I Made a Large Fragrance-Loading Error During Commercial Production — How Do I Decide Whether Correcting the Batch Is Cheaper and Safer Than Rejecting It?
शेयर करना
Does the correction lose me candles? The opposite. The batch grows from 22.4 kg to 32.4 kg, so a 200 g fill goes from 112 candles to 162, and material per candle falls from ₹198.32 to ₹168.43.
What is the catch? Three things the material sum ignores: labour (about 8.25 hours here — put in your own rate), remelt loss (do not take a number from a blog; weigh the wax in and the wax out), and the fact that a corrected batch is an untested batch until it has been cured and burned.
When should I reject instead? When the delivery date is inside about 17 days, because you cannot cure and retest in less; when the correction breaches a safety line; or when you cannot sell the extra 50 candles the correction creates.
The correction, to the gram
Start by fixing the convention, because at commercial scale a convention error is itself a loading error. This page states loads as a percentage of wax weight, which is how CSI product pages express them — the Luxury Soy Wax Chunks page says "8% to 13% of total wax weight", and CSI 464 states "Fragrance Load Up to 10%". On that convention, 1 kg of wax at 8% takes 80 g of oil. On the finished-candle convention — load as a share of the whole candle — the same 8% wants about 87 g per kilo of wax. That is a 9% difference in oil, and on a 20 kg batch it is 140 g of oil, which is real money and a real load shift. Write the convention on the batch sheet.
Now the formula. With W grams of wax already in the batch, an actual load of L1 and a target of L2, the oil in the batch is F = W × L1, and:
| Step | Expression | Substituted | Result |
|---|---|---|---|
| Oil already in the batch | F = W × L1 | 20,000 × 0.12 | 2,400 g |
| Wax required at the target load | F ÷ L2 | 2,400 ÷ 0.08 | 30,000 g |
| Wax to add | (F ÷ L2) − W | 30,000 − 20,000 | 10,000 g |
| Same thing, one line | F × (1 ÷ L2 − 1 ÷ L1) | 2,400 × (12.5 − 8.3333) | 10,000 g |
| Batch weight before | W + F | 20,000 + 2,400 | 22,400 g |
| Batch weight after | (F ÷ L2) + F | 30,000 + 2,400 | 32,400 g |
| Candles at a 200 g fill | batch ÷ 200 | 22,400 ÷ 200 → 32,400 ÷ 200 | 112 → 162 |
The part that surprises people is the candle count. Correcting a load error by adding wax does not shrink the batch — it inflates it, because you are keeping all the oil and buying the wax it should always have had. Fifty extra candles is fifty extra jars, fifty extra wicks and fifty extra labels, which is either an opportunity or a second problem depending entirely on your order book.
| Line | Faulty batch at 12% | Corrected batch at 8% | Change |
|---|---|---|---|
| Wax | 20,000 g · ₹10,148 | 30,000 g · ₹15,222 | +₹5,074 |
| Oil | 2,400 g · ₹12,064.32 | 2,400 g · ₹12,064.32 | ₹0 |
| Batch material | ₹22,212.32 | ₹27,286.32 | +23% |
| Batch weight | 22,400 g | 32,400 g | +45% |
| Candles at 200 g | 112 | 162 | +50 |
| Material per candle | ₹198.32 | ₹168.43 | −₹29.89 |
| Inside the wax page's stated load? | No — CSI 464 states up to 10% | Yes — 8% | — |
On the floor, the order of operations matters as much as the sum. Six steps, and the first two happen before anything is heated.
The threshold rule for your own numbers
Your batch will not be 20 kg at 12%. Here is the rule in a form you can apply to whatever you are holding, and it is pleasingly simple because everything cancels.
