My Candle Smells Different After Being Stored for Six Months — How Do I Determine Whether the Fragrance Changed or the Wax Affected How I Perceive It?
Share
Why would a candle smell different rather than just weaker? Because a blend is a mixture, and the lightest materials are the ones that most readily leave a solid wax surface. If the brightest part of the top recedes at the surface, what reaches you is proportionally more base — so the candle reads different, not quieter. Soy wax also has a faint character of its own, which becomes relatively easier to notice as the top eases back.
Can I test the material rather than the surface? Yes — shave 5 g from the aged candle and 5 g from a control, melt each separately, and smell the melt. That bypasses the solid surface entirely and is the single most informative half-hour in this guide.
Is the stock dispatchable? Inspect, then decide. A character change is a quality judgement against your own tolerance. Cracked glass, a wick touching the wall, a clogged wick and a jar that runs uncomfortably hot must never ship, whatever it smells like.
The verdict on the stock in hand
The verdict is inspect, then decide. Nothing about a candle smelling different is a safety matter, and nothing about it can be settled by a supplier's decree — it depends on what you promised the customer and on how far your product can drift and still be the product. What can be settled from a distance is the list of faults that block a dispatch regardless.
| Finding | Category | Action |
|---|---|---|
| Cracked, chipped or starred glass | Must not ship | Withdraw; inspect the carton and shelf position it came from |
| Wick leaning on or touching the glass | Must not ship | Withdraw — a votive's small diameter makes this more likely, not less |
| Wick clogged, buried or drowned | Must not ship | Withdraw and burn-test a sibling before releasing the batch |
| Jar uncomfortably hot in a test burn | Must not ship | Stop the release; wick-to-vessel, not storage |
| Fails your own written descriptor tolerance | Decide | Discount, use as samples, or re-label — but say what you are doing |
| Passes the triangle test as indistinguishable | Dispatchable | Release and log the panel result on the batch sheet |
| Wet film or droplets on the surface | Decide | A separate diagnosis; note the load and the shelf position |
| Frosting, faint crystals, a rougher top | Cosmetic | Soy crystal behaviour. Set and publish your own tolerance |
Twenty units is a little over 3% of a 600-unit votive hold and takes about an hour. The row that matters commercially is the fifth one, and it only works if the tolerance exists in writing before the panel runs. A sentence is enough: "we release votives whose blind panel cannot be distinguished from a control at five of six, whose descriptor profile is within one point on every note, and which pass the physical list." Write it once and it settles a dozen future arguments.
Why a blend can change character, not just volume
This is the part that makes "different" an intelligible observation rather than a mystery, and it needs no invented chemistry. A fragrance oil is a mixture of many materials, and CSI's pages publish that structure openly — 88 of 107 pages state a full top, heart and base pyramid, and a further 14 state a scent profile. The materials named in a top are, as a class, lighter and more volatile than those named in a base. That is the whole basis of the pyramid convention.
Now add the one physical fact from the storage side: a cold candle releases fragrance from a solid surface at room temperature, and the materials most able to leave a solid surface at room temperature are the lightest ones. Put the two together and the consequence is about proportion. If the surface's brightest, lightest element becomes less prominent, then what reaches your nose from that surface is proportionally more heart and base — so the candle reads different, and often heavier, rather than simply quieter. No rate is needed for that argument, and none is available: no CSI page states one.
| Oil | Stated top | Stated base | What a proportional shift would read as |
|---|---|---|---|
| British Rose | Single-note page — states a rose profile rather than a three-tier pyramid | Not stated as a tier | Less scope for a tier shift; a change here is more likely to be surface or reading |
| Bergamot Amber | "Sparkling bergamot and citrus" | "Woody musk and smooth vanilla" | Warmer, rounder, less sparkling — a character change, not a volume change |
| Palo Santo | "Soft citrus" within the stated main aroma | "Warm amber, and a touch of incense" | More resinous and smoky, less lifted at the opening |
And there is a second contributor that gets forgotten: the wax is not odourless. Soy has a faint character of its own, mild enough to sit underneath a fresh fragrance and noticeable enough to register when the brightest part of the opening eases back. Nothing has gone wrong in that scenario — the candle has simply changed the balance it presents. This is also why a "different" reading is much more informative when it comes with descriptors than with a single stronger-or-weaker verdict, and why the same question in a fresh candle, where the wax choice does the same thing, is treated in why the same fragrance smells strong in paraffin but weak in soy.
