I Sell Both Cotton-Wick and Wooden-Wick Candles — Should I Maintain Completely Separate Burn-Test Records for Each System?

Completely separate is the wrong shape, and one shared sheet is worse

Your question has two obvious answers and both of them are wrong in the same place. Keep completely separate records and you will maintain two copies of the jar, the wax grade, the fragrance, the load, the pour temperature, the cure window and the lot numbers — facts that are identical on both candles — and the two copies will drift. Keep one shared sheet with a wick column and you will eventually filter that column away, or sort by it, or read a row at speed, and a wooden result will be read as approval for the cotton version. That is the failure this whole page exists to prevent, and it is the expensive one.

The correct shape has three layers. Layer one is the formulation record, and there is exactly one of it per candle recipe, shared by both systems, because every fact in it is genuinely the same. Layer two is the configuration approval, and there is one per wick family and wick count, never shared, because the wick is what the approval is about. Layer three is the burn sheet and the complaint log, one set hanging under each approval, because observations belong to the configuration that produced them and to nothing else.

Counted in fields, that is 14 shared, 10 per approval and 13 per burn sheet. One system is 37 fields; two systems is 60. The duplication is 62.16% more fields, not double, and 23.33% of what you keep is genuinely shared. The answer to your title is therefore: not completely separate — separate where it counts.

None of this needs software. The three layers work in a ring binder with three dividers, in three tabs of one spreadsheet, or in a notebook with the formulation on the left-hand page and the two approvals facing it. What matters is not the tool but that the three things are visibly three things, so that nobody can write a wooden observation onto a cotton approval without noticing they have turned a page. If you are working on a phone between pours — which is how most of this actually gets recorded in a small Indian workshop — a spreadsheet with three tabs beats one long sheet precisely because switching tabs is a deliberate act and scrolling right is not.

Layer one: the formulation record both systems genuinely share

Here is the run this page records, so the fields below have real values behind them. Premium Coconut Soy CSI 468 at the 10 kg pack, which is ₹637.00 per kilogram; Vanilla Orange Blossom at the 1kg pack, which is ₹8.29 a gram; and the Clear Glass Jar with Golden Lid at ₹94.40 a piece in the Pack of 10. The jar page publishes a 150g fill capacity, an 8cm height, a golden metal lid included, and describes the glass as hot-pour safe to 85°C.

The fill reading this page carries forward is 150 g of total fill, taken from the jar’s own published capacity, which at 8% of wax weight divides into 138.89 g of wax and 11.11 g of fragrance oil. That is a deliberate choice and it is stated here because the other reading — 150 g of wax plus 8% on top — gives ₹298.92 of materials against ₹284.47, a difference of ₹14.45, which is 1.53 times the entire wick line. The reading goes in the formulation record as a field, not as an assumption, because two people working from the same recipe have to pour the same candle.

The published positions that govern this candle also belong here. Vanilla Orange Blossom is a perfume-grade oil whose page lists candles as one use among several, with a recommended candle load of 7-10% and the instruction to start at 7%; Premium Coconut Soy CSI 468 states a load of up to 13% of total wax weight, a melting point of 50-53°C and a pour temperature of 58-60°C. Where two published figures apply the lower one governs, so the ceiling here is 10%, and 8% sits above the oil page’s own starting point rather than below it — which is a fact worth having in the record, because if the configuration later disappoints, 7% is a documented step you have not yet taken. The oil page also states a flash point above 93°C, which sits above the wax’s published pour window, and notes that its high vanilla, tonka and honey content can cause moderate to significant browning in white or light-coloured wax over time.

The scheme has to survive your tenth formulation, not just your first, so give the formulation record a number rather than a name. F-01 is this jar with this wax, this oil and this load. A new fragrance in the same jar is F-02, not a variant of F-01, because the fragrance is one of the things the approval is testing. A new load in the same jar with the same oil is F-03. This feels wasteful at four formulations and is the only thing that still works at forty, when the alternative is a record called “vanilla jar small new” that two people interpret differently. Names drift; numbers do not.

