Three Customers Say My Candle Glass Becomes Extremely Hot Halfway Down — Should I Stop Selling the Batch and Repeat Full-Life Burn Testing?
Aktie
Clear 150 g jar with golden lid ₹944 / pack of 10 · CSI 464 from ₹260 / 500gEco Candle Wicks from ₹59Made & stocked in India · ships nationwide
CSI complaint-handling guides
Three customers have told you something your test data cannot answer. That is the finding — and the decision that follows is yours, not ours.
★★★★★
"The jars were of good quality"
Soumili GopeVerified buyer · Mar 2026
Clear Glass Jar with Golden Lid
★★★★★
"One of the best wax for container Candles. Easy to work with, especially if you are a beginner."
ThenirvaofficialVerified buyer · Nov 2025
Luxury Soy Wax CSI 464
★★★★
"Quality of the wick is good for jar candles, team can work on the pricing and delivery timelines."
Monika DeepVerified buyer · Oct 2025
Eco Candle Wicks
★★★★
"I like the material of tin jar no doubt but some are little dent but the Mandala design is awesome…"
Jinal MonilVerified buyer · Mar 2026
Tin Containers 85 & 135g
★★★★★
"Nice glass jars, they look cute. I like the packaging as well."
AshwiniSep 2025
White Matte Glass Jar
★★★★★
"Good product"
Biny AbrahamVerified buyer · Feb 2026
Black Glass Jar with Black Lid
★★★★★
"The jars were of good quality"
Soumili GopeVerified buyer · Mar 2026
Clear Glass Jar with Golden Lid
★★★★★
"One of the best wax for container Candles. Easy to work with, especially if you are a beginner."
ThenirvaofficialVerified buyer · Nov 2025
Luxury Soy Wax CSI 464
★★★★
"Quality of the wick is good for jar candles, team can work on the pricing and delivery timelines."
Monika DeepVerified buyer · Oct 2025
Eco Candle Wicks
★★★★
"I like the material of tin jar no doubt but some are little dent but the Mandala design is awesome…"
Jinal MonilVerified buyer · Mar 2026
Tin Containers 85 & 135g
★★★★★
"Nice glass jars, they look cute. I like the packaging as well."
AshwiniSep 2025
White Matte Glass Jar
★★★★★
"Good product"
Biny AbrahamVerified buyer · Feb 2026
Black Glass Jar with Black Lid
Reviews collected and published through Judge.me on candlemakingsuppliesindia.store. Products are as recorded by Judge.me: Clear Glass Jar with Golden Lid, Luxury Soy Wax CSI 464, Eco Candle Wicks, Tin Containers 85 & 135 g, White Matte Glass Jar and Black Glass Jar with Black Lid. One longer review is shortened with an ellipsis; wording is otherwise unedited. None of these reviews describes container temperature, a full-life burn, or a safety concern of any kind — they are ordinary product-quality reviews and nothing on this page should be read as customer evidence about heat.
✓ Published fragrance ceilings per wax, where the wax maker states one✓ 92 fragrance oils in stock✓ GST invoicing · MSDS & IFRA on request
Customer Complaint Guides · Container Heat
Your customers have found the exact place your testing stopped, so you cannot settle this from your development records. Here is the protocol that generates the missing data, the variables worth checking, and what each option costs in real materials. The decision to hold or ship is yours, and so is the liability.
4%
of a 300-unit batch's material value buys the full-life burn record you do not have — ₹3,037.92 for a twelve-unit comparison against ₹75,948.00 of materials sitting in the batch · CSI live pricing, September 2026
Quick answers — read this first
Should I stop selling the batch? Not a question this page can answer, and not one any supplier should answer for you. It is a commercial and legal judgement belonging to you as the manufacturer of the finished candle — where the stakes are serious, take qualified professional advice rather than a blog's. What we can give you is the framework, the arithmetic and the protocol.
Is the complaint useful? Very. "Halfway down" is precisely the part of a candle's life that early-burn testing never reaches.
What is the first free step? Check the recorded formula against the fragrance ceiling published for your wax — 13% for Luxury Soy Wax Chunks, 10% for CSI 464, 13% for Coconut Soy CSI 468, with no ceiling published on the other thirteen waxes — then the add temperature against the listing.