In the candle, or in the reading?
Eight candidates, split into the two families the question asks about. Each has a distinguishing test, and several can be true at once. Work the whole table and write every answer down; the actions are completely different and one of them costs nothing.
| Candidate | Family | The test that separates it |
|---|---|---|
| Proportional shift — lighter material less prominent at the surface | In the candle | Descriptor sheet: the top descriptors fall while base descriptors hold |
| Real change in the material itself | In the candle | Shave-and-melt: melted shavings differ blind as well as the jars |
| Surface change only — crystals, bloom, a rougher top | In the candle | Shave 1–2 mm off one top; re-run the triangle on the freshened surface |
| Wax character now relatively more noticeable | In the candle | Melt a sample of unfragranced wax from the same lot and smell it alone |
| Headspace released at first opening weeks ago | In the reading | Both lids off for 30 minutes, then repeat the triangle |
| Room temperature, airflow or competing scent | In the reading | Four-hour temperature match, one unscented room, same door and window state |
| Adaptation and session order | In the reading | Randomise the odd jar and the order; three jars maximum per session |
| Memory drift — comparing with a recollection | In the reading | The triangle test itself: it never asks you to recall anything |
The last row is why the triangle test is the right instrument for this particular question rather than the paired comparison used elsewhere in this cluster. A paired test asks "which of these two is stronger", which is answerable. Your question is "has it changed from how it was", which a paired test cannot ask, because the original is gone. The triangle reframes it into something physical: can anyone reliably distinguish the aged candle from a candle of the same pour stored differently? That question has a yes-or-no answer and a known chance of being answered correctly by luck.
The blind triangle test and its pass bar
Six steps, six sessions, six retained units. Run the sessions across two or three days rather than back to back, because a nose that has done three triangles is not the nose that did the first one. The pass bar and session count are a defensible method rather than a standard — but the probabilities attached to them are plain arithmetic and they are the reason this bar and not a lower one.
| Score | Chance by guessing | As a fraction | Reading |
|---|---|---|---|
| 6 of 6 correct | 0.14% | 1 in 729 | A clear difference. Move to the shave-and-melt test. |
| 5 or 6 of 6 | 1.8% | 13 in 729 | The pass bar used here. Act on it. |
| 4 or more of 6 | 10.0% | 73 in 729 | Not a finding. Convincing to a human, unremarkable to arithmetic. |
| 3 or fewer of 6 | 90.0% | 656 in 729 | Indistinguishable. A releasable, loggable result. |
The descriptor sheet
The triangle test tells you whether there is a difference. It does not tell you what the difference is, and "different" is the whole complaint. For that you need a descriptor sheet: five named notes taken from your own oil's stated profile, each rated 0–5, filled in for the aged unit and the control unit separately and blind.
| Descriptor | Where it sits in the stated profile | Pattern that points at a proportional shift | Pattern that points at the reading |
|---|---|---|---|
| Bright / sparkling opening | Top | Falls by 2 or more against the control | Falls equally on both units |
| Floral or fruity middle | Heart | Falls by about 1 | Unchanged on both |
| Warm / resinous depth | Base | Holds, or reads higher relative to the rest | Unchanged on both |
| Sweetness | Heart to base | Holds or rises relatively | Tracks the assessor's mood, not the jar |
| Waxy / neutral background | Not the fragrance at all | Reads higher as the top eases back | Reads the same on both units |
The signature of a proportional shift is a pattern, not a single number: top descriptors down, base descriptors level or relatively higher, and the waxy background more noticeable. The signature of a reading problem is that both units move together, which the control makes visible immediately. Use the descriptors your own oil's page names, so the sheet is about your product — British Rose, for example, is one of the 14 single-note pages that state a profile rather than a three-tier pyramid, so its sheet needs descriptors drawn from the rose character itself rather than a borrowed top-heart-base template. Fragrance-family behaviour over a cure, which is the closest well-documented analogue, is set out in how cure time changes the smell of different fragrance families.
The shave-and-melt test
The last instrument, and the one that splits "the candle changed" from "the surface changed". It takes about thirty minutes and destroys nothing you were going to sell.
