How Do I Calculate Raw-Material Buffer Differently for Wax and Fragrance Versus Fragile Glass Jars?

The short answer: weigh one, count the other

Most buffer mistakes come from using one rule for every material. "Add 10% to everything" sounds tidy, but wax, fragrance oil and glass jars lose quantity in completely different ways, and the buffer should follow how each one is actually lost. Wax and oil are continuous: you weigh them, losses are small fractions of every batch (oil and wax left in the jug, on the thermometer, on the jar rim; candles filled a few grams heavy), and leftovers can be weighed into the next order. Glass jars are discrete and fragile: each jar is either usable or not, losses come in whole units, and they are lumpy — one dropped pack is ten jars, not 0.2% of the order.

So the two buffers take different shapes. For wax and oil the buffer is a percentage of the net requirement — pour loss and a fill tolerance — applied to the grams the candles poured will hold, and then the bag sizes round it up; often the rounding is a large part of the buffer. For jars the buffer is a count: the candles you will pour divided by the share of jars that survive, and then a check that it covers at least one lumpy loss. Neither buffer touches the other: an empty jar that breaks uses no wax, and wax left in a jug breaks no jars.

Why the two materials need different arithmetic

Two rows matter most. Pack rounding: wax and oil come in bags and bottles that rarely match your need, so the bag overshoot already gives you some buffer, and stacking a large percentage on top of it buys a second bag you may not need. Lumpiness: the average breakage rate understates the risk for glass, because losses come in clumps. A buffer that covers 2% on average can be wiped out by one dropped pack. The 2% breakage post shows the probability of running short when breakage is random; this page adds the case where it is not.

The wax and oil buffer, worked

The example is 750 candles of 200 g in Luxury Soy Wax CSI 464 with Roasted Coffee at 8% — 184 g of wax and 16 g of oil per candle. The White Matte jar's fill is not published, so 200 g is illustrative. The reject allowance (3%) and 6 test candles come first and are counted as whole candles: 780 candles poured. Then the consumable buffers apply to the grams those candles hold: a 4% pour-loss allowance and a 1% fill tolerance for candles filled slightly heavy (both labelled working assumptions).

Notice the fill-tolerance line. Overfilling is a quiet consumer of wax and oil: 2 g extra in each of 780 candles is 1.56 kg of finished candle. It is the reason some makers run out of wax on the last tray despite a pour-loss allowance. If your filling is tight, drop it; if your fill line drifts, measure it by weighing ten filled jars and use your own figure. Fragrance oil follows the wax because it is weighed into the same melt — the same percentage applies — but oil is usually fragrance-specific, so a slow-selling oil should carry less spare than a shared wax. The combined-versus-separate BOM post shows how that changes when several fragrances share one wax.

The glass buffer, worked

For jars, start from the same 780 candles poured. At 2% breakage (a labelled example — use your own tally) the rate-based figure is 780 ÷ 0.98 = 795.92, so 796 jars. That is 16 spare. Now the lump check: the White Matte Glass Jar is listed in packs of 5 and 10, and one dropped pack is 10 jars. The rate covers routine breakage spread across the run; a single dropped pack is a separate event on top of it. So add one pack to the rate figure: 796 + 10 = 806 jars, which the 10-packs cover as 810. If your largest single loss has ever been bigger than one pack, use that figure instead.

The rate rule alone gives 16 spare jars, which one dropped pack would mostly use up before routine breakage had even started. The lump rule adds 10 more for ₹708.00. The naive +10% gives more again, but for the wrong reason, and it is calculated on the order (750) rather than on the candles you pour (780). Glass does not age, so spare jars carry into the next order — but they take shelf space, and a slow-moving jar style is still money on a shelf. The 20-day post shows when a glass buffer must be held as finished candles instead.

The same order with one +10% for everything

It is worth seeing what the one-number habit actually buys. Applying +10% to the order of 750 for every line gives the plan below. The totals end up close, which is exactly why the habit survives — but the quantities are wrong in opposite directions for different materials.

The +10% plan buys more wax than it needs and more jars than the breakage justifies, and its oil is computed on 750 candles rather than 780. It looks safe because the totals are similar. The material-specific plan buys each line for the way that line is actually lost — and if your breakage is higher or your pours lossier, you change one number in one row instead of inflating everything. That is the point of separate logic: it tells you which material to adjust.

Setting both buffers from your own records

The priced CSI list with material-specific buffers

CSI 464 is ₹507.40 a kilo from 1 kg to 10 kg, so the bag mix only affects the overshoot, not the rate. Roasted Coffee is ₹3.74 a gram at 1 kg, ₹3.78 at 500 g and ₹3.90 at 100 g, so the kilos carry it and the small packs top up. The jar line is in 10-packs at ₹70.80 a jar.

The rest of the 500–1,000 candle cluster

This page sets the rule; the neighbours apply it. The breakage post sizes the glass line in detail; single-wick and double-wick posts size the wick count; colour trials add a dye and wax line; the 20-day post adds a deadline reserve; combined or separate BOMs decides which buffers are shared across SKUs; and the procurement sheet gives every buffer its own column. Safety stock for shared wax versus a slow oil takes the same logic into ongoing stock.

Frequently asked questions