What is the deliverable? A full-life burn record on retained units, run alongside matched units from the previous batch. Twelve units cost ₹3,037.92 in CSI materials.
The short answer
Credit the complaint: three independent reports about the same stage of the burn is a signal, and the stage they name is the one your development testing skipped.
Then gather, in this order: report rate against units shipped; whether the three cluster on a shared wax lot, wick lot, jar shipment, pour session or oil; the recorded formula against the published ceiling for your wax; then a full-life burn on retained samples.
Then decide — you, not us. We will not tell you the batch is fine, because no published figure exists that would let anyone check such a claim. Replace the three customers' candles regardless of cause, as goodwill and not as an admission.
A candle's life in thirds, with the observation points a full-life test would record. Workshop testing almost always ends inside the first third: the pour looks right, the first burn opens a full pool, the unit goes into production. The complaint sits in the last third. No product is named, no temperature plotted and no property claimed — the point of the diagram is the absence of data, not its content.
Three customers report the glass becoming extremely hot halfway down. Should the batch be pulled, and should full-life burn testing be repeated?
Repeat the testing: yes, and on a full-life basis — the instinct is correct and the reason is structural. Early-burn tests cover the first third of a candle's life; the complaints are about the last third, so you hold no data on the period being described. Whether to stop selling is a different question, and it is yours. You are the manufacturer of the finished candle, so the commercial exposure and the legal responsibility sit with you, and no raw-material supplier — CSI included — can carry that judgement for you. What this page supplies is the arithmetic to frame it: the report rate against units shipped, the clustering checks that would point at one wax, wick or container lot, the free formula checks against the published fragrance ceilings, a full-life burn protocol on retained units, and what holding, re-pouring and testing each cost. Where the stakes warrant it, take qualified advice — and replace the three customers' candles now, as goodwill rather than as an admission.
One line: the complaint has located your data gap precisely — generate the missing full-life record on retained units, and make the hold-or-ship call yourself with the numbers in front of you.
A full-life test needs two instruments and a habit. Pen Thermometer ₹354 and Candle Making Weighing Scale ₹354 — ₹708 for the pair — plus retaining a fixed number of units from every batch from now on.
Three people, independently, have named the same stage of the burn — and in doing so have located the weakness in almost every small maker's testing regime, yours and your competitors' alike.
Workshop testing has a shape and it is always the same: pour a trial batch, cure it, light one, check the pool reaches the wall, look for soot, note the finish, and into production it goes. Almost nobody burns a trial unit to the bottom before committing a batch, because it takes days of attendance and blocks the bench. The record you hold covers the beginning of the candle's life; the record you do not hold covers the end of it.
The end is genuinely different, by ordinary physics. As the wax column shortens the flame sits lower, so the part of the vessel nearest it moves from the upper wall towards the base; at the same time the remaining wax and glass have less mass to absorb and spread the heat produced, and less unmelted wax below the pool to insulate. A candle can behave one way in its first third and differently in its last, both consistent with the same recipe.
This is a gap, not a failure of care — and the corollary matters most. You cannot resolve it from your existing development data, because the relevant observations were never made. New data is the only route, which is why the title's instinct is sound, and the test is already defined by the complaint: burn units of this batch from full to finished, observing at intervals, against a comparison set from a batch shipped without reports. Related: how to verify wax quality before you buy in India.
What this page will not do
It states no temperature at which glass is safe or unsafe to handle, no maximum burn time, and no heat-tolerance or thermal-shock figure for any CSI container — none is published, and a claim of that sort would be a safety verdict on your product. For your supply file, CSI provides GST invoicing, MSDS and IFRA documentation on request on WhatsApp at +91 7397976926.
The hold-or-ship framework — yours to apply
What follows is a list of things to weigh, and it deliberately does not end in a recommendation. We are not going to conclude for you, and be suspicious of anyone who does — a supplier who has never seen your batch, your jar carton or your customers cannot responsibly tell you whether to keep selling. Start with the rate, because three is a count, not a rate.