Shave 5 g from the top of an aged unit and 5 g from the top of a control unit, keeping them in separate covered containers. Melt each one separately and smell the melt. A Mini Electric Wax Melter makes this repeatable — its page states Gear I at approximately 70–85°C and Gear II at 100–120°C, with the note that "the ideal melting range for most candle waxes is 80-90°C" — but a small pan over hot water does the job as long as both samples get identical treatment. Clean the vessel between samples, or use two identical vessels.
| Jars, blind | Melted shavings, blind | Interpretation | What you change |
|---|---|---|---|
| Distinguishable at 5 of 6 | Indistinguishable | The difference is at the surface, not in the material | Storage, handling, lids, less opening. Nothing in the formula. |
| Distinguishable at 5 of 6 | Also distinguishable | A change in the material is now the leading candidate | Hold length, packaging, storage conditions — then, last, the formula |
| Indistinguishable | Indistinguishable | No detectable difference. The reading was the variable. | Your test procedure and the sniffing room. Release the stock. |
| Indistinguishable | Distinguishable | Unusual. Suspect the melting step — different temperatures or a contaminated vessel | Re-run with two identical vessels and one gear setting |
What the panel is poured from
The reference build for this guide is deliberately small and cheap, because a blind panel wants several identical units rather than a few large ones. Each card lists every pack size and the per-candle figure at the 70 g, 8% basis.
| Size | Price | Per piece |
|---|---|---|
| Pack of 5 | ₹159.30 | ₹31.86 |
| Pack of 10 | ₹318.60 | ₹31.86 |
| Size | Price | Per 100g | 70g candles @ 8% |
|---|---|---|---|
| 15 gm | ₹94.40 | ₹629.33 | ≈2 |
| 50 gm | ₹212.40 | ₹424.80 | ≈8 |
| 100 gm | ₹342.20 | ₹342.20 | ≈17 |
| 500 gm | ₹1,593 | ₹318.60 | ≈89 |
| 1 kg | ₹3,068 | ₹306.80 | ≈178 |
| Size | Price | Per kg | 70g candles @ 8% |
|---|---|---|---|
| 1 kg | ₹399 | ₹399 | ≈15 |
| 5 kg | ₹1,996 | ₹399.20 | ≈77 |
| Size | Price | Per piece |
|---|---|---|
| Default Title | ₹2,360 | — |
Put against the hold, the whole apparatus is small. Six retained votives are ₹495.72 and a 600-unit hold is ₹49,572, so the panel is 1% of the batch. The fragrance for that run is 3,360 g, which at the listed packs is three 1 kg bottles plus four 100 g — ₹10,572.80, since 1 kg is the largest listed fragrance pack and the listed per-kilo price is the ceiling. Raising the load from 8% to 10% on the next 600 would cost ₹86.85 a candle instead of ₹82.62 — ₹2,538 across the run — which is five times the panel and answers none of its questions.
CANDLEMAKINGSUPPLIESINDIA supplies raw materials, not finished candles. A supplier who wanted to sell more oil would tell you the fragrance has degraded and recommend a heavier dose. This page tells you that the most likely finding is no detectable difference, and gives you the arithmetic to prove it — which sells nothing except a pack of votive jars. We write it this way because the alternative is makers scrapping sound stock and chasing a character change that only exists in the gap between a jar and a memory.
Every price, pack size and specification comes from live CSI stock in September 2026 and links to the product page, which is always authoritative on current pricing. Stated figures are quoted only where a page states them: British Rose "Flashpoint 192°C" and "Vanillin content 0% - no discolouration or yellowing in wax"; Eco Soy CSI 400 "add your fragrance or essential oils at 85°C (185°F), stir gently for 2 minutes, and pour directly" and a stated fragrance add temperature of 80 to 90 degrees, with no maximum load stated; White Matte Votive Jar "Capacity: ~70g wax fill"; Mini Electric Wax Melter "1 litre capacity, dual power modes (300W and 600W)", "Gear I temperature Approximately 70-85°C", "Gear II temperature Up to 100-120°C", "The ideal melting range for most candle waxes is 80-90°C", "1-year warranty". Scent-structure counts are from CSI's own published pages: 88 of 107 state a full top, heart and base pyramid, 14 state a profile without one, and 5 state neither. Loads are formulation limits, not regulatory limits. No CSI fragrance-oil page states a shelf life, an expiry or a storage instruction — zero out of 107 — so this guide publishes no ageing rate, no percentage lost, no migration rate and no accelerated-to-real-time conversion. The proportional-shift argument here is about direction only; the only numbers on this page are CSI prices and the binomial arithmetic of a forced-choice triangle.