The correction's material cost is F × (1 ÷ L2 − 1 ÷ L1) × wax price per gram. The material at risk if you reject is dominated by the oil, worth F × oil price per gram. Divide one by the other and F disappears. Correction beats rejection on material whenever (1 ÷ L2 − 1 ÷ L1) is less than the oil's price per gram divided by the wax's price per gram. The batch size is irrelevant. Only the two loads and the two prices matter.
| Actual load → target load | 1 ÷ target | 1 ÷ actual | Left side of the rule | Wax to add per 100 g of oil in the batch |
|---|---|---|---|---|
| 10% → 8% | 12.50 | 10.00 | 2.50 | 250 g |
| 12% → 10% | 10.00 | 8.33 | 1.67 | 167 g |
| 12% → 8% | 12.50 | 8.33 | 4.17 | 417 g |
| 16% → 10% | 10.00 | 6.25 | 3.75 | 375 g |
| 20% → 10% | 10.00 | 5.00 | 5.00 | 500 g |
| 24% → 8% (a trebled dose) | 12.50 | 4.17 | 8.33 | 833 g |
| Fragrance oil | 1 kg price | Per gram | Right side of the rule | Correction wins on material up to… |
|---|---|---|---|---|
| Citrus Lemon Fragrance Oil | ₹2,832 | ₹2.832 | 5.58 | a 20% → 10% error, but not a trebled dose |
| Roasted Coffee Fragrance | ₹3,738 | ₹3.738 | 7.37 | a 20% → 10% error comfortably |
| Sandalwood Fragrance Oil | ₹5,026.80 | ₹5.0268 | 9.91 | almost a trebled dose |
| Lavender Fragrance Oil | ₹5,221 | ₹5.221 | 10.29 | a trebled dose |
| Vanilla Bean Fragrance Oil | ₹7,646 | ₹7.646 | 15.07 | any realistic error |
| Santal Noir | ₹9,120 | ₹9.12 | 17.97 | any realistic error |
Read the two tables together and the conclusion is blunt. The worked case is 4.17 against 9.91 — the correction costs 42% of the oil's value, so on material it is not close. Even a trebled dose corrected back to 8% sits at 8.33, still under Sandalwood's 9.91 and well under Vanilla Bean's 15.07. The material test almost never says reject. Which means that if you have rejected a batch on cost grounds, the real reason was probably one of the next three sections — and it is worth knowing which, so you can fix the right thing.
Labour and remelt loss
These are the two lines that turn a clear material win into a marginal decision, and they are also the two lines nobody can supply for you. Everything in this section is either your own rate or your own measurement, and it is labelled as such.
Labour, first. The size of the job depends on whether the fault was caught in the pot or after pouring. Both cases are worth costing because the difference is large.
| Task | Caught in the pot | Caught after pouring | Why |
|---|---|---|---|
| Melt cycles | 2 h | 4 h | 32.4 kg is 3.24 fills of a 10 kg melter, so four cycles; a 20 kg melter halves it |
| Emptying 112 poured candles | — | Included in the melt time | Cured wax has to be dug or warmed out before it can be remelted |
| Cleaning 112 jars | — | Your own figure | Highly variable by jar shape; measure it once on ten jars and multiply |
| Re-pouring | 1 h for the 162-candle pour at 50 an hour | 3.25 h | 162 candles rather than 112, plus rewicking every jar |
| Weighing, records, batch coding | 1 h | 1 h | A corrected batch needs a new batch code and a new sheet |
| Assumed total | ≈3 hours | ≈8.25 hours | — |
| At an assumed ₹150 per hour | ₹450 | ₹1,237.50 | Substitute your own rate — this figure is illustrative |
Remelt loss, second — and this one you must measure. There is no honest published figure for how much wax a remelt costs you, because it depends on your melter, your pouring jug, your scrapers, how patient you are and how cold your workshop is. So do not take a percentage from anyone, including us. Weigh the wax going into the pot. Weigh what comes out as poured candles plus what is left in the jug. The difference, divided by what went in, is your loss for that batch on that equipment. Write it on the sheet and you will have a real number for every future decision.
| Measured loss | Wax and oil lost | Candles lost at 200 g | Value at ₹0.8422 per gram | Candles remaining | Material per candle |
|---|---|---|---|---|---|
| 2% | 648 g | 3.2 | ₹545.73 | 158 | ₹172.70 |
| 5% | 1,620 g | 8.1 | ₹1,364.32 | 153 | ₹177.30 |
| 8% | 2,592 g | 13.0 | ₹2,182.91 | 149 | ₹182.79 |
| 12% | 3,888 g | 19.4 | ₹3,274.37 | 142 | ₹191.40 |
Put the three lines together at an illustrative 5% loss and the correction totals ₹5,074 of wax + ₹1,237.50 of assumed labour + ₹1,364.32 of measured loss = ₹7,675.82, producing about 154 candles — roughly ₹49.84 of correction cost per recovered candle, on a batch whose material per candle finishes at ₹177.30 against ₹198.32 before. Against rejection at ₹22,212.32, the correction is still cheaper by a wide margin. Even at a 12% loss the sum comes to ₹9,585.87, which is 43% of what rejection destroys.
