How Do I Test Whether Preheating Jars Actually Improves Adhesion Instead of Adding Unnecessary Production Labour?

The proof has two halves: fewer spots, and labour that pays for them

Most makers who try preheating judge it by eye: one batch in warmed jars looks better, so the jars get warmed from then on. That skips both halves of the proof. The first half is whether warmed glass really reduces wet spots in your jar, wax and room, or whether the better batch was luck, weather or a cleaner carton. The second half is whether the reduction is large enough to pay for the minutes you will now spend on every tray, every day, for as long as you make the candle. Either half can fail on its own. Warmed glass can reduce spots by too little to cover the labour; or the labour can be trivial and the reduction imaginary.

So the answer is: run a confirmation test big enough that chance cannot explain the result, time the preheat step with a stopwatch, and compare the jars it rescues with the line on the chart above. The single cheapest first step is the stopwatch. On your next production day, time the preheat handling for one tray — loading, measuring, unloading, any extra re-seating of wicks — without changing anything else. That gives you the cost side in your own numbers before you have poured a test jar. Then pour two arms of five jars on each of two production days: ten warmed, ten at room temperature, from one melt per day. Twenty numbered jars, two days, two results that either agree or do not.

This page uses the series default candle — 150 g in CSI's Clear Glass Jar with Golden Lid (150 gram), 12 g of Fresh Strawberries and 138 g of Luxury Soy Wax Chunks at 8%. CSI's Chunks page gives a pour of 75–80°C in its spec table (its step list also shows 70–80°C) and the jar's page states hot pour up to 85°C, so the whole page range is usable in this jar. If you have not yet run a small screen, the glass-temperature ladder or the two-arm preheat test come first; this page is for the maker whose screen looked promising and who now has to decide whether to make preheating permanent.

Why three jars cannot prove a labour decision

Wet spots are lumpy. Two jars poured a minute apart from the same jug can come out differently for reasons nobody controls — a slightly different spot on the tray, a draught, a jar from the bottom of the carton. That is why a screen of three jars per arm is a screen: it tells you where to look, not what to adopt. Before committing labour to every future tray, it is worth knowing how easily chance alone produces a result that looks like a win.

The table below does that arithmetic for a few example outcomes. It asks: if preheating did nothing at all, and the spotted jars just happened to fall where they fell, how often would at least this lopsided a split land on the control side? The figures are a one-sided Fisher exact test — standard probability, computed, not a CSI measurement — and the outcomes are examples, not results anyone has recorded. Read the right-hand column as "how easily luck could have done this".

Two lessons fall out of the arithmetic. With three jars per arm, even the most lopsided possible result — all three controls spotted, no warmed jar spotted — would happen by chance about 5% of the time if preheating did nothing; a two-versus-none split, 20%. With ten jars per arm, a six-versus-one split drops to about 2.9%. More jars buy you certainty, and certainty is what a permanent labour step deserves. That is why the confirmation run below uses ten jars per arm, across two days. The "reading" thresholds in the last column are a labelled working practice, not a statistical law; what to record when comparing adhesion covers how to keep the records that make such arithmetic possible.

The confirmation run: twenty jars over two days

Each production day pours one melt of Luxury Soy Wax Chunks with Fresh Strawberries at 8% into ten numbered Clear 150 g jars: five warmed to your chosen glass temperature (the one your screen suggested; an example arm is 35–45°C, measured and recorded, below the pour and well below the jar's stated 85°C) and five at room temperature. Same wick, same pour temperature inside the Chunks page's 75–80°C, same tray, alternated positions. The second day repeats the first with a fresh melt.

Turn the result into money: the net labour sum

The cost of preheating is easy to see: at the example 9.0 minutes per tray of 20 and ₹150 an hour, it is ₹1.12 a jar, or ₹112.50 per 100 jars. The cost of not preheating is spread out. Each spotted jar may get a heat-gun touch-up or a re-sort — an example 1.5 minutes, ₹3.75 — and then either a bigger opaque label (example ₹12.00) or a discount (example 30% off an example ₹899.00, ₹269.70). Those two treatments are kept apart because they are very different sums: a hidden spot costs a little money and some of the premium look; a discounted jar costs a lot of money.

On those example numbers preheating pays in both treatments, but notice how differently. Where spots are hidden with a label, it needs 7.14 rescued jars per hundred to break even; where spots mean a discount, 0.41. A maker who labels over spots can sensibly decide preheating is not worth it if the confirmation run rescues only a handful of jars; a maker who discounts spotted stock almost certainly should preheat if the run shows any real effect. The chart above draws both lines across a range of handling minutes so you can place your own stopwatch figure on it. The material cost of the candle itself — ₹225.44 at 5 kg Chunks and 1 kg oil rates, with the example ₹5.90 wick and the ₹94.40 jar, excluding labour, packaging, freight and GST — does not change with preheating; counting preheating labour as a manufacturing cost folds it into the full candle cost.

The priced confirmation kit

Listed, in-stock CSI packs. Grams use the CSI convention (oil = W × L) and a 5% pour-loss working allowance. Jars are bought as two packs of 10 rather than the pack of 20 — the same ₹1,888.00 either way, and the pack of 20 is listed in limited stock. The wick line is an example; use the wick you have burn-tested in this jar.

For comparison, a six-jar screen in the same jar costs about ₹1,657.23 in jars, Chunks at the 500 g rate and oil at the 15 g rate. The confirmation run costs more because it has to: it is buying the certainty that justifies a daily labour step. Use production stock wherever you can — the jars and candles from a confirmation run are saleable if they pass, which makes the true cost of the run mostly the spotted jars from the room arm, which you would have made anyway.

The decision rule

Apply after the 7-day grading of both days; check the 30-day photos before writing the SOP.

Whichever way the rule falls, keep the log. The stopwatch figure, the two days' grades and the net-per-hundred sum are what you will need next winter, or when a new jar arrives, or when someone on the team asks why the jars go in a cabinet before pouring. Splitting one wax batch fairly between arms covers the pouring discipline that makes the comparison trustworthy, and whether a wax that removes preheating could justify a higher price looks at the same labour from the wax-buying side.

The CSI products in this run

Frequently asked questions