How Do I Create a Quality-Control Checklist for Glass Candle Jars Before They Enter Production?

Three layers: the delivery, every jar, a sample

Glass jar faults come in two kinds, and they need different checks. Random faults, such as a chipped rim, a crack from a knock in transit or a stone in one jar's wall, can be in any jar, so a sample will miss them; only looking at every jar catches them. Systematic faults, such as a whole lot that is 2 mm wider, a heavier base, a lid that sits loose on every jar, are shared across the delivery, so measuring every jar wastes time; a spread-out sample finds them. A good checklist puts the quick checks on every jar and the slow ones on a sample.

This is an incoming-inspection checklist for jars you have already approved. Choosing and approving a jar in the first place is a different job with different tests: the measurements to take before a first bulk order are in the bulk measurement checklist, and heat and burn-down approval of a new container is in the pre-500 heat test protocol and the small-brand safety programme. If you are bringing in thousands of imported jars under a formal sampling plan, the imported-jar approval guide goes further than we do here.

The checklist itself, item by item

Print this, cut what does not apply to your jar, and add anything your jar needs (a lid gasket, a printed pattern). The class column tells the inspector where a failure goes without having to ask.

Why a few of these lines are there

Chips, cracks and surface stones are critical because they are commonly cited as places where cracks start once glass is heated; a candle is a heat source sitting inside the jar for hours. Uneven wall or base thickness is major, not minor, for the same reason: uneven glass is commonly linked to uneven heating and thermal stress, and a jar that looks lopsided from the rim is a jar you have not tested. Wobble matters twice: a rocking jar is a candle that can tip, and it often signals a base that is not level, which puts the wick off-centre to the pool. The detail on why a chipped rim is a reject and never a "use with care" is in the chipped-rim guide; how uneven glass behaves under heat is in the heavier-glass guide.

The "right product" line exists because listings, ours included, are not always consistent. On 26 September 2026 several CSI listings carried text that contradicts the title: the Black Glass Jar with Black Lid (150 gram) is titled with a black lid while its description mentions a "Golden metal lid"; some coloured 200 ml Salsa listings have web addresses naming a different colour from the jar; a few 150 g listings still carry a stale "Deliveries Starting 5th Dec" in the title. If a colour or lid matters to your product, confirm it on WhatsApp before ordering and check it at the gate when the jars arrive.

How many jars to check, and when to check more

For the measured checks we use a simple rule of our own: 10 jars or 2% of the delivery, whichever is more, drawn from as many cartons as possible (for example, two from each of five cartons, not ten from the top carton). It is not a statistical standard; it is enough to see whether a lot has shifted, which is what those checks are for. Formal acceptance-sampling tables exist and are worth learning if you receive thousands of jars at a time; the imported-jar guide linked above introduces them.

Tighten the sample (double it) for a first delivery from a supplier, a first delivery of a new jar, a delivery that shows a new lot or date mark, or the delivery after one that had problems. Relax back to the base rule after three clean deliveries in a row. Never relax the every-jar layer: a chip in jar 347 is not visible in the sample.

Critical, major, minor: what each failure means

A three-class system stops the inspector negotiating with each jar. The classes are about consequence, not appearance: a tiny chip is critical because of what it can do under heat, while an obvious seam line can be minor because it cannot.

Two judgement calls come up often. Bubbles: small seed bubbles inside the glass are common in mass-produced jars and usually cosmetic; a bubble that breaks a surface, or a solid speck with fine cracks around it, is a different thing and goes in the critical pile. If you sell a minimal clear-glass candle where every flaw shows, you may choose to grade visible bubbles as major for appearance, not safety. Finish: coloured, matte and frosted jars can be coated or treated, and the pages don't say which process was used, so a rub check at intake is about appearance only. Whether a finish tolerates repeated heating is a separate test, covered in the frosted-coating heat test.

Setting up the QC bench: a walkthrough

Clean glass is also where some cosmetic problems start or stop. Wet spots, the patches where soy wax pulls away from the glass, are commonly linked to cooling and temperature differences and to glass cleanliness; a clean, dry, room-temperature jar is one of the usual mitigations, though never a certainty. The full treatment is in the clear-glass wet-spot guide.

Records that make the checklist worth keeping

The record has three jobs: to show a supplier exactly what arrived, to show you whether a jar is drifting over time, and to let you trace a customer complaint back to a delivery. One line per delivery is enough if it has the right columns.

Keep the reference jars. For each product we would keep three unpoured jars from the approved lot, which in our worked example costs ₹283.20 of a 500-jar order worth ₹47,200.00 (0.6% of it). They are what every future delivery is compared against, and they are your evidence if a supplier's jar quietly changes. The questions to ask a supplier so that jar stays the same for years, and what to keep alongside the reference jars, are in the reorder-for-years guide.

What QC costs, and what it saves

Here is the arithmetic for one worked example: 500 Clear Glass Jars with Golden Lid (150 gram) at ₹94.40 each (the pack of 20 is ₹1,888.00, the same ₹94.40 per jar), made into 150 g candles at 8% with Luxury Soy Wax CSI 464 (138 g, ₹70.02), 12 g of Jasmine Fragrance Oil at the 1 kg price (₹47.08) and an Eco C2 wick (₹9.44). One finished candle's materials cost ₹220.94; excluding labour, freight, packaging and GST.

The time cost is real but small: about 15 seconds a jar for the look-touch-stand layer (our estimate), so roughly 2 hours for 500 jars, plus about half an hour for a 10-jar sample. Every flawed jar caught at the carton rather than after pouring saves ₹126.54 of wax, oil and wick in this example (₹759.22 for the six), before labour and before any complaint; that is what the two hours buy. The six-jar count is illustrative; your own record sheet will tell you your real numbers within a few deliveries.

One number the table can't hold is the safety side. A chip or crack is commonly cited as a place where glass can fail under heat. A candle-maker's checklist exists first to keep those jars away from a flame, and only second to save money. For what happens when a jar fails during a burn anyway, see the cracked-while-burning guide.

What to buy for the QC bench

Everything below was in stock on 26 September 2026. You don't need all of it: the scale is the one piece of kit the sample checks can't do without. Prices are CSI live pricing; per-unit figures are pack price ÷ units.

Frequently asked questions