The 17-day clock
A corrected batch is an untested batch. The load changed, which changes the fuel, which changes how the wick behaves — so the burn test that qualified the 8% recipe does not qualify this batch, and the burn test that never happened on the 12% batch certainly does not. Before any of the 162 candles ships, two of them have to be cured and burned.
| Day | What happens | Why it cannot be compressed |
|---|---|---|
| 0 | Error found. Weigh everything, calculate the correction, order 10 kg of wax. | The arithmetic is quick; the wax has to arrive. |
| 1–2 | Remelt, add wax, add nothing else, stir, re-pour, rewick. | Four melt cycles at 10 kg capacity, plus the pour. |
| 2–16 | Cure. Lids on, same room, two sample candles marked. | The wax page asks for a minimum of 48 hours and ideally 1–2 weeks; the Sandalwood page states 48–72 hours for soy. On a corrected commercial batch, take the fortnight. |
| 16 | Burn both samples to a full melt pool, up to three hours each. Record pool time, flame, soot, throw. | A three-hour burn is the minimum that shows whether the pool reaches the glass. |
| 17 | Read the result. Release, re-wick and repeat, or divert. | — |
So the rule is arithmetic of a different kind: if the order ships in fewer than about 17 days, the correction is not available to you. It is not that correcting is unwise in that window; it is that what you would ship is untested stock, and that is not a saving — it is a deferred and larger cost. The honest options when the clock has run out are to pour the order from a recipe you have already proven and park the faulty batch for correction afterwards; to correct part of it and ship the proven part; or to reject.
What you must not trade for salvage
The question asks whether correction is "cheaper and safer", and the safety half deserves a plain answer. A dilution correction is, in itself, a conservative move: you are bringing a batch from outside a stated formulation range back inside it. That is the right direction. The failures happen at the edges of the job, and there are four lines this page will not cross.
| Shortcut | Why it fails | Do this instead |
|---|---|---|
| Push the wax hotter to speed up four melt cycles | You would be working past what the page states. CSI 464's page gives a melting point of 80–90°C; that is the figure to work to, not a floor to exceed because you are behind schedule. | Accept four cycles, or melt in two batches over two days. Time is the cost, not heat. |
| Sell the 12% batch as-is because "more fragrance is better" | At 12% the batch is above the load CSI 464's page states, which means it is outside the range the wax is formulated to hold. Excess oil does not bind; it migrates during cure and can drown the wick. | Correct it, or divert it. Do not ship a batch you cannot describe as within its wax's stated range. |
| Rewick the 112 cured candles in place and ship them | A wick set in cured wax cannot be replaced and vouched for. You cannot see the interior geometry afterwards. | Remelt and recover. The wax and oil come back; the wicks do not. |
| Ship the corrected batch without a burn test to make the date | The correction changed the fuel. The previous burn result belongs to a different formula. | Cure, burn two, read the result. If the date cannot accommodate that, the date decides, not the batch. |
One more honest note about what this page is not claiming. Correcting to 8% brings the batch inside the load the wax page states, and that is a formulation statement made by the wax's manufacturer — not a regulatory threshold, not a certification, and not a safety test result. What makes the corrected batch safe to sell is the burn test: a full melt pool reaching the glass in a sensible time, a steady flame of normal height, no soot and no mushrooming. Two candles, day 16. Everything else on this page is money.
