I’m Changing from Soy Wax to a Soy-Coconut Blend but Keeping Every Jar and Fragrance the Same — How Much of My Existing Wick-Testing Data Can I Still Use?

What a wax change resets, and what it can't touch

A wick's job is to draw melted wax and fragrance up to the flame at a rate the pool can supply. Change the wax and you change how it melts, how fast the pool spreads, how the melt moves up the wick and how much heat the candle stores late in its life. CSI's own pages show how far apart two container waxes can sit: Luxury Soy Wax CSI 464 states a pour temperature of 80°C–90°C and "Fragrance Load Up to 10%"; Premium Coconut Soy Wax CSI 468 states a melting point of 50°C–53°C, a pour of 58°C–60°C and "Fragrance Load Up to 13%". Those are different materials, not different batches.

What that does to your wicks is a tendency, not a rule. Many makers find coconut-soy blends form a pool faster and wider than straight soy at the same wick, so the old wick is commonly the same size or one rung too big, rarely too small. Some blends behave the other way, and fragrance still matters. Vanilla Bean's page puts it plainly: "Wick testing is recommended based on your container size and wax type." CSI publishes no wick-performance data for either wax, and no CSI oil has measured wick data in either.

What a wax change cannot touch is the glass. Your jars' inside diameters, wall, base, taper and fill line are measurements of objects that have not changed. Neither has the geometry of where you put the wicks: a two-wick layout's distance to the wall and between flames is arithmetic on the jar. That is the line this whole page draws. Everything about the object carries; everything about the burn restarts.

The carry-over map, item by item

Go through your soy batch sheets and test logs and give every field one of three labels. The table is our judgement, built from what the two wax pages state and from the tendencies above. If a field is not on your sheet, that is a gap to fill now, before the new wax arrives; the batch-sheet post lists what a traceable sheet records.

Notice that nothing in the "carries" pile is a wick result. That is uncomfortable if you spent months bracketing in soy, but it is the honest reading: those months told you where to start, which jars are hard, and which oils misbehave. That is worth a great deal of saved testing. It is not worth a sale.

Choosing representatives: the edge, not the favourite

A representative is not your bestseller or the oil you like most. It is the member of a group whose soy record sat nearest the edge in the direction you expect the blend to push. If the blend tends to make pools faster and flames livelier, the light-group representative is the oil that already ran warmest in soy, and the heavy-group representative is the one with the most carbon build-up or the slowest edge, because that is where a step down is riskiest. Choosing representatives from records rather than names is the core of the behaviour-groups post; how far grouping can collapse a large matrix is the grouping-maths post.

Eucalyptus Spearmint's page says "No vanillin - zero risk of discoloration" and states "Moderate to strong throw"; Vanilla Bean's warns of "Potential discoloration" at 2.5% vanillin. Those page facts are supporting evidence for the grouping, not proof of how either oil burns. The records decide, which is why a member that behaves unlike its group in Stage B leaves the group rather than dragging the group's wick with it.

Stage A: re-brackets on representatives only

Stage A re-brackets each representative in each jar on the approved soy wick and, where CSI has one, the rung below. Everything else is held: same jars from the same lot, the soy layout and spacing, 8% load on total weight, the CSI 468 page's 58–60°C pour, one cure for all, one still room. Two candles per arm keeps one odd candle from deciding.

Stage B: every member still earns its own full life

Once a representative's cell is decided, each other member of its group in that jar gets two candles on the winning wick, burned through full life. No second arm: the representative has already told you where to start, so the member's job is to confirm it. In the worked example that is British Rose and Lavender in the light group, Dark Vanilla Oud and Roasted Coffee in the heavy group, in both jars: 16 candles.

A member passes if both candles clear full life inside the same standard as the representative. If one or both fail, that member has told you something the soy data did not: it leaves the group in CSI 468 and gets its own two-arm bracket, in the direction it failed. Record it in an exceptions log so that the next wax or jar change starts from the truth. If two members of the same group fail the same way, your grouping does not survive the blend; re-bracket the group on its other edge member before confirming anyone else.

The same principle, applied to a single wick-type change rather than a wax change, is the cotton-to-wooden requalification: what carries (fill, load, cure, jar) and what restarts (wick, count, spacing, full life). And if you are also thinking about using the higher ceiling, wait: the 8% to 10% load-change plan is a separate requalification that should start only after this one is finished.

Twenty-eight candles against a full restart

A full restart would re-bracket all 12 approved candles on their own: the six light-group candles on one arm each (nothing below their wick), the six heavy-group candles on two arms, two candles per arm: 36 candles. The representative plan pours 28: 12 in Stage A and 16 in Stage B. That saves 8 candles, half of them large, because the 10 cm jar takes 77% of the wax.

Oils are costed at each oil's 1 kg price per gram, the size most brands keep in production stock. Each member uses 104 g, which is 4 g more than a 100 g bottle; plan the stock check before pouring. Per candle, CSI 468 costs more than CSI 464 at 10 kg prices: ₹17.88 more wax cost in a 150 g candle and ₹59.62 in a 500 g candle. That difference belongs in your pricing decision for the switch itself, not in this test budget.

The process sheet changes too

The wick plan is only half of a wax change. These lines on your batch sheet change on day one, before any burn result exists.

Two of those lines tempt makers into changing a second variable. The 13% ceiling invites a stronger candle; at 13% of a 150 g candle that is 19.5 g of oil instead of 12 g, and it would make every Stage A result uninterpretable. And the lower pour temperature invites a new pouring routine; change the temperature because the page says so, but keep the jar pre-warming, cooling and room the same so that only the wax has changed. Whether a higher ceiling should ever mean a higher load is the 12%-capacity question.

What to buy, and what to skip

Four items, chosen on page statements, price and stock on 29 September 2026. Your jars and oils carry over from stock, so they are not on this list. CSI publishes no wick-performance data, so nothing is ranked on burn results.

Frequently asked questions