One field in that record is doing quiet work worth naming. Storage conditions during cure looks like padding until the first time a candle cured in a Mumbai monsoon disagrees with the same formulation cured in a Pune February. We publish nothing about how either wick family behaves under Indian heat or humidity, and nothing about how cure conditions change what you will see on the sheet, so your own record of the room is the only evidence that will ever exist for your candle. Note the month. Note whether the room has a fan or an air conditioner. It takes four words and it is the field you will be most glad of in a year.

Layer two: the approval, which is never shared and never inferred

An approval is a decision, not an observation. It says: this exact combination of vessel, wax, fragrance, load and wick is one I am willing to sell, judged against thresholds I set before I lit it. Everything above the wick is in layer one. The approval adds the wick and the verdict, and it is the layer that must never be shared, inferred, or carried across from one family to the other.

Why never inferred: changing the wick family changes the only variable the approval was testing. A wooden result tells you nothing about the cotton version of the same candle and the reverse is equally true. This is not a cautious position, it is the definition of the test. If a wooden pass could stand in for a cotton approval there would be no reason to run two.

Two rows there carry most of the weight. Thresholds set in advance is the row that makes an approval an approval rather than a rationalisation; a threshold invented after you have seen the result is not a threshold, it is a preference. And exposed length at pour matters far more on the wooden record than the cotton one, because trim length is one of the few levers the wooden family actually gives you — the wicks come in one width, so there is no gauge to step to and the adjustable variables are trim, load, wax grade and family.

Layer three: the burn sheet and the complaint log

Layer three is where observations live, and observations are the property of one configuration. There are two kinds: the burn sheets you fill in during the trial, and the complaint log that fills itself in after the candle ships. Both hang under an approval ID. Neither ever sits in a shared file.

The burn sheet, and the rows each family does not have

Most of the sheet is common to both systems, which is convenient and also dangerous — it is precisely why the sheets look interchangeable at a glance. The rows below are the ones worth keeping, and the last two columns show which family each row actually applies to. Every cell you write is yours; there is not one figure from us anywhere in it.

Two rows are family-specific and they are the reason a single shared sheet never quite works: mushrooming is a cotton observation and char on the exposed wood is a wooden one. Keep both rows on both sheets with not applicable in the inapplicable one and the sheets stay the same shape, which makes them easy to compare and easy to fill in. The ambient row matters more in India than almost anywhere, because the same candle in a Mumbai July and a Delhi January is not being tested under the same conditions, and a ceiling fan over a burning candle changes what you are looking at. Note the fan. Note the season.

The complaint log, which is where most records quietly fail

A complaint is the only burn data you will ever get from a real room, which makes the complaint log the most valuable record you keep and the one most likely to be a chat thread. It needs three fields before anything else: the approval ID, the batch, and what the customer actually described. Without the approval ID a complaint about a wooden candle is just a complaint about your brand, and it teaches you nothing about which configuration produced it.

Keep the logs separate by approval even though that feels like over-engineering at three complaints a year. The point is not volume, it is attribution: five complaints spread across two systems look like bad luck and five complaints concentrated in one approval look like a configuration to withdraw. Your records are the only thing that can tell the difference, because we publish nothing that can.

The cross-reference: put the family in the ID, not in a column

This is the single change that does the most work, and it costs nothing. Every identifier below layer one carries the wick family inside it. F-01 / C2 × 1 and F-01 / WD × 1 are two strings that cannot be confused by sorting, filtering, a narrow phone screen, a photocopy, or somebody reading quickly at seven in the evening. The formulation is the stem, the family and count are the branch, and the branch is never optional.

Use the same string everywhere the configuration appears: on the approval, at the head of every burn sheet, on the sticker under every test candle, in the batch record, on the production instruction, and in the complaint log. When the string is identical in all six places you have a cross-reference without maintaining one, and the only way to lose track of which system a result belongs to is to delete the ID outright.

What the twin approval costs, and what a mis-read one costs

Records feel like an overhead because they have no invoice. Here is the invoice. On the costed candle, materials from CSI come to ₹284.47 with the Thick (C2) wick and ₹334.03 with a wooden wick from the Pack of 10. At 3 candles per configuration, approving both properly is ₹853.40 on the cotton side and ₹1,002.08 on the wooden side — ₹1,855.49 in total, which is the entire price of keeping the two systems straight at the only point where it is cheap to do so.