Step one · three reports expressed as a rate
The same count against different shipped volumes
Units of this batch shipped
Three reports as a rate
What it does to the decision
40
7.5%
High enough that a systematic cause is the obvious hypothesis, not the unlikely one
120
2.5%
Well above a stray-complaint background; clustering checks become the priority
300
1.0%
Ambiguous alone. The clustering pattern and the retained-sample burn carry the weight
1,200
0.25%
Low, but low is not zero — and a heat report is not a scent report
One caveat: the observed rate is a floor, not the true rate — only some proportion of customers who notice something write in, and nobody publishes a figure for that proportion, so do not multiply your three by an invented number. Then look for clustering, because a systematic cause usually leaves a fingerprint. If the three share an input, that is where the investigation goes; if they share nothing, that strengthens either a design-level cause affecting every unit of the recipe or a usage pattern you have not characterised.
Step two · the clustering checks
What to look up, where it lives, and what a match would mean
Check
Where the record should be
What a shared value would point at
Wax lot
Supplier invoice date and the bag the wax came out of
The wax supply for that period rather than the recipe
Wick lot
Wick pack and the invoice it arrived on
The wick supply — worth checking physically against retained wicks
Jar shipment
Carton, invoice and any date code on the glass
The container — the variable most often invisible because the jars look identical
Pour session
Batch sheet: date, time, ambient temperature, who poured
Process on one day — a mis-weighing, a temperature drift, a substituted step
Oil and recorded load
Invoice, bottle and the load in grams as weighed
Load or a specific oil — the cheapest thing on this list to rule out
Date shipped and channel
Order records
Transit or storage conditions rather than manufacture
Then weigh the category of complaint, and be honest that this part is not arithmetic. A report that a candle smells weaker than expected has a bounded downside; a report about a container becoming unexpectedly hot does not, which is a legitimate reason to act on weaker evidence. A cautious maker may reasonably hold stock while investigating even on thin evidence, and that is defensible rather than an overreaction — but it is a judgement, made by you, not the output of a formula, and this page will not pretend otherwise.
The CSI principle
A complaint is data. A decision is yours. Nothing a supplier publishes converts the first into the second.
You are the manufacturer of record. The framework, the protocol and the costs here are ours to supply; the call to hold, ship, re-pour or recall — and the liability attached to it — are yours. Where the stakes justify it, take qualified professional advice, legal or technical, rather than relying on any blog including this one.
Two default answers to refuse, both tempting. The first is "the customer burned it wrong" — do not start there. It is the cheapest explanation, it stops the investigation before it begins, and it costs you most if it is wrong. The second is "we have a defect, pull everything and apologise", which concedes what you have not established and creates a written admission that may be just as inaccurate. The honest position right now is "not yet known", and that is a professional thing to say, alongside a replacement and a note that you are investigating.
Do this today, regardless of cause
Replace the three customers' candles in full, without arguing about cause. It is right while the answer is unknown, it costs a fraction of what the argument costs, and it is goodwill rather than an admission — say exactly that. Ask for the batch number in the same message.
The full-life burn protocol — the thing you should already have had
This is the page's real deliverable: not complicated or expensive, only slow, which is why it gets skipped.
1
Assemble two matched sets of sixSix retained units from the suspect batch, and six from the last batch you shipped without a heat report — same recipe, jar, wick and load, different batch. The comparison set is what makes the test interpretable, because no published figure exists to compare against; your only reference is your own earlier product. If you hold nothing from the earlier batch, pour six comparison units to the recorded recipe and note it. If you hold no retained samples at all, stop here: that gap is now the first thing to fix, permanently — a shelf, a marker pen, and a rule that a fixed number of units per batch is never sold.
2
Fix the session length and keep it identical across all twelveThree-hour sessions are a common choice because they make the arithmetic tidy, not because three hours is any kind of limit — the figure you pick is a test parameter, and this page states no maximum burn time. The rule that matters: the session length you test must be the one your care card tells customers to use. Record ambient temperature at every session start, and stand every jar on the same surface.
