How Do I Build a Commercial Candle Manufacturing System From Validated Formula to Bulk Procurement, Incoming QC, Batch Production, In-Process Checks, Finished QC, Traceability and a 10,000-Unit Monthly Scale?

1 · The validated formula, with its denominator written on it

A validated formula names products the way the invoice names them. Not "soy 464" but Luxury Soy Wax CSI 464 (Flakes) For Container Candles, 10 kg, ₹5,074.00. Not "cotton wick" but Eco Candle Wicks, Thick (C2), 1,000-pack at ₹9,440.00. Not "a white jar" but White Matte Glass Jar, pack of 10 at ₹708.00. CSI publishes no wick sizing data of any kind — only Thin (C1) and Thick (C2), with no jar-diameter recommendation, no burn-diameter figure and no chart anywhere — so the only authority for "which wick" is your own burn record. Name the variant you validated, or you cannot reorder it.

Then write the denominator — the costliest arithmetic error in the trade, because the catalogue uses both conventions. Luxury Soy Wax Chunks states its ceiling as "13% of total wax weight"; CSI 464 states "Fragrance Load Up to 10%" with no base named. On a 200 g candle those differ: 8% of finished weight is 16 g oil and 184 g wax, which is 8.70% of wax weight. Go to 10% of finished weight and you are at 20 g and 180 g — 11.1% of wax weight, above CSI 464's stated figure read that way. A sheet saying "10%" alone has lost the argument before batch one.

Two fields people leave blank. Cure: no cure time is published for any CSI wax, so cure is a constant you choose and hold identical — its job is to make batch 1 and batch 60 comparable, and changing it is a revalidation. Fill weight: most jar families publish no capacity. The Apothecary Glass Candle Jar with Dome Lid at ₹212.40 publishes 500 gm; the 200 ml Salsa Glass Jar publishes a volume, and a volume is not a weight. "A 200 g candle" is an assumption until you pour one and weigh it. Method: the repeatable mixing and curing protocol.

2 · Pilot before production — and the 20 kg wall that sets your batch size

Scale-up fails physically, not arithmetically — a larger mass cools slowly, the last jar pours cooler than the first, and a hand-stirred vessel mixes unevenly in a way a small pot hides — so the bench pilot proves the formula and the first full-size batch proves the process (the argument in full). Now the hard boundary. The largest melter CSI publishes is 20 kg / 20 litres: the Mini Electric Wax/Soap Melter, 1 litre, ₹2,360.00; the Electric Wax Melter, 20 Litres, ₹23,600.00; and the WAX MELTER (Imported) at 10 kg, ₹21,240.00, or 20 kg, ₹25,960.00. There is no 25, 40, 50 or 100 kg melter, no jacketed holding tank, no automated filler, no pump filler and no conveyor in the catalogue. A 50 kg day is therefore multiple cycles, or two melters in parallel, and the design question becomes cycle time, not machine size — see commercial candle making equipment in India and equipment needed to manufacture candles commercially.

That ceiling sets the batch. A 20 kg finished batch at 8% of finished weight is 18.4 kg of wax plus 1.6 kg of oil, and exactly 100 × 200 g candles — so 10,000 units is about a hundred batches, and 1,840 kg of raw wax over a 20 kg melter is 92 full melts. Neither figure is a throughput claim: CSI publishes no output, cycle-time or candles-per-hour figure for any melter, funnel or tool, so time three cycles yourself and fix whichever step everyone waits on. Trade-off: calculating the optimal candle batch size.

3 · Bulk procurement: the real pack arithmetic for 10,000 units a month

At 8% of finished weight, 10,000 × 200 g candles consume 1,840 kg of wax and 160 kg of fragrance oil before process loss; 5,000 units is 920 kg and 80 kg, a thousand is 184 kg and 16 kg. The published wax ladder is 500 g, 1 kg, 5 kg and 10 kg — no 25 kg sack and no 100 kg pack of any wax. The sole exception is Natural Soy Pearl Wax at 50 kg for ₹21,240.00 — ₹424.80/kg — of which 1,840 kg is thirty-seven packs, landing on 1,850 kg. Every other wax is 184 ten-kilogram packs.

Fragrance is tighter, and it is this series' hardest constraint. The ladder is at most 15 g, 50 g, 100 g, 500 g and 1 kg. There is no 2 kg, 5 kg or 10 kg fragrance pack anywhere on the store. So 160 kg is 160 one-kilogram bottles — not a drum, and not something that appears if you buy enough. The published price does not fall automatically above 1 kg, so ask for a quotation but never plan a month around a discount nobody has offered. And note what there is no point negotiating against: CSI publishes no MOQ, no quantity slab, no rate card, no pallet price, no annual volume agreement, no credit term and no lead time. Published price × quantity is the whole comparison.

