How to Maintain Consistent Candle Quality as Production Grows

What actually drifts when the batch gets bigger

The first thing to be clear about is that growing output does not degrade candles through some general loss of care. It degrades them through a small number of specific, physical mechanisms, each of which can be named, measured and checked. If you cannot name the mechanism, you will end up changing your wax, which is expensive, disruptive and usually wrong.

Start with the one that catches almost everybody. A twenty-jar pour is over in a few minutes; a fifty-jar pour is not. The wax in the pitcher loses heat the whole time, so the last jars are poured cooler than the first. The Luxury Soy Wax Chunks page gives a pour temperature of 75–80 °C, and a band five degrees wide sounds generous until you realise a long pour can walk out of the bottom of it without any single decision being made. The candles at the start of the batch are fine. The candles at the end have rougher tops and worse adhesion, and because they are the ones still on the shelf when the good ones have sold, they are over-represented in complaints.

The second drift is weighing. At twenty candles a week you weigh the fragrance because it feels like part of the craft. At five hundred a month you are doing it ten or thirteen times, often while three other things are happening, and the bottle starts getting poured by eye. Fragrance load in this series always means a percentage of wax weight, which is how the product pages state their ceilings, and an eyeballed pour is exactly how a batch intended at 8% ends up at 9.5%. The Chunks page states up to 13% of total wax weight with thorough test burning at that maximum, so a drift like that is not necessarily a disaster — but it is a change you did not choose, and it will make one batch smell different from the last.

The third is cure. It is the easiest thing in a candle business to give away, because nothing visible happens when you take it. A candle poured on Monday and shipped on Thursday looks identical to one shipped a fortnight later. The Chunks page itself says cure a minimum of 48 hours and ideally one to two weeks before burn testing, and standard practice for soy container candles is 14 to 21 days by fragrance family. Under a festive deadline, a batch that should have sat for two weeks sits for four days, and the customer who smelled your stall sample gets a candle that has not finished developing.

The fourth is wick centring, which suffers purely from speed. Forty jars in a row invites approximation, and an off-centre wick burns towards one wall, puts heat against the glass and tunnels on the opposite side. It is the cheapest fault to prevent — ten seconds looking down the row from one end before the wax sets — and one of the most expensive to have shipped.

The five checks, and what each one costs

Four of the five checks are free, in the sense that they consume no materials and need no equipment you do not own. The only one that costs money is the burn test, because a burned candle cannot be sold. A Pen Thermometer at ₹354.00 is the one instrument the whole system depends on, and a second one at ₹708.00 for the pair is a sensible redundancy once two people are pouring — a thermometer that has been dropped reads confidently and wrongly, which is worse than no reading at all.

Check one: two weights per pour

Weigh the first jar and the last jar of every pour and write both on the sheet. Everything in between is assumed to lie along the line joining them. More than about 5 g of drift across a batch at a 150 g fill means the pour is getting away from you — usually because the pitcher is emptying and the last jars are being over-filled to use it up. The two numbers take fifteen seconds and they are the single best early-warning indicator in a candle workshop.

Check two: two temperatures per pour

One reading at the first jar and one halfway through. Two readings turn a single data point into a rate, and a rate is what tells you whether a longer batch will finish inside the band. This is also the check that decides whether you need more melting capacity: if the halfway reading is consistently at the bottom of your band, the answer may be a Mini Electric Wax Melter at ₹2,360.00 holding the wax at temperature while you work, or a larger unit such as the imported 10 kg melter at ₹21,240.00. Be honest about what equipment does, though: a melter holds temperature, which removes one drift. It does not centre a wick, weigh an oil or wait out a cure.

Check three: a visual pass on every candle at 24 hours

Not a sample — every candle, once, the day after the pour, in daylight. Surface, wick position, adhesion, fill level, vessel condition. It costs nothing but attention, it is the only check that sees 100% of output, and it is where the 3% visual reject assumption in this page's budget comes from. Record the reject count and the reason on the batch sheet, because a reject count with no reason attached teaches you nothing next month.

Check four: a sampled burn test

The only check that consumes stock. Burn a sample of every batch after the cure, in a room of a normal size, and watch the melt pool at two hours, the throw, the flame height, sooting and mushrooming. Sample rates are in the next section. Always burn two, not one, whenever a wax lot changes, an oil lot changes or a wick size changes — one candle cannot distinguish a lot problem from an unlucky candle. At 500 a month this is 20 candles and ₹4,636.86 of material.

Check five: the pre-dispatch gate

The last look, done as the candle is packed rather than as it is made. It catches damage that happened after the visual check — a chipped lid, a smudged label, a candle that sat in a sunny window for a week. It is also the only check that happens with the customer's expectation in mind rather than the maker's.

How many candles to burn, and what that costs

Burn testing is the only quality check that destroys the thing it checks, which makes it the only one where the rate is a real decision. The rate should fall as volume rises — not because bigger businesses need less information, but because a fixed number of tests per batch gives you more information per candle once batches are more alike.

