My Soy Candles Start Sweating During Indian Summer — Should I Change My Wax or Reduce My Fragrance Load?

The question offers two routes. There are three, and the third is free

A candle that behaved from November to February and started beading in April has not developed a manufacturing fault. The same wax, the same oil, the same percentage and the same process are now being asked to hold together twenty degrees warmer. Sweating is fragrance oil leaving the wax it was dissolved into and collecting on the surface; heat is what makes the wax softer and the oil more mobile, so the amount a given wax can hold in solution is lower in May than it was in December. Every one of the three routes below works on that single sentence from a different side.

Route 2 lowers how much oil the wax is being asked to hold. Route 3 changes what is doing the holding. Route 1 changes how warm it gets. Only Route 1 attacks the thing that actually changed, and it is the only one of the three that costs nothing — which is why a supplier who would happily sell you a different wax is putting it first.

Ten minutes of triage before you spend anything

Pick up three jars from the batch that is misbehaving and answer these in writing. The answers decide which of the three routes is even worth testing.

Route 2 — reduce the load: how much, and what it really costs

Read the Luxury Soy Wax Chunks page carefully and it does not simply give you 13%. It gives a ladder: 8% as a starting point, 10% as the recommended figure, 13% as the maximum with the instruction to burn-test thoroughly at that level. A maker sitting at 11% is between the recommended figure and the maximum, which is a legitimate place to be in a cool month and an ambitious one in May. Premium Coconut Soy CSI 468 also states up to 13%; Luxury Soy Wax CSI 464 states up to 10%; Eco Soy CSI 400 does not publish a figure, so ask on WhatsApp before you dose it.

Check the oil's page as well, because it carries its own range and the lower of the two always governs. Cinnamon Vanilla states up to 13% for candles, so at 11% the pair sits inside what each page allows — which is exactly why this is a heat question and not a dosing one. Many other CSI oils state 6–10% for soy, and a maker running one of those at 11% has a more ordinary overload on their hands.

Two points off the load saves ₹18.22 a candle and ₹1,821.92 across a hundred. That is the pleasant part. The honest part is what you might lose: some scent strength, sometimes. Whether you lose any is a question the twelve-jar test answers directly, because Set B is scored for hot throw at day 14 alongside the control. In a well-wicked jar with a full melt pool, the difference between 11% and 9% is frequently inaudible, and a candle that does not sweat is worth more than two percentage points that nobody can smell.

Route 3 — change the grade: what CSI's pages actually let you compare

This is the section where a raw-material supplier is supposed to sell you an upgrade, and there is a limit to how much of that we can honestly do. Melt point is the figure that would tell you how a wax behaves in a 42°C room, and across CSI's soy shelf it is largely not published. Premium Coconut Soy CSI 468 publishes a melting point of 50–53°C. Eco Soy CSI 400 does not publish one. Luxury Soy Wax Chunks publishes working temperatures rather than a melt point, and Luxury Soy Wax CSI 464's page gives 80–90°C as the working range for melting and pouring. If you need those figures for a specification sheet or a buyer's questionnaire, ask for them on WhatsApp rather than inferring them from a blog.

What you can compare honestly is what the pages do state, and one of those comparisons cuts against the obvious upgrade. CSI 468's melting point of 50–53°C is the lowest published working figure in the range, and a closed delivery van standing in an Ahmedabad car park, or a metal-roofed godown in Nagpur in May, is not far from that number. It is an excellent container wax with a page describing an ultra-smooth finish, minimal frosting and excellent glass adhesion, and it costs ₹10.20 more per 200 g jar than Luxury Soy Wax Chunks. What it is not is a heat-resistance upgrade, and buying it as one would be an expensive misreading.

There is one more option on the shelf that people ask about, and it deserves a straight answer rather than a sale. Vybar Additive is ₹156.00 for 50 g, and its page says it is compatible with soy at 0.5–2% of wax weight — the lower end, 0.5–1%, for soy — added to the melted wax before fragrance and colour, with claims of improved fragrance retention, a smoother surface and less sweating. Our position across this series is unchanged: treat it as an optional, controlled experiment, not as a substitute for fixing temperature or storage, and note that single-pour soys such as Eco Soy CSI 400 are sold as needing no additives at all. If you want to test it, test it as a fifth set in the same run, not as a change to production.

Route 1 — storage, packing and dispatch: the free one

This route has no product to sell and it is the one we would put money on. The aim is simple: keep finished stock below the temperature at which the wax stops holding its oil, for as much of its life as you control. You control the cure shelf, the storage cupboard, the packing bench and the hour of dispatch. You do not control the van, and pretending otherwise is how brands end up blaming waxes.

Be honest with yourself about the last leg. A parcel can sit in a closed vehicle at midday in Delhi, Ahmedabad or Nagpur at temperatures no container wax on any supplier's shelf is designed to shrug off, and CSI does not sell insulated packaging or temperature-controlled freight. What a maker can do is reduce the number of hot hours: pour smaller festive batches closer to the date, book earlier dispatch slots, and tell the customer in the box that a candle which has travelled should be stood upright somewhere cool for a day before it is lit. That last sentence has rescued more summer orders than any wax change.

The twelve-jar test: four sets, three checkpoints, fourteen days

Pour twelve jars on the same afternoon: three each of four sets. Set A is your current recipe in your current storage — the control. Set B is the same wax with the load cut by two points. Set C is a different grade at its own stated ceiling. Set D is the control recipe held in cool, sealed storage. Materials come to ₹2,573.92, which is roughly ₹214.49 a jar, and the comparison you get is between jars that lived through the same fortnight of weather — the only comparison that means anything in a season that changes week to week.

Reading the result — and what to do when it is a tie

Write these rules on the scoring sheet before day 3, so that the decision is made by the test and not by whichever route you had already half-decided on.

Two rules of interpretation matter more than the table. First, a gap of one point out of three is a tie, not a result — noise in a tissue test is easily that large, and a tie should always be settled in favour of the cheaper option. Second, run the whole thing again in a different week before you rebuild production around it. A fortnight in early May and a fortnight in late June are not the same experiment, and a result that repeats is worth five times a result that does not.

What each route costs across 100 candles — and what guessing costs

Read the third row carefully, because it is the most interesting line in the table and the least comfortable one for us to print: the cheapest route on paper is the grade change to CSI 464, and it saves ₹2,742.96 across a hundred candles — but it saves that money mostly by capping your load at 10%, which is Route 2 with a new invoice attached. If the twelve-jar test says the load was the problem, you can have that saving without changing wax at all, simply by dosing Luxury Soy Wax Chunks at 9% and keeping the pour window, the wick and the process you already know.

Set the whole table against the calendar. Diwali 2026 falls on 8 November 2026, and the sweating season in most of India is behind you by then — but the wedding and corporate gifting run from April to June is not, and neither is next summer. Running this test in September, when the weather is moderating, gives you a clean control and an answer in hand before the April orders arrive. A test run in the third week of May, on stock that is already leaving the studio, is a test you will not trust.

Frequently asked questions