3
Record at defined intervals, not only at the startMeasure remaining wax depth with a steel rule before and after every session, and take a full observation set at six points: full, three-quarters depth remaining, half, one-third, one-quarter, and the final session. Record four things at each: the melt pool — reached the wall or not, depth, centring; the wax depth in millimetres; the flame — height against session one, steadiness, leaning, mushrooming; and the container, in your own words on a scale written down before you start. Trim between sessions with a wick trimmer at ₹177 and record it.
4
Describe the container against your own reference, because no published limit existsNo figure exists anywhere in the catalogue for how warm any jar becomes or is meant to become, so your record is comparative and descriptive. Write the scale down first: four or five levels in your own words, from the coolest you would expect to the point at which you would put the jar down. Apply it identically to a suspect unit and a comparison unit burned side by side, same room, same session. The test can tell you whether this batch differs from the batch you were shipping without complaint. It cannot tell you whether any level on your scale is acceptable — that is yours to determine and, where it matters, your adviser's.
5
Photograph the same three frames at every observation pointOverhead with the flame lit, thirty minutes in: pool width, centring, carbon on the wick. Side on with a rule against the glass so remaining depth is readable in frame. The wick tip before you trim it. Pay closest attention from the halfway point onwards — that is where your customers are pointing. A pen thermometer and weighing scale at ₹354 each belong on the bench for the pour record; the burn record needs a rule, a notebook and a phone.
Batch traceability, which is what makes any of this possible
A batch number is only useful if it connects to something: the wax, oil, wick and jar products with their invoice dates and cartons; the load in grams as weighed; the pour date, time, ambient temperature and who poured; the measured fill weight; the cure period; and the retained-sample shelf location. Without those a complaint is uninvestigable. Print the number on your labels — a Candle Care Stickers Roll is ₹413 per 500 at 1 inch (₹0.83 each) or ₹699 per 1,000 (₹0.70 each).
Five candidates, explained mechanically — checked one at a time
None of these is an accusation against any product, CSI's or anyone's; they are the places in a container-candle design where the late burn can differ from the early one.
One. The wick's fuel demand against what the design can dissipate late in life. Early on the heat is produced high in a tall column with a large mass of wax and glass around and beneath it; late on, the same consumption happens in a short column with little mass left below. A wick can carry a design comfortably for two-thirds of its life and be working at the edge of it in the final third. Note what cannot be said: Eco Candle Wicks ship as Thin (C1) and Thick (C2) — the listing's own words — and CSI publishes no jar-diameter recommendation, no burn diameter, no melt-pool width, no flame height and no burn rate for any wick. See also the wick guide by diameter.
Two. A jar shipment that differs from the one you approved, at identical external dimensions. This variable hides best, because the jars look the same on the shelf and in the photograph. Glass can vary in wall thickness and base geometry between production runs while the outside profile stays in tolerance, and those are what govern how heat leaves a burning container. The check takes ten minutes: weigh ten empty jars from the suspect carton and ten from a retained carton of the approved shipment, and record mean and spread. Then pour and weigh the fill, because most container families publish no fill capacity; where a listing names one — the Clear Glass Jar with Golden Lid is listed as 150 g — verify it on your scale anyway. And mind millilitres: the 200 ml Salsa Glass Jar is described by volume, and 200 ml is a volume, not a 200 g fill.
Three. Fragrance and dye are fuel, and they are not wax. Everything in the pour that is not wax burns differently from wax, and the proportion of non-wax fuel matters more as the column shortens. This is the only variable here with a published limit, so it is the first to check. The published ceilings: Luxury Soy Wax Chunks up to 13% of total wax weight; CSI 464 up to 10%; Coconut Soy CSI 468 up to 13%. The other thirteen waxes publish none, and these are formulation limits set by the wax maker — not regulatory limits. Check your recorded grams against the ceiling for the wax you used, then the temperature against the listing: CSI 464 states pour at 80–90°C, CSI 468 pour at 58–60°C, Luxury Soy Chunks fragrance addition at 80–90°C with two minutes of stirring. More: premium versus cheap fragrance oils.
The arithmetic of a load error, which is easy to make
The ceilings are worded differently, so read your numbers back carefully. The Luxury Soy Chunks listing states its 13% of total wax weight, so 1 kg of that wax takes up to 130 g of oil. The CSI 464 listing says "Fragrance Load Up to 10%" and names no base, so the conservative reading is 10% of the finished candle: a 150 g unit is 135 g wax and 15 g oil. Weigh 15 g into 150 g of wax instead and your unit is 165 g at 9.09% — auditable only if you wrote both figures down at the bench.