Correct part of it, write off the rest
At commercial scale the answer is frequently neither "correct all of it" nor "reject all of it". Splitting the batch lets you buy the thing you are actually short of, which is usually time or demand rather than money.
| Split | Wax to add | Cost of that wax | Candles at 200 g | What happens to the remainder |
|---|---|---|---|---|
| Correct all 22.4 kg | 10,000 g | ₹5,074 | 162 | Nothing left over. Needs 162 jars and wicks and a fortnight. |
| Correct half (11.2 kg) | 5,000 g | ₹2,537 | 81 | 11.2 kg stays at 12%, labelled and parked for later or diverted to melts. |
| Correct a quarter (5.6 kg) | 2,500 g | ₹1,268.50 | 40 | 16.8 kg parked. Useful when the order is small and the deadline is tight. |
| Divert all of it to wax melts | None | ₹0 | — | 22.4 kg of melts at 12% of wax weight, inside the 10–15% melt range the Lavender page states and the up-to-12% the Sandalwood page states. No jars, no wicks. |
The half-correction is the workhorse. It halves the wax outlay to ₹2,537, halves the jars and wicks you need to find, and still produces 81 candles — and the parked 11.2 kg keeps its options, because scented wax at a known load and a known oil is a recoverable asset as long as the label survives. The quarter-correction exists for the case where the order is 40 candles and the rest of the batch is a problem for next month.
The melt route deserves its own mention because it removes the two lines that make a poured correction expensive. A melt needs no jar and no wick, and 12% of wax weight is inside the melt ranges those two oil pages state, so no dilution is needed at all — the "faulty" load is a perfectly ordinary melt load. If your batch was over-dosed and you sell melts, the cheapest correction on this page is to stop calling it a correction.
What the correction costs to buy
Three purchases: the wax that does the work, the equipment that decides your labour line, and the instrument that prevents the next one. Each card carries every listed size with the per-kilo or per-piece maths.
| Size | Price | Per kg | 200g candles @ 8% of wax wt |
|---|---|---|---|
| 500 g | ₹260 | ₹520 | ≈2 |
| 1 kg | ₹507.40 | ₹507.40 | ≈5 |
| 5 kg | ₹2,537 | ₹507.40 | ≈27 |
| 10 kg | ₹5,074 | ₹507.40 | ≈54 |
| Size | Price | Per piece |
|---|---|---|
| 10 kgs capacity | ₹21,240 | — |
| 20 kgs capacity | ₹25,960 | — |
| Size | Price | Per piece |
|---|---|---|
| Default Title | ₹354 | — |
CANDLEMAKINGSUPPLIESINDIA supplies raw materials and does not make or sell finished candles. We benefit from the answer this page gives — the correction is 10 kg of wax — so the honest thing is to be explicit about the three places where we tell you not to buy it: when the deadline is inside 17 days, when your order book cannot absorb the extra 50 candles, and when your wax is dear relative to your oil. The half-correction and the melt diversion both sell less wax than the full correction, and both are on this page because both are often right.
Every price, pack size and specification is from live CSI stock in September 2026 and links to the product page, which is always authoritative. Loads and temperatures are quoted only where the page states them: CSI 464 states a fragrance load of up to 10% and a melting point of 80–90°C; Luxury Soy Wax Chunks state 8–13% of total wax weight with 10% recommended, add 80–90°C, pour 70–80°C, and a cure of minimum 48 hours ideally 1–2 weeks before burn testing; Sandalwood Fragrance Oil states up to 10% for soy with a 48–72 hour cure and up to 6% for paraffin. Loads are wax formulation limits, not regulatory limits, and correcting to a stated load is not a safety certification — the burn test is what qualifies the batch.
The labour rate and the remelt-loss percentages on this page are assumptions, clearly labelled. We have not measured your workshop and we will not publish a loss figure as though we had. Weigh the wax in and the wax out, and use your own number. For bulk pricing above the listed packs, GST invoicing or documentation, message us on WhatsApp at +91 7397976926.
Frequently asked questions
- Testing finished samples from a 100-candle batch
- Recalculating 5 kg of scented wax at the wrong percentage
- Rewicking or remelting an entire production batch
- Can too much fragrance oil make a candle smell weaker?
- How to price candles in India 2026
- How much fragrance oil is needed for 500 Diwali candles?