Three facts there cut against our own interest. Luxury Soy Wax Chunks' 10 kg pack is ₹0.90/kg dearer than its 1 kg — ₹599.90 against ₹599.00 — which across 1,840 kg is ₹1,656.00 lost by buying the big bag. CSI 464 is flat at ₹507.40/kg from 1 kg up, so the 10 kg pack buys fewer parcels and nothing else; its 500 g is the outlier at ₹520.00/kg, so buy a part-kilogram tail as a 1 kg. Eco Soy CSI 400's ladder stops at 5 kg and its 5 kg is ₹0.20/kg dearer than its 1 kg — 1,840 single kilograms is ₹734,160.00 against ₹734,528.00 as 368 five-kilogram packs. Know it exists rather than believe "buy bigger".

Then the part-pack problem, which is pure traceability arithmetic. A 20 kg batch needs 1.6 kg of fragrance and the largest bottle is 1 kg, so every batch opens a second bottle and leaves 400 g in it. If part-bottles are never carried forward, 100 batches open 200 bottles against the 160 actually consumed — forty bottles of needless purchase, ₹113,280.00 on Citrus Lemon at ₹2,832.00. Per candle: a kilogram at 16 g a candle is 62 whole candles with 8 g left behind, so 160 bottles dosed in isolation yield 9,920 units. The fix is not a bigger bottle; there isn't one. It is a line on the batch sheet for the part-pack and its code.

Two gaps must be named rather than dropped. CSI sells no standalone lids, no retail or gift boxes, no shipping cartons, no printed labels, no inserts, no cushioning and no packing tape — the Candle Care Stickers Roll is the only published consumable of that sort, so cost those lines with your own printer first. Some jars ship with a lid, so never price a lid twice. And CSI publishes no lead time or dispatch time: confirm it on WhatsApp before building a schedule. Wider picture: the bulk raw-material package and choosing a wax supplier.

4 · Incoming QC when no listing carries a lot code and no spec sheet exists

Stated honestly: no CSI listing publishes a lot code, a manufacture date or an expiry date, and there is no certificate of analysis, specification sheet or batch-level test result for anything on the store. Three waxes publish a fragrance ceiling and two a shelf life; beyond that there is no number to receive goods against. So incoming QC is not "check it against the spec" — it is three records you create as the parcel is opened.

One: a receiving code you assign. Date, invoice, sequence — 2026-10-04/INV8821/W-03 for the third wax pack on that invoice — written on the pack in marker before it reaches the shelf, because the shelf is where identity gets lost; the convention does not matter, using the same one every time does. Two: a retained sample from every pack — a hundred grams of wax in a sealed coded jar, ten grams of oil in a coded bottle, ten wicks in a coded bag. If a batch misbehaves in February you need the material as it arrived in October, not a memory of it. Three: a burn check before the pack enters production, one candle to your locked formula against your approved reference.

That is 344 receiving codes and 344 retained samples a month — a lot, until you price the alternative. One unexplained batch of 100 candles is ₹9,336.16 of wax (18.4 kg of CSI 464 at ₹507.40/kg) plus ₹4,531.20 of fragrance (1.6 kg of Citrus Lemon), before jars and wicks. One lost batch pays for a year of notebooks.

5 & 6 · Batch production and in-process checks: one sheet, filled at the bench

A batch record is one page, filled in during the pour rather than from memory, and the full field list is in the table below: date, batch number, finished SKU, wax and fragrance products with their variants and receiving codes, dose in grams with the denominator used, dye product and dose, wick and jar products with variants, melt, fragrance-addition and pour temperatures, mix time, ambient temperature, units filled, fill-weight checks and the operator's name. The denominator is the field people drop and the one that costs money: write "1,600 g — 8% of finished weight (16 g in a 200 g candle)", never "8%". See converting a recipe into a commercial manufacturing SOP.

In-process checks are three habits. Fill by weight on a tared scale, not by eye to a line. There is no filler, pump or filling machine in the catalogue — the Wax Pouring Funnel with 4 nozzles at ₹1,888.00 is the only published filling aid — so accuracy lives entirely in the operator and the scale. Overfilling by 5 g on 10,000 units gives away 50 kg of wax a month; underfilling by 5 g is a complaint you cannot defend. Weigh jars 1, 25, 50, 75 and 100, spread across the pour, because drift is what you are looking for.

The third habit is pouring inside a window rather than at a number, because a 20 kg batch cools while you fill it — so record the first jar and the last. The citable figures are few, and they are instructions rather than specifications: CSI 464 pours at 80–90°C; CSI 468 pours at 58–60°C with a published melting point of 50–53°C; Eco Soy CSI 400 specifies fragrance addition at 85°C (185°F); and Luxury Soy Wax Chunks gives two pour ranges on one page, whose overlap is 75–80°C. CSI 468's two-degree window is worth noticing before you commit. Beyond those, no heat-stability figure, no softening point and no storage-temperature limit is published for any wax, so no page can tell you how your candle behaves in a warehouse in May — the wax selection page treats it the same way.