Four. A wick lot that differs from the one you approved. Wick is a manufactured textile and lots can differ; a difference will show where the design has least margin, which is late in the burn. Compare retained wicks against the suspect pack physically, and if you hold both, burn two otherwise identical units, one on each. Five. The wax lot. Same logic, same check, invoice date as lot marker. Note how the wax was stored between arrival and pour: CSI 464 and CSI 468 each specify cool, dry storage and a 60-month shelf life.
One variable at a time, and this is not a stylistic preference. Change the wick and the load in the same corrective batch and you learn nothing you can rely on: you cannot tell which change did it, nor whether one helped while the other hurt, and you have destroyed the evidence that would have told you. Worse, you now have a second recipe in the field whose late-burn behaviour is also untested. Change one input, run the protocol, record it, then decide whether to keep the change.
Need your paperwork in order for the investigation file? GST invoices, MSDS and IFRA documentation for anything you have bought from CSI are supplied on request, and it is worth collecting them into the batch file now rather than when somebody asks.
Your care card almost certainly says something about how long to burn the candle. The question worth asking is where that sentence came from.
Most small makers copy burn-duration wording from another brand's label, a forum, or whichever supplier's blog they read first. Ask whether the figure on your card reflects anything you actually tested — this recipe, this jar, this wick. If not, it is not an instruction; it is decoration handed over as guidance. Part of the point of the protocol above is to make that sentence true.
Now the harder half, directly. Instructions cannot be used to move a design problem onto the customer. If a candle behaves in a way you would not accept when burned exactly as your own label tells people to burn it, tightening the label is not a fix — it is a relabelling of the same product. The reverse is equally true and gets forgotten: a candle burned for many hours continuously is outside the tested envelope of almost every small maker, including, before this week, yours. That is not the customer being unreasonable; it is you not knowing where your envelope is.
A care card is a record of what you tested. If you did not test it, you did not write an instruction — you wrote a wish.
— CandleMakingSuppliesIndia
Practically: hold the wording in draft until the protocol is finished, then set it to match what you observed and are willing to warrant. Ask the three customers what their session lengths actually were, without a hint of accusation — you are gathering a variable, not building a defence. If they were far longer than anything you have tested, the answer is more testing, not a sterner label. Reprinting is never the reason to leave a label wrong: care stickers are ₹413 per 500, ₹0.83 a unit.
What each option costs, in materials you can price
Numbers make the decision cleaner even though they do not make it for you. Everything below is built on one reference unit at CSI live pricing, September 2026: a 150 g finished candle in a Clear Glass Jar with Golden Lid (150 g), poured in CSI 464 at that wax's published 10% ceiling — 135 g wax, 15 g oil — scented with Lavender, on a Thick (C2) Eco Candle Wick with a sticker. Substitute your own figures; the method is the point.
The reference unit · 150 g finished candle
Material cost per unit at best per-gram pricing, and across a twelve-unit test
Component
Rate used
Per unit
× 12 units
CSI 464 wax, 135 g
₹5,074 / 10 kg = ₹0.5074/g
₹68.50
₹822.00
Lavender oil, 15 g
₹5,221 / 1 kg = ₹5.221/g
₹78.32
₹939.84
Clear 150 g jar with golden lid
₹944 / pack of 10
₹94.40
₹1,132.80
Eco Candle Wick, Thick (C2)
₹944 / 100 wicks
₹9.44
₹113.28
Wick sticker
₹100 / sheet of 40
₹2.50
₹30.00
Total materials
—
₹253.16
₹3,037.92
Your three options, priced · 300-unit batch
₹253.16 per unit, so ₹75,948.00 of materials sits in the batch
Option
Cost in CSI materials
Arithmetic
Run the full-life comparison on twelve retained units
₹3,037.92
12 × ₹253.16 — 4% of the batch's material value
Hold all 300 units while you investigate
₹75,948.00 of finished stock frozen
300 × ₹253.16, plus your own retail margin deferred
Re-pour all 300 with new jars
₹75,948.00
300 × ₹253.16 — the batch cost paid twice
Re-pour 300 reusing cleaned jars
₹47,628.00 plus cleaning labour
300 × ₹158.76 (₹253.16 − ₹94.40 jar)
Keep shipping and change nothing
₹0 today
The exposure is unpriced, and it is yours
Read those five rows together and the shape of the decision becomes clear even though the decision itself does not. The test costs 4% of what the batch is made of — one twenty-fifth of a re-pour, one twenty-fifth of what is frozen if you hold. That is why the second half of the title's question is easy to answer yes to, and the first half is neither cheap nor answerable from here. Add ₹708 for a pen thermometer and weighing scale, plus ₹118 for five stirring sticks. See also how to price candles in India 2026.
