I Need 1,000 Candles but Expect 2% Jar Breakage — How Many Extra Jars Should I Include in the Purchase Order?

The short answer: divide by 0.98, and only for jars

If 2% of the jars you handle will break, then only 98% of the jars you buy become candles. To end up with a fixed number of good jars you divide by 0.98; you do not multiply by 1.02. For the bare question — exactly 1,000 jars wanted, nothing else in the way — that is 1,000 ÷ 0.98 = 1,020.41, so 1,021 jars. The shortcut "add 2%" gives 1,020, and 2% breakage on 1,020 jars leaves 999.6 on average: a fraction of a jar short before anything else has gone wrong. On a single order that difference is one jar. The bigger mistakes on this question are elsewhere, and this page is about them.

The first is what the 2% is a percentage of. You do not pour exactly 1,000 candles for a 1,000-candle order. In this example you pour 8 test candles first and carry a 3% reject allowance (both labelled working assumptions), so the line pours 1,040 candles. Every one of those needs a jar that survived, so the breakage divides 1,040, not 1,000: 1,040 ÷ 0.98 = 1,061.22, rounded up to 1,062 jars. The second is where the 2% happens. A jar that cracks in the carton on arrival costs you a jar. A finished candle that cracks in the courier van costs you the jar, the wax, the fragrance, the wick and the sticker — and it belongs in a different line of the budget entirely.

The allowance stack for 1,000 candles, in order

This series applies allowances in one fixed order so that nothing is counted twice. The example candle is a 200 g White Matte jar candle at 8% in Luxury Soy Wax CSI 464 with Japanese Cherry Blossom — 184 g of wax and 16 g of oil per candle. The 200 g fill is illustrative: CSI does not publish a fill weight for the White Matte jar, so weigh a test pour before you order 1,000 of anything. The 8 test candles, the 3% reject rate and the 3% pour loss are working assumptions you replace with your own; only the 2% breakage comes from the question.

Notice what step 4 does not touch. Wicks follow the candles poured (1,040, plus a small damaged-wick allowance that the single-wick post sets properly). Wax and oil follow the candles poured plus pour loss. If you multiplied the wax by 1.02 as well "for breakage", you would buy about 3.9 kg of wax that has no jar to go into. On a 200 g candle at CSI 464's 10 kg rate that is roughly ₹2,000.18 of wax bought for a breakage that cannot use it. The buffer-logic post explains why consumables and glass need different arithmetic in general; this page applies it to one number.

Where your 2% happens changes what you buy

Ask where the breakage figure came from. If it is your experience of cartons arriving with cracked or chipped jars, or jars lost while washing, wicking and pouring, it is glass breakage and this page's arithmetic applies. If it is your courier's breakage rate on finished candles, it is not a jar allowance at all — it is a replacement-candle allowance, and every replacement needs wax, fragrance, wick, sticker and jar. With the example candle, a finished candle carries about ₹250.03 of CSI material (wax at CSI 464's 10 kg rate, oil at the 1 kg rate, an Eco Thin C1 wick, the jar and a sticker), so 2% of 1,000 dispatched candles is 20 candles and ₹5,000.67 of material — before labour, packaging and courier, which are yours to add.

That split matters because it stops two allowances covering the same jar. A jar that cracks when hot wax goes in is already a reject: it took wax, oil and a wick with it. If you count it in the 3% reject line and again in the 2% breakage line, you have bought two jars for one failure. The cleanest rule is: glass breakage covers empty glass only. Anything that breaks with wax in it is a reject or a replacement, sized with the whole-candle maths.

Why the average is not enough: the odds of running short

A breakage rate is an average. If your jars break independently at exactly 2% — a simplifying model, labelled as such, because real breakage comes in clusters such as one dropped carton — then buying 1,062 jars for 1,040 candles gives you enough good jars only about 62% of the time. That is what "on average" means: in roughly four runs out of ten you break a few more than average and finish a handful short. The table shows how many jars buy you more certainty under the same model.

