How Much Safety Stock Should I Keep for a Wick That Is Used Across Several Bestselling Candle SKUs?

Your shared wick, made precise

The worked example, labelled. Four SKUs, each a 200 g candle at 8% (16 g of oil and 184 g of wax, load as a percentage of total weight) in your own 8 cm jar, all approved on Eco Thick (C2) after full-life burns: Lavender, Woody Oud, Citrus Lemon and British Rose. Woody Oud was the hardest to approve; its first wick sooted by hour three and C2 was the rung that passed. These four sell steadily, so the C2 bin empties steadily, and last month it emptied two days before the next pack arrived. Someone suggested wicking a Woody Oud batch on Thin (C1) "just this once". It didn't happen, but it shows the real cost of a stock-out: a substitution that undoes the test that fixed the soot (the master formula fields that forbid it are 6996).

Your question is how many C2 to hold so that doesn't happen again, without filling a cupboard. The records you need are the wicks drawn from the bin each week for each SKU (candles plus wicking rejects), how long your supplier takes from order to door, and how long your new-lot check takes. Eight weeks of the worked example's draw:

Why one wick across four SKUs changes the arithmetic

Demand pools. Each SKU's week-to-week draw wobbles: Lavender between 55 and 80, British Rose between 18 and 35. Buffer each one separately and you hold enough for every SKU's bad week at the same time, which rarely happens. Add the draws together first and some of the highs and lows cancel: the combined total's standard deviation (21.1) is clearly less than the four separate ones added up (8.9 + 6.3 + 8.7 + 5.6). That's the whole case for one pooled buffer. It's also the honest case for holding fewer wicks than four separate calculations suggest: 93 against 66 for the demand part.

But it pools less than theory says. If the four SKUs were independent, the pooled buffer would be 47. Calculated from the real weekly totals it's 66, because your SKUs move together: a gifting week lifts all four. That's why the method below uses the variation of your actual weekly totals rather than a formula that assumes independence. In festive weeks they move together even more; plan those separately.

Risk concentrates. The same pooling that saves wicks puts four top-selling SKUs on one lot. If a new C2 lot burns taller than the retained sample, all four formulas are affected at once, and Woody Oud, already the closest to its soot limit, shows it first. So a shared wick needs two buffers with different jobs: a demand buffer for normal variation, and a decision about what you'll do when a lot fails. For how the wick-SKU count itself is decided, see the published post on reducing wick SKUs without forcing every fragrance into one wick.

When the shared wick runs short or gets held: causes, ranked

The ranking is our judgement for a small maker with one wick shared across several SKUs. The table separates causes that come from how customers burn your candles from causes in what you make and buy, because one of them often triggers a hold that empties the bin for no reason.

Cause 5 deserves its row. A single customer photo of a black rim is usually about trimming, burn length or a fan (that's the soot cluster's whole lesson, the soot hub), but a nervous maker can quarantine the whole shared lot and stop four SKUs. One retained candle from the complaint batch, burned the way your label says, settles it in a session (what to retain and how to use it). Cause 3 is the one most small makers miss: they set a reorder point on the courier's lead time and forget the ten days the new pack sits waiting for its lot check.

The pooled safety stock, step by step

This is a standard textbook method, labelled as such, not a candle-industry rule. It assumes weekly demand is roughly bell-shaped and that eight or more weeks of records are representative. If your history is shorter, or dominated by one big order, treat the result as a starting point.

Two things in that table usually surprise people. The lead-time spread matters more than the sales spread. Half a week's uncertainty in when the pack arrives adds 89 wicks, more than the whole demand buffer. If you want a smaller buffer, measuring and steadying your lead time does more than forecasting sales. Ask us on WhatsApp for a realistic dispatch estimate to your city; we'll say what's typical and won't promise what we can't. The reorder point is compared with your stock position, not the bin: wicks on hand + packs ordered but not arrived + packs arrived but still in their lot check. Otherwise you'll reorder twice.

Then the money, honestly. Pooling saves 27 wicks of demand buffer against four separate calculations, worth ₹254.88 at ₹9.44. That's small, because wicks are cheap. The reason to pool isn't the rupees; it's that one well-sized buffer is easier to keep honest than four, and it shows you the shared-lot risk clearly. Argued against our own interest: don't round a 155-wick buffer up to a 1,000-pack "to be safe". The C2 1,000-pack is ₹9.44 a wick like the 10-pack; it only ties up cash. How large each order should be (a month, six weeks, six months) is 6998; the full purchase-quantity formula is 6999.

The second risk: one lot stops four SKUs

A demand buffer protects you from ordinary wobble. It does nothing if the new pack burns differently, because the buffer and the new pack may be the same lot. Two rules handle that. First, never open a new lot until it's released, and keep the previous released lot as the working stock while the new one is checked; that's why the check time sits inside the effective lead time. Second, decide in advance what happens if a lot fails. A failed check means another order, another supply wait and another check: one more effective lead time, about 594 wicks of demand. You can hold that as a lot-failure reserve (₹5,607.36), taking the reorder point to 1,343, or accept the gap and plan to pause the four SKUs for about 3.5 weeks. The failed check alone leaves about 255 candles' worth of wicks unusable for its 1.5 weeks. Our judgement: hold the reserve for the festive run-up, when a stopped bench is expensive, and not for the rest of the year.

The release test itself is the lot-comparison method in 6994, which owns the measurements and the paired limits. What a shared wick changes is which formulas you test in: not all four, and not one at random, but the two that bracket the range.

Materials for the release test, at 1 kg oil and 10 kg wax prices with eight new-lot wicks: ₹2,222.07, excluding retained wicks and your jars. Testing every one of the four SKUs to the floor instead would roughly double that and the bench time, for little extra information if the bracket is chosen from real approval records (our judgement). If you're unsure which formulas bracket your range, test all four the first time and keep the records.

The CSI wick buy list and the step to production

The production step: each time the stock position (on hand + on order + in its lot check) falls to 749, order. In the worked example that's a 1,000-pack about every 5.9 weeks; with the 3.5-week effective lead time there is nearly always one pack on its way or in its check, which is normal for a shared wick. When it arrives, measure and burn it per the release test, then move it behind the old lot. Rotate the sealed safety packs into use every few months so they never become the oldest lot in the building. Argued against our own interest, twice: outside the festive run-up, you don't need the lot-failure reserve; and pooling means you need fewer wicks than four separate buffers, not more. If a single SKU is about to leave the pool (a formula moving to C1, say), recalculate before the next order rather than after.

CSI's range is short: Eco Thin (C1), Eco Thick (C2) and Cotton Braided as sized, tabbed wicks, with nothing in between and one wooden width. That makes pooling natural, because most small ranges end up on one or two wicks. It also means a failed lot has no "nearby size" to borrow, which is the reason the release test and the reserve matter. For the bigger question of buying production quantities only after every fragrance on a wick has passed, see the published post on bulk-buying one wick size.

What to buy, and what to skip

Four items for a shared-wick pool, chosen on page statements, price and stock on 7 October 2026. CSI publishes no wick, soot or lot data, so nothing is ranked on performance, and each card says when not to buy.

Skip for now: a 1,000-pack bought only as buffer; a second wick spec held "in case" without a test record; and any oil for a release test that isn't your production oil and load. A lot released in the wrong formula hasn't been released.

Frequently asked questions