1
Wax · the only variable with a published ceiling
Luxury Soy Wax CSI 464 (Flakes)
The reference wax, because it is one of only three publishing a fragrance ceiling — up to 10% — which makes a load audit possible rather than guesswork. The listing also states pour at 80–90°C, cool dry storage away from sunlight and moisture, and a 60-month shelf life. Quote those as the listing's own statements. No heat-tolerance or softening figure is published for this or any other wax on the store.
Size
Price
Per gram
135 g units
500 g
₹260.00
₹0.52
≈3
1 kg
₹507.40
₹0.5074
≈7
5 kg
₹2,537.00
₹0.5074
≈37
10 kg
₹5,074.00
₹0.5074
≈74
₹68.50 of wax per 135 g unit at 10 kg pricingBuy CSI 464
2
Wick · Thin (C1) and Thick (C2), the listing's own words
Eco Candle Wicks
Two thicknesses, and that is the entirety of what is published. No sizing chart, no jar-diameter recommendation and no burn-rate figure exists for either option, so which belongs in your jar is something only your burn record establishes. Keep a retained sample from every wick pack, marked with the invoice date. The Thick (C2) 10-wick pack is out of stock; the 20-wick pack and above are the buyable Thick variants.
At 15 g per 150 g unit — the 10% ceiling published for CSI 464 — lavender is ₹78.32 of the ₹253.16, the largest single line in the reference unit and the one worth auditing in grams rather than remembered percentages. If your load sits above the ceiling published for your wax, that is a finding you can act on today. A cheaper comparison: Citrus Lemon at ₹2,832 per 1 kg is ₹2.832 a gram, so 15 g is ₹42.48.
Size
Price
Per gram
150 g units @ 10%
15 g
₹105.02
₹7.00
1
50 g
₹270.00
₹5.40
≈3
100 g
₹540.00
₹5.40
≈6
500 g
₹2,620.00
₹5.24
≈33
1 kg
₹5,221.00
₹5.221
≈66
₹78.32 of fragrance per 150 g unit at 1 kg pricingBuy Lavender
Where we argue against our own interest, and you should hold us to it. The commercially convenient thing for a supplier to say here is that your materials are fine and the batch is fine, because that keeps you pouring and reordering. We are not going to say it: CSI has published no glass-temperature figure, no heat-tolerance rating and no thermal-shock specification for any container, so no number exists on which such a reassurance could be checked — by you, by us or by anyone. Apply that standard to every supplier including this one: a heat claim that cannot be traced to a published figure is an opinion. The honest offer is documentation, invoices and a method — not a verdict.
Why trust this guide
CANDLEMAKINGSUPPLIESINDIA supplies raw materials, not finished candles. That distinction is why this page refuses to answer its own title. You chose the recipe, the wick, the container and the label, so the safety judgement on a product sold under your name belongs to you together with the liability attached to it. A supplier who told you a batch was safe to keep selling would be making a claim about a product they have never seen, and you would be relying on it. Where the stakes are material, take qualified professional advice rather than a blog's.
