My Supplier Discontinued a Critical Material — How Do I Document the Replacement and Validation so Future Employees Know Why the Formula Changed?

Why the reason for a change disappears first

When a supplier stops making a material, the business moves fast: someone finds a replacement, someone tests it, the candle goes back on sale. All of the important knowledge — why the old material had to go, which replacements were tried, which failed and why, what the tests showed, who decided — sits in messages, notebooks and one person's memory. Then that person leaves, or simply forgets, and a new employee looks at a formula that no longer matches the original launch notes. The natural instinct is to “fix” it back, or to try the candidate that was quietly rejected eighteen months ago. A discontinuation file is the defence against both.

This page is about the documentation, not the retest itself. The retest for a wax replacement is worked in full in the wax-supplier hub, the ten-fragrance discontinuation post and the twenty-fragrance qualification post; for an oil it is the discontinued-lavender hub and the three-oil blend post. What follows is how to capture the change so that the evidence and the reasoning outlive the people who produced them.

Section 2: the runway maths that sets the clock

The first number in the file decides everything after it: how long the old material lasts. The figures here are the reader's example and should be replaced with yours. Suppose the notice arrives today, 29 September 2026, with 23 kg of the old wax on the shelf, and the SKU sells 300 candles a month. At 8%, each 200 g candle takes 184.0 g of wax.

That gap is the single most important fact in the file, because it explains every compromise that follows. A future reader who sees that the candle was released after a 7-day checkpoint rather than a 14-day one needs to know it was a decision made against a 7.1-day hole, not carelessness. The choices for bridging the gap — pausing the SKU, selling from finished stock, taking pre-orders, or releasing on an early checkpoint with a watch item — are weighed in the substitution vs partial vs waiting post. Whatever you choose, the file records the choice and the reason. The cure checkpoints (48 hours, 7 days, 14 days) are common working checkpoints, not specifications.

Section 3: candidates, and why the rejects matter

List every replacement you considered, not only the one you chose. For each, write why it was screened in or out and whether it was tested. In the example, two CSI container waxes go forward to a side-by-side screen. The reasons below are the kind of reasons that belong in the file — observable facts about the listing, not claims about behaviour, because CSI publishes no melt point, hardness or burn data for any of its waxes.

Note the language in the last column. “Out of stock on the 24 September 2026 pull” is a fact with a date; it is not a claim that the product is discontinued, and the file should never turn one into the other. “Changes wax family; would need a wider retest” is a reason about test scope, not a claim that the wax is worse. Keep the reasons factual and dated, and the file stays useful even when stock and prices change. The direct-replacement post covers how to handle a supplier's claim that a new wax is a drop-in, and the 1,000-candle emergency post shows the minimum screen when the clock is shorter still.

Section 4: the retest matrix, with a reason per cell

The retest matrix records which earlier results survive the change. For a future reader, the reasons matter as much as the verdicts: a REPEAT with no reason looks like over-caution, and a KEEP with no reason looks like a shortcut. The example below is for replacing the discontinued wax with either candidate, keeping the Jasmine oil, Eco Thin C1 wick and White Matte jar unchanged.

The screening protocol the file references

Section 5 of the file does not summarise results; it points to them. So the screen itself has to produce records worth pointing to. The protocol below pours three candles on each candidate and three controls from the remaining old wax — possibly the last controls you will ever be able to make, which is a reason to keep extra unburned ones.

The safety lines are comparative because CSI publishes no numeric flame-height or jar-temperature limit. Acceptance is no worse than the approved control on every safety line, within your jar supplier's guidance and the recognised candle fire-safety standard you work to (ASTM F2417 is a commonly cited one). A candidate that throws better but runs the jar hotter fails, and the file says so in plain words.

The priced screening kit

Nine candles — three per candidate and three controls — at 200 g and 8% need 144 g of Jasmine, 151.2 g with a 5% pour-loss allowance. Each candidate wax needs 579.6 g with the same allowance, so a 1 kg bag of each. The control wax is the old stock, so it is not priced.

The oil is bought as 100 g + 50 g + 15 g (₹763.00, 165 g) rather than a 500 g bottle (₹1,961.00) because 151.2 g is all the screen needs. If Jasmine is a production oil you already stock at 1 kg, use that and the kit drops to the wax, wicks and jars. The file's section 2 should note the kit cost: it is part of what the discontinuation cost you.

Sections 6 to 9: decision, diff, cut-over, watch items

The decision section is short and exact: which candidate, pass / fail / conditional, the date, the name of the person approving, and any conditions (“released on the 7-day checkpoint; 14-day storage and end-of-life results due by [date]”). A conditional release is honest when the runway forced it; an unconditional release on partial evidence is not.

Where the file lives and who reads it

Keep the discontinuation file in the SKU's folder, in front of the batch records, with the retained control candles boxed and labelled alongside. Give it a number and put that number on specification v2, so anyone reading the spec can find the reason in one step. The reproducibility record covers the batch records the file refers to, the simple change-control form is the front page for smaller changes that do not need a full file, and the full change-control system shows how a discontinuation fits beside planned changes and cost-driven ones.

The file has three readers. The first is the person approving the change today, who needs sections 2, 4 and 6. The second is the person investigating a complaint next year, who needs section 8. The third is the new employee who thinks the formula looks wrong, who needs sections 1 and 3 — the reason it changed, and the list of things that were already tried. Pre-qualifying backups is how you make the next discontinuation shorter: a file that already contains a tested reserve candidate turns a crisis into a cut-over.

Frequently asked questions