How Many Candles Should I Make from a New Soy Wax Lot Before Releasing It into Full Commercial Production?

Six candles from the lot, two holds — for a wax you have already approved

Start by separating two situations that often get mixed up. Approving a new wax — a new product or a new supplier — is a full qualification: control candles, full-life burns, wick brackets if needed, a pilot and a second delivery. The supplier-selection hub sets that out, and it takes dozens of candles. Releasing a new lot of a wax you have already approved is a much smaller question: is this delivery the same material you approved, and does it behave the same way at production scale? That takes far fewer candles, because you are checking for difference, not proving fitness from scratch.

For that second question, a practical release for a small business is two holds. R1, the bench hold: pour 3 candles of your standard recipe from the new lot beside 3 control candles from approved stock or a retained sample, cure them to your normal checkpoint, burn them in matched pairs, and take one of each to the end of usable wax. R2, the first-batch hold: make your first production batch from the new lot as normal — 24 candles in the example — but hold the whole batch; pull 3 candles (first, middle and last poured) and burn them beside a control before the batch ships. If both holds pass, the lot is released into full production and you keep counting rejects against its lot code.

That is six candles taken from the lot for testing. On a 10 kg lot that is about a ninth of what it makes; on a 50 kg lot it is about two in a hundred. The rest of this page explains why six is enough for this question and not for others, when to add more, what the holds cost, and how the release fits into the system of lot labels, retained samples and incoming QC the rest of this cluster builds.

How many candles depends on what is new

The lead table sets the candle count by situation. Find your row before you count anything. The counts are labelled working practice for a small business, not a standard; increase them whenever the cost of a mistake is high.

The first row is the everyday case and the one this page prices. Everything else either adds to it or sends you to a sibling. The common mistake is the opposite one: treating a new supplier's first delivery like a routine lot, and releasing it on two candles because "it is the same type of wax". Why a look-alike wax still needs full testing explains why that does not hold.

Why six candles is enough for this question — and not for others

Small numbers of test candles are good at catching large, consistent differences: a lot that sets with a visibly different top, runs a clearly hotter jar or tunnels where the control does not. They are poor at estimating low defect rates. A standard statistical rule of thumb, the "rule of three", says that if you test n items and none fails, the true failure rate could still be as high as about 3 ÷ n with roughly 95% confidence. The table applies it to test candles. It is general statistics, not a CSI measurement.

So six candles do one job well: they catch a lot that is grossly different from the one you approved. They cannot tell you whether the lot's reject rate is 2% or 4%, and that difference matters — the ₹60-per-kg post shows how about 4.4% extra rejects cancels a typical price saving. That job belongs to production records: count rejects per batch, write them against the lot code, and compare lots over time. The delivery record has the columns for it.

The release, step by step

CSI publishes no numeric flame-height or jar-temperature limit; the safety rule is always relative — no worse than your approved control, within your jar supplier's guidance and the candle fire-safety standard you work to (ASTM F2417 is a commonly cited US standard). The standard test candle post fixes the recipe so every lot is tested the same way.

What the release costs, lot by lot

The bench stage uses 579.6 g of the new lot (₹246.21 of Natural Soy Pearl at ₹424.80 a kilo) plus the same of your control wax. The cash outlay is oil, wicks and jars. The first-batch hold costs nothing in materials — the batch is sold if it passes — except the 3 pulled candles.

Two things stand out. The release costs the same six candles whatever the lot size, so larger lots carry it more lightly: about 11.1% of a 10 kg lot but 2.2% of a 50 kg lot. And the first-batch hold protects ₹5,619.82 of materials in the example 24-candle batch for the price of 3 candles (₹702.48). That is the trade the hold makes: a few days' delay on one batch in exchange for never shipping a whole batch from an untested lot.

The release decision

For the example maker buying Natural Soy Pearl, the purchasing consequence is straightforward. Because the release costs six candles per lot regardless of size, fewer, larger lots spread it best — but only if you can store an opened bag well and use it within a reasonable time. A 100 kg month could be 10 × 10 kg bags from one delivery (₹42,480.00, one lot code, ten numbered bags) or 2 × 50 kg if WhatsApp confirms stock (₹42,480.00, the same ₹424.80 a kilo). Either way it is one delivery, one release and one retained sample.

Fitting the holds into a production calendar

The holds add calendar time, not much labour. The practical answer is to keep one lot ahead: while lot 2 is in production, lot 3 arrives, is labelled, checked and goes through R1. By the time lot 2 runs out, lot 3 has already passed its bench hold and only R2 remains. Setting a reorder point from consumption and lead time and safety stock at 50 kg a month work out how much released stock you need to make that possible.

Frequently asked questions