What is verifiable here is verifiable. Every price is taken from live CSI stock in September 2026 and links to the product page, where current pricing and specification take precedence. The fragrance ceilings quoted are the wax manufacturers' stated maximum loads — 13% for Luxury Soy Wax Chunks, 10% for CSI 464, 13% for Coconut Soy CSI 468, with none published on the remaining thirteen waxes — and they are formulation limits describing how much oil the wax will hold, not regulatory limits. CSI publishes no wick sizing chart, no jar-diameter recommendation, no melt-pool or burn-rate figure, and no glass-temperature, heat-tolerance or thermal-shock figure for any product in the range, and nothing here should be read as supplying one by implication.
For GST invoicing, MSDS and IFRA documentation, invoice copies for a batch file, or help reading the stated figures on a listing, message us on WhatsApp at +91 7397976926. We can tell you what a listing states and send the paperwork. We cannot tell you whether to stop selling your batch.
Frequently asked questions
Should I stop selling the batch?
That decision is yours, and not one any raw-material supplier can responsibly make for you — you are the manufacturer of the finished candle, so the commercial and legal responsibility sits with you. Work the framework in order: express the three reports as a rate against units actually shipped; check whether they cluster on a shared wax lot, wick lot, jar shipment, pour session or oil; audit the recorded formula against the ceiling published for your wax; run a full-life burn on retained units with a matched comparison set. Then weigh holding against continuing, knowing a heat complaint is not in the same category as a scent complaint. A cautious maker may reasonably hold stock on thin evidence while investigating — a judgement, not a calculation. Where the stakes warrant it, take qualified professional advice.
Why can't I settle this from my existing burn tests?
Because the observations you need were never made. Early-burn testing covers the first third of a candle's life and the complaints describe the last third. As the wax column shortens the flame sits lower and progressively less wax and glass mass remains to absorb and spread the heat produced, so late-burn behaviour is genuinely different and a candle can pass every early test. Re-reading your development notes will not produce data that was never recorded. The only route is new data — the full-life protocol on this page, twelve units at six observation points each, with a comparison set from a batch you shipped without reports.
Is my wick too large for my jar?
Nobody can answer that from a catalogue, and be wary of anyone who tries. Eco Candle Wicks ship as Thin (C1) and Thick (C2) — the listing's own words — and CSI publishes no jar-diameter recommendation, no burn diameter, no melt-pool width, no flame height and no burn rate for any wick. What the mechanism does tell you is that a wick's fuel demand is tested hardest late in the burn, when least mass remains to dissipate the heat, which is why a wick that looked right in a first-burn test is still a legitimate candidate. Establish it by burning two otherwise identical units on the two thicknesses, full life, side by side.
Should I just add a stronger warning to my label instead?
No, and this is worth being blunt about. Instructions cannot transfer a design problem to the customer. If a candle behaves in a way you would not accept when burned exactly as your care card tells people to burn it, tightening the card is a relabelling rather than a fix. The useful question is the other one: does the burn-duration wording already on your card reflect anything you actually tested, on this recipe, in this jar, with this wick? A candle burned for many continuous hours is outside almost every small maker's tested envelope — a reason to find out where your envelope is, not a reason to blame the customer. Run the protocol, then write a card you can stand behind.
What does the test actually cost me?
On the reference unit priced on this page — a 150 g candle in a clear 150 g jar with golden lid, 135 g of CSI 464 and 15 g of Lavender on a Thick (C2) Eco Candle Wick with a sticker — materials are ₹253.16 per unit. Twelve units is ₹3,037.92, which is 4% of the ₹75,948.00 of materials in a 300-unit batch. Re-pouring all 300 with new jars is ₹75,948.00, or ₹47,628.00 if you reuse cleaned jars and absorb the labour. Holding the batch freezes ₹75,948.00 of finished stock plus your own deferred margin. Add ₹708 for a pen thermometer and weighing scale if you do not own them. CSI live pricing, September 2026.
Should I replace the three customers' candles before I know the cause?
Yes. Replace them in full, promptly, and say in writing that it is goodwill while you investigate rather than an admission of a fault — both halves of that sentence are true and both worth stating. It costs a small multiple of ₹253.16 in materials, less than the time you would spend arguing. In the same message, ask for the three things that make a report usable: a photograph of the candle as it stands, roughly how long each burning session lasted, and what surface the jar was standing on. Ask without a hint of accusation — you are collecting variables, not building a defence.
Close the gap, then decide
The test costs 4% of the batch. The decision costs whatever you judge it to cost — and it is yours.