For help designing a blind panel for a specific product, bulk pricing above the listed packs, GST invoicing, MSDS or IFRA documentation, message us on WhatsApp at +91 7397976926.
Frequently asked questions
- Weaker cold throw at six months — can fragrance performance change in storage?
- Strong when burned but weaker cold — has the fragrance actually been lost?
- Citrus against woody — why fragrance profiles can age differently
- How cure time changes the smell of different fragrance families
- Why does the same fragrance smell strong in paraffin but weak in soy?
- How to compare two fragrance oils using a controlled burn test
- Why should candle makers test fragrance after curing?
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; two longer reviews are shortened with an ellipsis and otherwise unedited.
Figures verified September 2026: Reference candle = 70 g finished at an 8% fragrance load, calculated as a share of finished candle weight (64.4 g wax + 5.6 g oil); the Luxury Soy Wax Chunks page expresses its own range as a share of total wax weight and that alternative convention exists — this series keeps the finished-candle basis. Eco Soy CSI 400 ₹399/1 kg and ₹1,996/5 kg (₹399.20/kg), 5 kg being the largest listed pack → 64.4 g = ₹25.70; 39 kg as seven 5 kg plus four 1 kg = ₹15,568. British Rose Fragrance Oil ₹94.40/15 gm, ₹212.40/50 gm, ₹342.20/100 gm, ₹1,593/500 gm, ₹3,068/1 kg (₹3.422 per gram at 100 g) → 5.6 g = ₹19.16; 3.4 kg as three 1 kg plus four 100 g = ₹10,572.80. Eco Candle Wicks Thin (C1) ₹590/100 → ₹5.90 each. White Matte Votive Jar 70 gram ₹159.30 per 5 and ₹318.60 per 10 → ₹31.86 each. Reference candle total ₹82.62; 6 retained units ₹495.72; 600 units ₹49,572. At a 10% load the same candle is 63 g wax + 7 g oil = ₹86.85, a difference of ₹4.23 per candle and ₹2,538 across 600. Pen Thermometer ₹354; Mini Electric Wax Melter ₹2,360. Probabilities: binomial for six independent forced choices with probability 1/3, denominator 3⁶ = 729 — exactly six correct 1/729 (0.14%), five or six 13/729 (1.8%), four or more 73/729 (10.0%), three or fewer 656/729 (90.0%). Stated specifications as published on product pages: British Rose "Flashpoint 192°C", "Vanillin content 0% - no discolouration or yellowing in wax"; Eco Soy CSI 400 "add your fragrance or essential oils at 85°C (185°F), stir gently for 2 minutes, and pour directly", "Fragrance Add Temperature 80 to 90 degrees", no maximum load stated; White Matte Votive Jar 70 gram "Capacity: ~70g wax fill"; Mini Electric Wax Melter "1 litre capacity, dual power modes (300W and 600W)", "Gear I temperature Approximately 70-85°C (158-185°F)", "Gear II temperature Up to 100-120°C (212-248°F)", "The ideal melting range for most candle waxes is 80-90°C", "1-year warranty"; Bergamot Amber stated notes of sparkling bergamot and citrus, warm amber and soft floral, woody musk and smooth vanilla; Palo Santo stated main aroma "Earthy palo santo with soft citrus, warm amber, and a touch of incense". Scent-structure counts across CSI's 107 fragrance-oil pages: 88 state a full top-heart-base pyramid, 14 state a profile without a pyramid, 5 state neither. Assumptions labelled as method, not standard: six triangle sessions with a five-of-six pass bar, a four-hour minimum temperature match, two breaths at about 5 cm per jar sniffed left to right twice, three sessions maximum per sitting, 5 g shaved per sample for the melt test, 20 units inspected per 600-unit hold, 6 units retained per pour, and 600 units as the worked example. What this guide does not state, because no source does: a shelf life or expiry for any oil or finished candle (zero of 107 CSI fragrance-oil pages state one), an ageing or migration rate, a percentage of any note lost over any period, a temperature threshold for any storage symptom, or any accelerated-to-real-time conversion. Prices and availability change — the linked product pages are always authoritative.