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

Why "does it work" is the wrong test

Here is what happens when a maker tests preheating without costing it first. They warm twenty jars, pour them beside twenty cold ones, and look at the results. The warmed half looks a little better — maybe two fewer patches, maybe one. The conclusion writes itself: preheating helps, add it to the SOP. And it did help, by a small amount. What the trial never asked is whether that small amount was worth forty seconds a jar for the next three years, which is the only question a production decision actually turns on.

The asymmetry is the problem. A benefit of "two fewer patches in twenty" sounds like a win and is almost impossible to evaluate by eye. A cost of "forty seconds a jar" sounds like nothing and compounds into the largest number on your cost sheet. At a thousand candles a year, forty seconds a jar is eleven hours of work. At five thousand candles it is fifty-six hours — more than a working week — spent on an operation that nobody has ever checked against a target.

So reverse the order. Work out the number preheating has to beat before you pour anything, write it at the top of the sheet, and then run the paired batch. Now the trial has two possible outcomes instead of one, and the second outcome — "it helped, and not enough" — is the one that saves you the most money. This is also why the comparison has to come from one wax batch on one day: a difference of two candles in twenty is small enough that any uncontrolled variable will swamp it.

Costing a minute of your own line

Two inputs, both yours: seconds per jar and rupees per hour. Neither is published by CSI or by anybody else, and both vary enormously between workshops, so the table below gives you the grid rather than a figure. Find your row and your column; that cell is what preheating costs you per jar, every jar, for as long as the step survives in your process.

Two things in that table matter more than the arithmetic. First, the spread between the cheapest and dearest cell is more than eleven times — ₹0.56 against ₹6.25 — which means how you warm jars changes the economics far more than whether you warm them. A carton moved once beside the melter and a gun used on each jar individually are not the same intervention with different thoroughness; they are different businesses. Second, the step that adds the least labour is also the one you cannot aim precisely, which is the genuine trade-off of this whole section.

What one cosmetic reject actually costs

Now the other side. A gap in clear glass is cosmetic — it does not change the burn, the melt pool or the fragrance, and it does not make the candle unsafe — so what it costs you depends entirely on what you do with the candle. There are three honest answers and they differ by a factor of ten, so pick the one that matches your practice rather than the one that makes the sums easiest.

The second and third rows are the ones most small makers are really in, and they make preheating harder to justify, not easier. A candle sold as a second at a ₹100 cut has cost you ₹100, not ₹249.80, so the step has to work almost twice as hard. That is an uncomfortable finding and it is the right one: if your patchy candles still sell, the commercial damage is smaller than it feels, and the case for a permanent labour step is correspondingly weaker. The reverse is also true — if a patch means a customer complaint and a replacement, cost it at the replacement, not the material.

The break-even table

Divide labour per jar by cost per reject and you get the percentage-point cut the step has to deliver. Here is the full grid for the scrapped case at three candle values and three handling times, at a ₹150-an-hour rate. Read your cell, write it at the top of your trial sheet, and do not look at it again until you have counted the two halves of your paired batch.

Those last three figures are the ones to carry around. On your biggest jar, preheating only has to rescue two or three candles in a thousand to pay for itself — which is a bar so low that if your paired batch shows any consistent improvement at all, the step is probably worth keeping. On a cheap votive it has to rescue fifteen and a half, which is a real, visible difference you would have noticed already. Same step, same seconds, opposite conclusions. The decision is not about preheating; it is about which product you are preheating.

And one worked example to show the scale when the problem is genuinely bad. Suppose you pour a thousand of the 150g clear jars a year and your logged reject rate is 8%. If a paired batch shows preheating takes that to 5%, you save 30 candles × ₹249.80 = ₹7,494 against ₹1,666.67 of labour — net ₹5,827, and an easy yes. If it only takes 8% to 7%, you save 10 candles × ₹249.80 = ₹2,498 against the same labour — net ₹831, which is a marginal yes that you might still decline because eleven hours of your year is worth more than ₹831. Both are defensible. Neither is available to a maker who never costed the step.

The timed paired trial, step by step

Sixty jars, not twenty. The reason is arithmetic: if the effect you are hunting is worth 0.67 of a percentage point, you cannot see it in twenty jars, and even sixty will only show you an effect large enough to matter commercially — which is exactly the effect you care about. Thirty warmed, thirty not, one pot, one day, everything else frozen.

Four ways this trial will lie to you

Every one of these has sunk a real trial, and all four are free to prevent.

Writing the decision down

Whatever the trial says, the output is a short written decision rather than a habit. Four lines are enough, and they should live with your recipe card: the method and its measured dwell time, the labour rate and seconds you used, the break-even figure, and the counted result with the date. If you ever hire, sell the business or simply come back to it after a summer, those four lines are the difference between a process and a superstition.

It is also worth recording the decision per product, because the break-even table says it should differ. It is entirely coherent to preheat for your Apothecary line and not for your votives — same workshop, same wax, same day, two different decisions, both justified by the same arithmetic. Makers resist that because it feels inconsistent. It is not inconsistent; it is the only conclusion the numbers support. If your range spans ₹107.53 and ₹720.17 of material per candle, a single process rule for both will certainly be wrong for one of them.

Frequently asked questions