CSI 464 from ₹260/500g with a published 10% ceiling, Luxury Soy Chunks from ₹299/500g at 13%, Coconut Soy CSI 468 from ₹325/500g at 13%, Eco Candle Wicks from ₹59 in Thin (C1) and Thick (C2), the clear 150 g jar with golden lid at ₹944 per 10, and 92 fragrance oils from 15 g to 1 kg. GST invoicing, MSDS and IFRA on request. No sizing chart, no glass-temperature figure and no safety verdict — none is published, and none would be honest.
About this guide: Written by the CANDLEMAKINGSUPPLIESINDIA team. This page contains no safety verdict. It states no temperature at which glass is safe or unsafe to touch, no maximum burn time, no glass-temperature or heat-tolerance figure, and no thermal-shock claim for any container — none of that is published, and a judgement of that kind belongs to the maker of the finished candle together with its legal consequences. Where the stakes warrant it, take qualified professional advice. The decision framework, burn protocol and cost arithmetic apply with any supplier's materials.
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 — Clear Glass Jar with Golden Lid, Luxury Soy Wax CSI 464, Eco Candle Wicks, Tin Containers 85 & 135 g, White Matte Glass Jar and Black Glass Jar with Black Lid. Names, star ratings, dates and products are as recorded by Judge.me, and only those Judge.me records as verified buyers are badged as such. One longer review is shortened with an ellipsis and is otherwise unedited. None of the six describes container temperature, full-life burning or a safety concern; they are ordinary product-quality reviews and are presented only as that.
Figures verified September 2026 · CSI live pricing: Published fragrance ceilings are the wax manufacturers' stated maximum loads and are formulation limits, not regulatory limits — Luxury Soy Wax Chunks up to 13% of total wax weight, Luxury Soy Wax CSI 464 up to 10%, Premium Coconut Soy Wax CSI 468 up to 13%; the other thirteen waxes publish no ceiling. Stated process temperatures: CSI 464 pour 80–90°C; CSI 468 pour 58–60°C; Luxury Soy Chunks fragrance addition at 80–90°C with two minutes of stirring. CSI 464 and CSI 468 each state a 60-month shelf life in cool, dry storage and cool, dry storage conditions; no other wax states a shelf life. Reference unit = 150 g finished candle: 135 g CSI 464 at ₹5,074 / 10 kg (₹0.5074/g) = ₹68.50; 15 g Lavender Fragrance Oil at ₹5,221 / 1 kg (₹5.221/g) = ₹78.32; Clear Glass Jar with Golden Lid 150 g at ₹944 per pack of 10 (₹94.40 each); Eco Candle Wicks Thick (C2) at ₹944 per 100 (₹9.44 each), Thin (C1) at ₹590 per 100 (₹5.90 each), Thick (C2) 10-wick pack out of stock; Candle Wick Stickers at ₹100 per sheet of 40 (₹2.50 each). Total ₹253.16 per unit; × 12 = ₹3,037.92; × 300 = ₹75,948.00; less the jar, ₹158.76 × 300 = ₹47,628.00. Other figures: Citrus Lemon ₹2,832 / 1 kg (₹2.832/g, 15 g = ₹42.48); Candle Wick Cutter / Trimmer ₹177; Pen Thermometer ₹354; Candle Making Weighing Scale ₹354 (₹708 together); Candle Care Stickers Roll ₹413 per 500 at 1 inch (₹0.83 each) and ₹699 per 1,000 at 1.5 inch (₹0.70 each); Bamboo Stirring Sticks ₹118 per pack of 5; Black Matte Glass Jar ₹354 per pack of 5; 200 ml Salsa Glass Jar from ₹200.60 per 5 — 200 ml is a volume, not a 200 g fill; Luxury Soy Chunks ₹299 / 500 g; Coconut Soy CSI 468 ₹325 / 500 g; 92 fragrance oils in the Fragrance Oils collection. Report rates: 3 of 40 = 7.5%; 3 of 120 = 2.5%; 3 of 300 = 1.0%; 3 of 1,200 = 0.25%. Prices and availability change — the linked product pages are always authoritative.