7 · Finished QC: burn a sample from every batch against your own approved reference

Finished QC on a hundred batches a month cannot mean burning everything, and it cannot mean trusting a specification, because none exists: CSI publishes no burn time, no melt-pool width, no flame height, no throw rating and no performance figure of any kind for any wick, wax or jar. What replaces it is a reference candle: at validation, set aside two or three from the approved batch, label them, store them unlit, and judge every later batch against those. It is the only standard available, and it costs three candles.

The per-batch method is fixed. Pull two candles — not the first two poured, which are least representative — and cure them with the batch. Burn one beside the reference, same room, started together, and record what you can compare: time to a full melt pool, pool depth, whether it reaches the wall, flame behaviour, sooting, the surface on resolidifying, and whether the room reads like the reference at thirty minutes and two hours. Keep the second unlit as the batch's retained sample.

Retained samples are units. Two per batch across a hundred batches is 200 candles a month you make and do not sell — 36.8 kg of wax and 3.2 kg of oil, or ₹18,672.32 at CSI 464's ₹507.40/kg plus ₹9,062.40 at Citrus Lemon's ₹2,832.00/kg. That is the price of being able to answer a complaint. Approval samples and replacements are units too: a 10,000-candle order is not a 10,000-candle production run, and a plan that treats it as one runs short every month.

8 · Traceability: the thread from receiving code to batch number to despatch

The thread has three joints and all must hold. Receiving code → batch sheet: every pack consumed is listed by its code, part-packs included. Batch number → despatch line: every order records which batch its units came from — the joint most makers never build, and without it the sheet is a diary. Retained sample → both ends: one per incoming pack, two per finished batch, labelled and stored so you can find them in four months.

What that buys is narrow and valuable. On a February complaint about a November candle, the despatch record names the order, the order names the batch, the sheet names five packs and the exact dose with its denominator, and the retained samples let you light the material as it was. So you can say something true — "made to formula, and here is the sample proving it burns like the reference", or "pack F-05 behaves differently, and here is the evidence". Both are defensible; "we think it was fine" is not. Note what you are not buying: traceability does not prove the material was good, because no incoming specification exists to prove it against. It proves what you did.

9 · Replenishment from measured consumption — and the honest verdict

By stage 9 replenishment needs nothing from anybody. A hundred batch sheets record the wax and oil that actually went in, loss included, so you know your consumption per unit — and you know the volume you have decided to produce. Multiply: at 200 g and 8% of finished weight that is 184 kg of wax and 16 kg of oil per thousand units, or 184 ten-kilogram packs and 160 one-kilogram bottles at 10,000. Then add whatever cover your own confirmed lead time requires: CSI publishes no lead time or dispatch time, so confirm it on WhatsApp at +91 7397976926. What this page will not do is tell you how much to hold, project demand, or call it "safe" to carry more stock. Threshold: when to stop buying soy wax like a hobby maker.

The one hard constraint on holding stock is shelf life, and it is thinner than makers assume. Exactly two waxes publish one: CSI 464 and CSI 468, both "60 months, cool dry storage". Every other wax publishes none and no fragrance oil publishes one at all. The only storage copy anywhere is qualitative — "Store in a cool, dry place. Avoid direct sunlight or moisture." on CSI 464, "Store in a cool, dry environment. Keep sealed when not in use." on CSI 468 — and since no listing carries a manufacture date, even those sixty months have no start point. Holding three months of material is a risk you size yourself, from your own receiving codes and first-in-first-out discipline.

So, the verdict. At 10,000 units a month three things bind on CSI's published catalogue. The 1 kg fragrance ceiling: 160 bottles, two opened per batch, and a part-pack discipline worth ₹113,280.00 a month on a cheap oil if you get it wrong. The 20 kg melting ceiling: no bigger vessel is published, so your month is about a hundred batches and your lever is cycle time, which only your own stopwatch measures. And the absence of published specification data: no lot code, no manufacture date, no expiry, no wick sizing, no heat-stability figure, no throughput, no mixing time for a production batch, and a shelf life on two waxes out of sixteen. None is solved by a supplier's assurance; all three are solved by the records in stages 4 to 8 — a notebook, a marker, three reference candles and ten minutes a delivery. And the most useful habit here: do not accept an unpublished figure from any supplier, this one included. If a number is not printed on a page you can check, treat it as a guess and measure it yourself — at a hundred batches a month your own measurement is the only authority still standing in February.

Frequently asked questions