It is worth seeing where that money actually sits, because it changes how seriously you take the records. On the cotton candle the jar is 33.18% of materials, the fragrance is 32.40%, the wax is 31.10% and the wick is 3.32%. The wick is the smallest line on the invoice and the largest single cause of a candle being wrong, which is the whole argument for keeping its approval separate. You are not protecting the ₹9.44. You are protecting the ₹284.47 around it, and the customer who lights it.

Now the other side. Suppose the cross-reference is missing and a wooden pass is read as approval for the cotton version of the same candle. You pour a batch of 50 and ship it. That batch is ₹14,223.39 of materials in jars — 7.67 times the cost of approving both configurations properly in the first place — and that figure counts only what CSI sold you. It does not count your labels, your boxes, your labour, the courier to your customer, the replacements, or the part of this that does not have a price at all, which is a customer who has already lit one.

Note which direction the error runs, too, because people assume it runs the safe way. It does not have a safe way. A wooden result read as a cotton approval ships an untested cotton candle; a cotton result read as a wooden approval ships an untested wooden one, and the wooden batch is dearer — ₹15,816.39 of materials at the Pack of 100 rate. Neither direction is recoverable by apologising well.

When a record has to be re-opened, and which layer it hits

A record is not a filing exercise, it is a claim that holds until something underneath it moves. The useful property of the three-layer shape is that it tells you immediately how far a change reaches. A change in layer one reaches both approvals. A change in layer two reaches one. A change in layer three reaches nothing but your confidence.

The row that surprises people is the vessel. A jar of the same published capacity from a different source is a different vessel, and the approval does not travel with the capacity figure. The Clear Glass Jar with Golden Lid page publishes a 150g fill capacity and an 8cm height, and that is the vessel this record is about — no CSI vessel page publishes an internal diameter, so an approval is tied to the named jar rather than to a measurement you can match elsewhere. Swap the jar, re-open both approvals.

There is a seasonal version of the same logic that matters in India. Nothing in your formulation changes between February and July, so nothing formally re-opens — but your ambient row changes completely, and a configuration approved in a cool month and never re-checked in heat is an approval with an unexamined condition attached to it. Put a dated re-check against your approvals rather than a re-approval, and let the ambient row in your burn sheets earn its place. The honest position is that we publish nothing about how either wick behaves at 40°C or in monsoon humidity, so your own seasonal records are the only evidence that will ever exist for your candle.

What these records cannot tell you

Nothing about how either wick performs. CSI publishes no burn times, melt-pool diameters, flame heights, consumption rates or soot measurements for any wick in the range, none exists in our data, and nothing on this page implies we have burned these candles. Every row in every sheet above is blank until you fill it, and every threshold is a judgement you make and write down in advance. If you were hoping for a number to compare your result against, there is not one, and a page that invented one would be worth less than this one.

Nothing about which wick to start with. A sizing table from diameter to wick is not something any CSI wick page carries. The guide by diameter is where a starting point comes from, and the burn test decides from there. On the cotton side the pre-tabbed range is exactly two sizes, Thin (C1) and Thick (C2), so a cotton approval that fails has one alternative within the family. On the wooden side there is one width, so a wooden approval that fails has none, and the levers are trim length, load, wax grade and wick family.

Nothing that makes a record a substitute for judgement about safety. Keep the guidance on both approvals conservative and general: trim before every burn, burn within the time stated on your own label, never leave a candle unattended, and discontinue any configuration that leaves the vessel too warm to hold comfortably. We will not tell you that a configuration is safe, and your records should not record that we did.

Nothing about lead times or document trails we do not publish. CSI publishes no lead time, so the gap between a reorder and your bench is a field you fill from your own order history. MSDS and IFRA documentation is available on request through WhatsApp, and if your records keep a documentation field — they should — the date you requested it belongs in layer one beside the lot numbers. For how a controlled comparison is run at all, running a controlled burn test and testing hot throw before a launch are the live pages to read beside this one.

Frequently asked questions