Three rules override the table. Never fewer than one candle per batch, whatever the rate says, because a batch with no burn test is a batch you are shipping on faith. Always two whenever a lot or a wick changes. And always burn from the end of the pour rather than the start, because the end of the pour is where the drift lives — testing jar 3 of 50 tells you about the easiest candle in the batch.

There is a genuine tension here that is worth naming. A burned candle is ₹231.84 of material and also a candle you could have sold, so the true cost to the business is higher than the material figure and depends on your own pricing, which is not something this page will put a number on. What can be said is that the alternative is finding out from a customer, and a customer who finds out for you costs a replacement, the postage on it, and the review.

The pre-dispatch gate, in four steps

The gate is deliberately boring and deliberately last. It happens at the packing table, one candle at a time, and it is the only point in the process where somebody looks at a candle as a customer would rather than as its maker.

One of those rows deserves a word about materials rather than process. Vessels arrive in cartons and cartons genuinely vary — one of the reviews on the CSI store says of a tin container that the material is good but some arrived a little dented and the lid is hard to open once a candle is finished in it. That is fair, useful feedback, and the production answer is the same for any supplier: check one vessel from each carton before it reaches the bench, because a dent found on the shelf is a supplier conversation and a dent found after filling is your loss.

The consistency budget at 120 and at 500 a month

Quality systems are usually argued for in the abstract and abandoned in the specific, because nobody puts a number on them in advance. So here is the number, built from the assumptions used throughout this page — a 3% visual reject rate, a 1% gate reject rate, one retained candle per batch, and the sampled burn-test rates from the table above. All four are stated assumptions, not measured data, and your own figures will differ.

Read the last two rows together, because they contain the only genuinely counter-intuitive thing on this page. The budget in rupees goes up as you grow — ₹4,405.02 becomes ₹11,592.15 — but as a share of output it falls from nearly 16% to 10%. Quality discipline gets cheaper per candle the more candles you make, which is precisely backwards from how most small makers think about it. The maker at 120 a month who says testing is a luxury they cannot afford is right that it costs them proportionally more, and wrong that the answer is to skip it, because the candles they are shipping untested are the ones building the reputation that has to carry them to 500.

Set against the budget: an unchecked month at 500 candles puts ₹1,15,921.50 of materials into the market with no evidence behind it. The ₹11,592.15 budget is exactly 10% of that — a tenth of the month's materials spent making sure the other nine tenths are what you think they are. Whether that is a good trade depends on facts nobody can supply for you, including your price, your repeat rate and how much a bad review costs you, none of which CSI can put a figure on.

When a second pair of hands arrives

A single maker carries most of a process in their hands. They know that the wax is ready when it looks a certain way, that this jar takes a little more, that you wait a beat before setting the wick. None of that survives a handover, and all of it was doing real work. This is why quality usually dips at the exact moment a business gains capacity, which feels like bad luck and is entirely predictable.

The fix is to spend one batch badly. Pour alongside the new person with both of you filling in a sheet independently, then compare the two sheets line by line. Every disagreement is something you knew and never said: the temperature you actually wait for, how long two minutes of stirring really is, how full is full. Fix those in the written formula rather than in a verbal correction, because a verbal correction lasts until the next busy afternoon.

Then split responsibility rather than duplicating it. One person owns the weigh-out and the sheet; the other owns the pour and the wicks. Both sign. Shared responsibility for a check means nobody does it, and this is true in a two-person workshop exactly as it is in a factory.

Holding the line through a festive ramp

Diwali 2026 is 8 November 2026. With a 14–21 day cure plus shipping, festive candles want to be poured by about mid-October, which loads September and early October with two or three times a normal month's pours. Every drift in this page is a drift that gets worse under time pressure, and the checks are exactly what a rushed workshop drops first.

Decide in advance which checks are non-negotiable. The two weights and two temperatures stay, because they cost seconds. The 100% visual pass stays, because it is the only thing standing between a rushed pour and a customer. The gate stays. What can flex is the burn-test rate — dropping from 1 in 12 to 1 in 20 during a ramp is defensible if, and only if, you are still burning two candles on every lot change. What cannot flex is cure, and the honest way to protect it is to pour earlier rather than to release sooner. At a 500-candle festive month that is ₹1,15,921.50 of material committed several weeks before the revenue arrives, which is a cash decision as much as a quality one.

The other festive-specific risk is heat, not haste. A candle cured in a workshop at 34 °C and then loaded into a delivery van in Ahmedabad is a different object from the same candle in February. Store finished stock away from sunlight and away from anything warm, and if a batch has been sitting in a hot room, put it through the visual check again before it is packed rather than trusting the check it passed a fortnight ago.

None of this needs new materials, and that is the point worth ending on. The checks in this page are a thermometer, a scale you already own, a reference candle you already made, and a decision about how many candles you are willing not to sell. If your candles have changed and the register says the wax lot has not, the answer is in the pour, not in the sack. For bulk pricing beyond the listed pack ladder, GST invoicing, or MSDS and IFRA documentation to file with your quality records, message WhatsApp +91 7397976926.

Frequently asked questions