Read the middle rows as a business choice, not a rule. Going from 1,065 jars to 1,070 — enough for about 95% confidence under the model — costs ₹354.00 more. Whether that is worth it depends on what running short costs you: a delayed dispatch, a rush reorder, a partial delivery to a corporate client. For a one-off order with a fixed delivery date, many makers would pay it; for a repeat retail run where a few leftover jars simply go into next month's stock, the 1,065 line is enough. Unused jars do not spoil, which is why glass is the one material where a slightly generous buffer is cheap insurance — see how buffers differ between wax and glass.

If your breakage is not 2%

Your rate may be lower with well-packed cartons and careful washing, or higher if jars travel a long way or your team is new. The table reruns the jar line for 1,040 candles poured, with the correct division beside the add-a-percentage shortcut and the White Matte pack rounding.

Two things stand out. At low rates the division and the shortcut agree to within a jar, so the method matters less than getting the base right (1,040 poured, not 1,000 sold). At 5% the gap opens to 3 jars, and at higher rates it keeps widening. Each extra percentage point of breakage costs roughly ₹736.32 in White Matte jars on this order — which is why packing and handling are worth fixing before you buy around them.

Rounding the jar line to real CSI packs

White Matte Glass Jar is listed as a pack of 5 at ₹354.00 and a pack of 10 at ₹708.00 (re-checked 4 October 2026) — ₹70.80 a jar either way, so there is no bulk break and the pack choice is only about covering the count. 1,062 jars is 106 packs of 10 plus 2, so the cheapest cover is 106 × 10-pack + 1 × 5-pack = 1,065 jars for ₹75,402.00, leaving 3 jars over the breakage-adjusted count. CSI does not publish the White Matte jar's fill weight; if your candle is not 200 g, weigh a test pour and redo the wax line, but the jar line does not change.

If you have not chosen the jar yet, the same jar count rounds differently on other CSI glass. Fill weights for these are not published either, so the table is a pack-rounding comparison only — it says nothing about which jar suits a 200 g candle.

The Salsa jar is sold in 5s, 50s and 100s and comes with a golden lug cap, so the 1,062 requirement rounds to eleven 100-packs or a mix; the Yankee 130 ml with clear cap is sold singly at ₹47.20, so it rounds exactly. Mini Glass Jars are a small-format jar; weigh a test pour before assuming any fill. Whichever jar you choose, the breakage maths is the same: candles poured ÷ (1 − breakage), then round up.

The priced CSI shopping list for the whole order

Here is the full order with the breakage allowance in its proper place. Wax and oil carry tests, rejects and pour loss; wicks and stickers carry tests, rejects and a 10-wick damage allowance; jars carry tests, rejects and the 2% breakage. Every line is rounded to the cheapest listed, in-stock CSI packs.

Three notes on the rounding. The wax line is 19 × 10 kg + 1 × 5 kg + 3 × 1 kg because CSI 464 is a flat ₹507.40 a kilo from 1 kg to 10 kg, so the cheapest cover is simply the smallest overshoot. The oil is 17 × 1 kg + 1 × 100 g + 1 × 50 g: Japanese Cherry Blossom is ₹4.84 a gram at 1 kg against ₹5.00 at 100 g and 500 g, so the kilos carry the order and the small packs top up. The wick line is a 1,000-pack plus a 50-pack because the Thin C1 10-pack is out of stock on the 4 October re-check. Leftovers are small: about 0.90 kg of wax, 11 g of oil and 3 jars.

Buy now, buy after testing

Jars are usually the easiest line to commit early on a client order: the client has chosen the container, it does not age, and early delivery gives you time to count breakage. The wick is the line to hold back, because the 1,000-pack is only right if your burn tests pass. Oil depends on whether the fragrance is already approved.

The rest of the 500–1,000 candle cluster

This page owns one allowance — glass breakage. The neighbouring posts each own another, and together they build the full procurement sheet. Single-wick allowances and double-wick allowances size the wick line; colour trials size the dye and trial wax; the new-fragrance post and the unvalidated-wick post decide what waits for approval; emergency remakes sizes the replacement candles that outbound breakage would draw on; the buffer-logic post sets different rules for wax and glass; combined or separate BOMs organises them; and the procurement sheet puts every line in one copyable table. Earlier in the cluster, pilot, test, reject and replacement allowances sizes the consumables side.

Frequently asked questions