A stability test is simple to describe. Take jars from one batch, in the package you sell it in, keep them in different places, some warm, some cold, some on an ordinary shelf, and on set dates write down what you see and measure. The report is the record of that: what was tested, where it was kept, what was checked, what changed, and what the maker concluded. Its value is in the numbers taken on day one, because "it looks a bit yellow" means nothing without the first look to compare with. Below is a report for a made-up maker's hand lotion at week 8, what each part is for, and the file. The storage temperatures, dates and limits are this maker's own. A protocol for a product sold widely is one to set with a lab or a cosmetic chemist.
The short version
- Test one named batch, in its real package.
- Keep jars in more than one place, and write down where.
- Check on dates set at the start, not when you remember.
- Look, smell, and measure: pH and the weight of the jar.
- Compare every check with the first one.
- End with a conclusion in plain words, signed and dated.
- TaroStack records which formula version each batch was made to. More below.
The top of the report
| Product | Shea hand lotion |
| Batch | L-1102 |
| Formula | Version 4 |
| Package | 4 oz jar, lined cap |
| Started | Sep 7, 2026 |
| Storage | Room, 70°F. Warm, 104°F. Fridge, 39°F |
| Checks | Weeks 0, 2, 4, 8 and 12 |
| Looked for | Appearance, color, odor, pH, jar weight, separation |
| This maker's limits | pH moved more than 0.5. More than 2% of weight lost. Any separation |
| Checked by | LK |
The package is part of the test. The same lotion in a pump bottle is a different test, because how much water a package lets out, and how much air it lets in, is the package's doing.
The checks, at week 8
| Check date | Week | Storage | Appearance | Color | Odor | pH | pH change | Jar weight | Weight lost | Separation? | Result |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Sep 7 | 0 | Room, 70°F | Smooth white cream | White | Light shea | 5.6 | 0.0 | 118.4 g | 0.0% | No | OK |
| Sep 7 | 0 | Warm, 104°F | Smooth white cream | White | Light shea | 5.6 | 0.0 | 118.6 g | 0.0% | No | OK |
| Sep 7 | 0 | Fridge, 39°F | Smooth white cream | White | Light shea | 5.6 | 0.0 | 118.2 g | 0.0% | No | OK |
| Oct 5 | 4 | Room, 70°F | Smooth white cream | White | Light shea | 5.6 | 0.0 | 118.3 g | 0.1% | No | OK |
| Oct 5 | 4 | Warm, 104°F | Slightly thinner | White | Light shea | 5.4 | −0.2 | 117.5 g | 0.9% | No | OK |
| Oct 5 | 4 | Fridge, 39°F | Thicker cold, normal once warm | White | Light shea | 5.6 | 0.0 | 118.2 g | 0.0% | No | OK |
| Nov 2 | 8 | Room, 70°F | Smooth white cream | White | Light shea | 5.5 | −0.1 | 118.2 g | 0.2% | No | OK |
| Nov 2 | 8 | Warm, 104°F | Thin, oil layer on top | Slight yellow | Faintly sharp | 5.0 | −0.6 | 116.0 g | 2.2% | Yes | LOOK |
| Nov 2 | 8 | Fridge, 39°F | Thicker cold, normal once warm | White | Light shea | 5.6 | 0.0 | 118.1 g | 0.1% | No | OK |
The week 2 rows are in the file. Each change is measured from the same jar's first check.
What the record shows
The warm jar failed at week 8. Its pH has fallen from 5.6 to 5.0, it has lost 2.2% of its weight, the color and smell have changed, and there's oil on top. Any one of those is past this maker's limit.
It was heading there at week 4. The pH was down 0.2, the cream was thinner and the jar had lost 0.9%. None of those was over a limit, so the result said OK. The trend was in the numbers a month before the failure was in the jar.
The room and fridge jars are unchanged. Eight weeks on a shelf at 70°F has moved the pH by 0.1.
The weight says something about the jar. The warm jar lost 2.6 g, most likely water leaving through the cap. A tighter liner is a packaging fix, separate from the formula.
The conclusion
The report ends with what the maker decided, in words anyone can read:
Batch L-1102, formula version 4. At week 8 the lotion is unchanged at room temperature and in the fridge. Kept at 104°F it thinned by week 4 and separated by week 8. Actions: test a cap with a tighter liner; trial a higher level of emulsifier as version 5 and put it on the same test; continue the room temperature jars to 12 months. LK, Nov 2, 2026.
What the record doesn't do is turn eight warm weeks into a number of months on a shelf. That's a judgment, and for a product sold through shops it's one to make with someone qualified to make it.
What this test isn't
It isn't a test of the preservative. Whether a formula holds off bacteria and mold is a separate test done by a lab: see preservatives for cosmetics. It also isn't the batch record, which says how this batch was made: see cosmetic batch record template. How the two fit into a maker's wider records is in cGMP for cosmetics, and what goes on the jar afterward in lotion shelf life and batch codes for cosmetics.
Get the template
Download the cosmetic stability test report. One sheet and a read-me: a row for each check of each jar, with the pH change and the weight lost worked out from that storage's first row, and LOOK against any row past the limits. Change the limits in the formula to your own, and write the product, batch, formula and package at the top of your copy. It's an ordinary Excel file with no macros, and it works in Google Sheets, LibreOffice and Numbers.
Where this stops working
The report is about one batch of one version. A maker who changes the emulsifier in January has a new formula and an old report. Six months later a customer writes about a jar that separated, and the first question is which formula was in it. The stability sheet can't say. Nor can it say which shops got the batch that was tested, or any other.
How TaroStack does it
TaroStack keeps the formula's versions and which batch was made to which.
Recipes are by weight, with versions. A changed emulsifier is a new version, and every batch records which one it used.
Every batch is a lot with a date, and it notes the lots of the ingredients it used, automatically.
Your checks are fields on the batch record: pH at filling, appearance, who signed.
What ships is recorded by lot, so one report shows which customers got a batch and what's left.
Starting a recall holds the lot that's still in stock.
| The question | Stability sheet | In TaroStack |
|---|---|---|
| Which formula version is in this jar | Not known | On the batch |
| Which ingredient lots went into it | The batch record, if kept | Noted automatically |
| Which shops got batch L-1102 | Read the invoices | One report |
| pH at filling, batch by batch | A notebook | Fields on each batch |
TaroStack doesn't run or schedule stability tests: keep the report for that. Recipes, batches and lots are on every plan, from $9.99 a month (5 recipes) or $39 a month (50 recipes). The first 14 days are free, with no credit card. If you'd like a hand setting up, ask, and we'll do it with you.
