method comes from not repeating the wick decision twenty times. It does not come from skipping later burns, controls or storage checks, and those are exactly the corners people cut under time pressure. A confirmer that looks fine after three cycles has still not shown you its last third, where heat builds and the wick's real margin is visible. Keep one candle of every fragrance burning to the end, even after release.

The second trap is letting the grid drift. If you add a twenty-first fragrance next month in a family whose anchor passed, it still needs its own confirmation in the new wax; family membership earns a shorter test, not a free pass. The third is changing the load during qualification because the new wax "allows more". A wax rated to 13% does not ask you to use 13%. Test the old formula first, and adjust only a fragrance whose blind throw dropped — the fragrance-capacity post explains why practical capacity can differ from the published ceiling.

Finally, if the reason for the wax change was cost, not discontinuation, count the qualification against the saving before you start. The ₹15-per-candle payback post does that arithmetic, and the master wax-change checklist is the single page to hand whoever runs the tests.

Frequently asked questions