- Scaling to 1,000 orders a month
Customer reviews: The six reviews at the top of this page are genuine, published Judge.me reviews left by CSI customers on their product pages. Names, star ratings, dates and products are as recorded by Judge.me; one longer review is shortened with an ellipsis and otherwise unedited.
Figures verified September 2026: Luxury Soy Wax CSI 464 500 g ₹260 (₹520/kg), 1 kg ₹507.40, 5 kg ₹2,537, 10 kg ₹5,074 — flat ₹507.40 per kilo from 1 kg to 10 kg. Luxury Soy Wax Chunks 1 kg ₹599, 10 kg ₹5,999. Premium Coconut Soy Wax 1 kg ₹920 (₹0.92/g). Sandalwood Fragrance Oil 500 gm ₹2,513.40, 1 kg ₹5,026.80 (₹5.0268/g). Citrus Lemon Fragrance Oil 1 kg ₹2,832 (₹2.832/g). Roasted Coffee Fragrance 1 kg ₹3,738. Lavender Fragrance Oil 1 kg ₹5,221. Vanilla Bean Fragrance Oil 1 kg ₹7,646. Santal Noir 1kg ₹9,120. Electric Wax Melter (Imported) 10 kgs capacity ₹21,240, 20 kgs capacity ₹25,960. Mini Electric Wax Melter ₹2,360. Candle Making Weighing Scale ₹354. Assumptions, clearly labelled as assumptions: the batch is 20,000 g of wax dosed with 2,400 g of oil, a 12% load expressed as a percentage of wax weight, against a target of 8% of wax weight; the fill is 200 g per candle; the labour estimate of about 3 hours in-pot and 8.25 hours after pouring, and the rate of ₹150 per hour, are illustrative and should be replaced with your own; the remelt-loss percentages of 2%, 5%, 8% and 12% are a sensitivity range, not measurements — measure your own by weighing wax in against wax out; jar cleaning time is not estimated because it varies by jar shape; jars, wicks, labels and packaging are excluded from every figure; a 14-day cure before burn testing is standard practice, and the Luxury Soy Wax Chunks page asks for a minimum of 48 hours and ideally 1–2 weeks. Arithmetic shown: F = 20,000 × 0.12 = 2,400 g; wax at target = 2,400 ÷ 0.08 = 30,000 g; added wax = 10,000 g = F × (1 ÷ 0.08 − 1 ÷ 0.12); added wax cost = 10,000 × ₹0.5074 = ₹5,074; material at risk = 20,000 × ₹0.5074 + 2,400 × ₹5.0268 = ₹10,148 + ₹12,064.32 = ₹22,212.32; ₹5,074 ÷ ₹22,212.32 = 22.8%, shown as 23%; corrected batch material ₹27,286.32 over 162 candles = ₹168.43 each, against ₹22,212.32 over 112 = ₹198.32; blended material rate = ₹27,286.32 ÷ 32,400 g = ₹0.8422 per gram; a 5% loss = 1,620 g = ₹1,364.32 and leaves 30,780 g = about 153.9 candles at ₹177.30 of material each; total correction at 5% = ₹5,074 + ₹1,237.50 + ₹1,364.32 = ₹7,675.82, about ₹49.84 per recovered candle; at 12% loss the total is ₹9,585.87, which is 43% of ₹22,212.32; threshold rule = correction wins on material while (1 ÷ L2 − 1 ÷ L1) < oil ₹/g ÷ wax ₹/g. Stated specifications quoted: Luxury Soy Wax CSI 464 — "Fragrance Load Up to 10%", "Melting Point 80°C – 90°C"; Luxury Soy Wax Chunks — 8–13% of total wax weight, 10% recommended, add 80–90°C, pour 70–80°C, cure minimum 48 hours ideally 1–2 weeks before burn testing; Sandalwood Fragrance Oil — up to 10% for soy with a 48–72 hour cure, up to 6% for paraffin, wax melts up to 12%; Lavender Fragrance Oil — candle load 6–10% and wax melt load 10–15% of total wax weight. Not claimed: no measured remelt loss, no labour figure for your workshop, no MOQ, no lead time, no price for a quantity above the listed packs, and no certification. Loads are formulation limits, not regulatory limits. Prices and availability change — the linked product pages are always authoritative.