My First Corporate Order Used 5% More Raw Material Than Planned — How Should I Incorporate That Historical Variance into Future BOMs?

Your 5%, as numbers

The title gives one figure: 5% more raw material than planned. That is your number. The rest is a labelled worked example built to land on it: a first corporate order of 1,000 candles, 150 g each at 8% on the total-weight basis (12 g oil + 138 g wax), in your own jar, with Luxury Soy Wax CSI 464, Japanese Cherry Blossom and one Eco C2 wick, planned with a 2% loss allowance (1% rejects, 1% heel) and no retained candles.

The "5%" is by weight, on wax and oil, which is how most makers notice it: the last bag ran out early. Pieces have their own lines and their own percentages, because a jar cannot be remelted and a wick cannot be weighed back. The actual figures come from the post-order review, counted rather than estimated; if you have not done that review yet, do it first, using the post on comparing estimated and actual consumption after a large order. This post starts where that one ends: you know the overrun; the question is how much of it belongs in the next BOM.

One-off or systematic: sorting the overrun

Go through the batch notes, reject log and the review with the people who poured, and give every extra kilo a cause. Then put each cause in one of two columns with one test: will it happen again on the next order unless we deliberately change something? If you have already changed something (pre-warming jars, a pitcher rest, a written colour approval), it is one-off. If nothing has changed, or the cause is part of how pouring works, it is systematic.

Three judgement calls are worth naming. The cold-jar rejects are one-off only because the fix (pre-warming clean jars, as commonly practised for soy wet spots) is in place and the reject rate fell back after day 1; if you are pouring on cold winter mornings again with no change, they are systematic. The colour trial was not a production loss at all: it was development work the client asked for and then cancelled. It cost ₹427.08 in CSI materials alone, and the honest response is commercial, not a buffer: next time, a written colour approval and a change fee for trials the client drops (practical judgement, not legal advice). The unexplained remainder, 0.16 kg of wax, goes into the systematic column; it is too small to chase and too real to ignore.

The one-offs total 4.17 kg of wax and 0.36 kg of oil, about ₹4,047.78 at CSI prices before your jars. That money is gone, but it should not be spent again on every future order as a permanent buffer. The systematic part is what your process actually loses, and that is what the standard has to carry.

Setting a standard loss % per material

For each material, take what was used, remove the one-offs, and compare it with what ended up inside good candles. Loss % = (used − one-offs − content of good candles) ÷ (used − one-offs). Wax and oil are weighed; jars and wicks are counted.

Wax and oil land on the same 3.9% here because every systematic cause in this order was a loss of finished mix (overfill, rejects, heel), which carries wax and oil in the 92:8 ratio. In a real workshop they often differ: oil left clinging in the weighing jug or bottles, or oil topped up "for strength", shows up only on the oil line. If yours differ, keep them separate; never average two materials into one "materials buffer".

Pieces came out on different sides of the plan. Jars ran over (3.2% against 3%), because every reject takes a jar and breakage was 15 rather than the 20 allowed; wicks came in under (2.3% against 3%) because only 6 were defective. A standard should come down as readily as it goes up; a wick buffer padded to match the jars would just leave pieces on the shelf. Why pieces deserve their own buffer at all is the post on jar and wick buffers.

Applying it: why you divide by (1 − loss)

The standard was measured as a share of what you issue: out of every 100 kg of wax issued, 3.9 kg did not end up in a good candle. So to end up with a required amount R inside good candles, you must issue R ÷ (1 − 0.039). Multiplying R × 1.039 applies the percentage to the wrong base and comes up short every time: for the next order, by 0.33 kg of wax. It is a small gap at 3.9%; at a 10% loss, the same mistake leaves you about 1% short on every line. The same logic for reject units is the post on a 4% historical reject rate.

The next order (worked) is 1,500 candles of the same formula plus 10 retains: 208.38 kg of wax and 18.12 kg of oil inside candles. Three ways to plan it:

Method B is the tempting one, and it is wrong twice: it charges the next order for a spill, a cold morning and a cancelled trial that will not recur, and it stacks the 5% on top of the 2% plan, buffering the same heel and rejects twice. Method A is the hopeful one; it repeats order 1's late top-up. Method C plans exactly what your process loses, and nothing else.

Wax is bought to the kilo because CSI 464 is ₹507.40 a kilo at 1, 5 and 10 kg. For the oil, 18 × 1 kg + 500 g + 4 × 100 g (₹91,656.00) is cheaper than 19 × 1 kg (₹91,998.00) despite the dearer small packs, because it buys 100 g less. These figures exclude your jars, labour, freight and GST. Before you give the client a date, confirm stock, lots and dispatch for 217 kg of wax on WhatsApp; we do not promise that one lot covers it, and a new lot gets a side-by-side check before it goes into the order.

Argued against our own interest: method B would sell more wax and oil. It would also leave ₹5,884.80 of material on your shelf after an order that went to plan, money that could have been your margin. Buy method C.

Re-checking after the next order

After order 2, repeat the sort, then pool rather than replace: pooled loss = (systematic loss on order 1 + on order 2) ÷ (issued without one-offs on order 1 + on order 2). Pooling weights each order by its size, so a small order with a bad week cannot swing the standard on its own.

The decision rule we use (practical judgement): if a new order's loss lands within half a percentage point of the standard, leave the standard alone; if it lands further out, update it to the pooled figure. In the worked re-check, order 2 ran at 3.1% against a 3.9% standard, outside the half-point band, so the standard moves to the pooled 3.4%. After four or five orders, many makers switch to the last five orders pooled, so that old habits age out.

Two things restart the history instead of extending it. A formula change (new wax, new oil, new jar or wick) is a new process: losses can move, and so can the burn result, so test first and treat the next order as a first order. The change-control side of that is the professional change-control system post. And a change of scale, say from hand-pouring 1,000 to a melter run of 5,000, changes heels and overfill; start the standard again from the larger run's numbers. The cost side of losses at scale is the post on hidden losses at production scale.

The materials and the one tool

The worked order's materials and the tool every loss % depends on, chosen on page statements, price and stock on 7 October 2026.

The honest summary: at CSI's prices, a too-generous standard costs you full price for every surplus kilo, because neither CSI 464 nor C2 wicks get cheaper in bigger packs. The saving is in buying the right quantity, not a bigger one.

Frequently asked questions