Should Every Commercial Batch Have Retained Candles From the Beginning, Middle and End of the Production Run?

Why three positions rather than one candle

A single retained candle is a snapshot. It shows what a typical candle from the batch looked and smelled like, which is useful when a complaint arrives. But the most common scale-up failures are not whole-batch failures; they are changes during the run. The wax cools and thickens in the pitcher, oil distribution shifts as the melt stands, volatile notes leave hot wax over a long pour, a refill introduces a step. One candle cannot see any of that.

Beginning, middle and end retains can. If a customer who bought a candle from late in the run complains about weak throw, you can burn the end retain beside the beginning retain and see whether the batch itself drifted. Without them, you are left guessing from memory and a batch sheet that records intentions, not results.

There is a second benefit. Three retains per batch, kept over months, build a library. When you change a wax pack, an oil lot or a wick box, you can compare the new batch against a retained candle from the last good one, poured the same way and aged in the same conditions.

Retained candles are not test candles

It is easy to blur the two, and that blurring causes trouble. Release-test candles are pulled so you can burn them now and decide whether the batch ships. Retained candles are pulled so you can look back later. If you burn a retain for release testing, you no longer have a retain.

In practice, pull both at each position: one to burn for release, one to keep sealed. That makes six candles per batch in total, of which three survive as retains. If six feels like too many for small batches, keep the three retains and pull release-test candles from beside them in the pour order.

Retained finished candles also differ from retained raw materials. A sealed sample of wax or oil tells you what came in; a retained candle tells you what came out of your process. Both matter, and they answer different questions.

What it costs

Take a 25 kg batch of 200 g candles at 8% of total weight, using CSI 464 and Jasmine, with ready-made cotton braided wicks. Each candle contains 184 g of wax and 16 g of oil.

Wax: CSI 464 is ₹507.40 per kg at every pack size, so 0.184 × ₹507.40 = ₹93.36. Oil: Jasmine at ₹3,923.00 per kg gives 0.016 × ₹3,923.00 = ₹62.77. Wick: ₹12.00. That is ₹93.36 + ₹62.77 + ₹12.00 = ₹168.13 per candle before the jar and labour. Three retains cost 3 × ₹168.13 = ₹504.39 per batch.

If you run 20 such batches a month, retains cost 20 × ₹504.39 = ₹10,087.80 in wax, oil and wicks. Measured against 2,500 candles of output, it is 60 candles a month — real money, which is why the decision should be deliberate rather than automatic for every tiny run.

The retain counts in the table are suggested starting points. Validate them against how often you actually open retains and what you learn when you do.

The usual objections, answered honestly

I do not have the space

Three 200 g candles per batch, twenty batches a month, is sixty candles — about two shoe-box-sized cartons a month, fewer once you discard on schedule. If space is genuinely tight, keep the full three for your bestsellers and your largest batches, and a single middle retain for short runs of slow sellers. A partial system you keep up is better than a complete one you abandon.

I can always make another candle to compare

You can make a new candle, but not an old one. A fresh candle made today uses today's wax pack, today's oil, today's thermometer and today's cure clock. When the question is "what did that batch in March actually contain?", only a candle poured in March can answer it.

I burn-test every batch anyway

Good — keep doing it. Release testing tells you the batch was acceptable at the moment you tested it. It does not let you revisit that judgement later against a complaint, and it usually samples one position, so a drift towards the end of the run can pass unseen.

What if I have never kept retains?

Start with the next batch. For past batches still in stock, pull one unsold candle from each as a makeshift retain, noting that its pour position is unknown. It is weaker evidence than a planned retain, but better than none if a question arrives about that batch.

How to take, label and store them

When a batch is poured from more than one melter fill, treat each fill as its own run and keep beginning, middle and end retains from each. A refill is exactly the kind of event that can split one "batch" into two slightly different products, and the retains are how you would ever know.

When to open a retain, and when to let it go

Open retains when a question needs an answer: a customer complaint, a suspected drift, a new material lot to compare, or a batch you want to reproduce exactly. Burn the relevant retain beside a fresh reference at a comparable age, blind-coded, and judge by smell.

Set a keeping period and discard on schedule. A suggested starting point is the time the batch will be on sale plus a margin that covers customer use and a complaint window. Review it every few months: if you have never needed a retain older than a certain age, you can shorten the period; if a complaint arrives after you discarded the evidence, lengthen it.

Write down what you find each time a retain is opened, even when the answer is "no difference". Over a year those notes show which products drift, which positions tend to be weakest, and whether your keeping period is right. They also give you something concrete to discuss with a customer, or with CSI on WhatsApp if a material question comes up, instead of a recollection.

When you discard a retain at the end of its period, record that too. A register showing that a retain existed and was disposed of on schedule is still useful when a late question arrives.

Retains are a small cost that turns guesses into comparisons. They are most valuable precisely when production is growing, batches are larger, and the distance between the first candle and the last is widest.

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