The Last Candles Poured From My Batch Have Weaker Fragrance Than the First — Could Extended Heated Holding Contribute Even If the Initial Mixing Was Correct?

Weak last candles have two rival explanations

When the tail end of a batch smells weaker, two mechanisms compete. One is loss: fragrance leaving the hot, open surface while the batch waits, so the wax that reaches the last jars genuinely contains less effective scent. The other is distribution: the oil is all still in the melter, but it is no longer evenly spread because the batch sat unstirred, so different draws carry different amounts. A third, duller possibility is that the late candles differ in some other way that happens to line up with pour order, such as a different jar carton or a shorter cure before testing.

You say the initial mixing was correct. It may well have been. But correct mixing at minute five does not guarantee an even mixture at minute 90, and it does not protect oil from time at temperature. The test below is designed so that a single run tells you which of these you are dealing with.

Why it matters: the two causes have opposite fixes. Loss is cured by shortening the hold or lowering the holding temperature within the wax maker's guidance. Distribution is cured by re-stirring before each draw. Pick the wrong one and you pay for a change that does nothing.

The worked example: 10 kg of lavender in coconut-soy

A maker mixes 10 kg of scented wax in Premium Coconut Soy Wax CSI 468 with Lavender at 8% of total weight: 9.2 kg wax + 0.8 kg oil. That fills 50 candles of 200 g before losses. Pouring alone, the maker takes about 95 minutes from oil addition to the last jar. After cure, candles 1 to 10 are strong and candles 41 to 50 are noticeably softer in the room.

What the batch cost, with the working: wax 9.2 kg at the 5 kg pack rate of ₹3,185.00 ÷ 5 = ₹637.00 per kg, so 9.2 × ₹637.00 = ₹5,860.40. Oil 800 g as 500 g ₹2,620.00 + three 100 g packs at ₹540.00 each (₹1,620.00) = ₹4,240.00, which is cheaper here than the 1 kg bottle at ₹5,221.00. Batch total ₹10,100.40, or about ₹202.01 per candle (₹10,100.40 ÷ 50). The ten suspect candles represent roughly ₹2,020.08 of material, which is why they deserve a proper test rather than a guess.

One run that separates loss from distribution

Rule out the dull explanations first

Before spending a batch on the split test, spend ten minutes with your records. Did the late jars come from a second carton? Were candles 41 to 50 wicked from a different reel or pack? Were they tested a day earlier than the rest because they were the last to be boxed? Was the room used for the burn test the same? Any one of these lines up neatly with pour order and can mimic a hold effect.

Also check the temperature record. If the wax reaching the last jars was well below the first pour, the late candles may have set differently at the top and around the glass, and a rough or pulled-away surface can change how a candle burns in its first hours. That is a temperature drift problem, related to the hold but with its own controls. If your notes cannot answer these questions, that is itself a finding: start writing pour times, jar cartons and wick packs on the batch sheet from today.

The repeat batch with three reference sets

Repeat the same batch, same oil lot, same wax lot, same jars and wicks. The only additions are three sets of reference candles poured at deliberate moments. The minutes and counts below are suggested starting points to validate against your own pace.

If the stirred late set matches time-zero, the oil was still in the melter; it had simply stopped being evenly spread. If even the stirred late set is weaker, stirring cannot put back what left the batch, and holding is contributing. Both can be true together, which is why the matrix has a row for it.

What to do with candles 41 to 50

Hold them. Do not ship the tail of the batch while the split test is running, and do not blend them into cartons with early candles, because then you can never trace a complaint back to pour order.

Once the test reads out, the disposition follows. If the stirred and unstirred late candles both matched time-zero, the weakness you noticed was probably not the hold, and you now know to look at jars, wicks or cure. If loss is confirmed, there is no reliable way to top up remelted candles with oil, because you do not know how much effective fragrance left; adding a guessed amount creates a new, untested formula. Honest options are to sell them as seconds with plain wording, to remelt them into a fresh batch at a small proportion and test that batch before release, or to write them off. For ten candles, the arithmetic above tells you what each option is really worth.

Controls that stop the tail fading next time

The controls depend on what the split test found, but three are worth adopting regardless. First, stir before every draw, and write that as a step on the batch sheet rather than a habit someone may skip when busy. Second, record the hold duration as a number: time of oil addition, time of the last pour. Third, keep a first and a last candle from every batch, labelled with pour time, so the next complaint can be checked against a retained sample instead of memory.

If loss was confirmed, the remedy is structural: mix smaller quantities more often so no wax waits as long. Two 5 kg mixes started 45 minutes apart in the same melter will usually beat one 10 kg mix held for 95 minutes, but prove it with one more reference set before you commit. Do not raise the fragrance load to compensate; you will pay for more oil and still pour a batch whose first and last candles differ. If you are unsure how to structure the split, CSI can talk it through on WhatsApp (+91 7397976926).

Where this page fits

CSI already has guides that cover parts of this question. This page covers the failure investigation for the question above; these go deeper on the rest:

Three things to buy for this job

Frequently asked questions