Before a cosmetic product goes into full production it needs four blocks of testing: stability, packaging compatibility, microbiological safety, and any test required by the claims you intend to print. Most private-label brands discover the gap late, because an OEM quotation typically covers formulation, filling and components but rarely covers the full test panel by default. The practical move is to ask your manufacturer, in writing, which tests are included in the quoted price, which are quoted separately, and who owns the resulting reports — then schedule the slow tests first, because accelerated stability and preservative challenge testing run on fixed timelines that no amount of money will shorten.
Key takeaways
- Testing is a schedule problem before it is a cost problem. A challenge test takes about 28 days and accelerated stability about 12 weeks; run in sequence rather than parallel, those two alone can add four months to a launch that was quoted as “90 days to delivery”.
- Stability and compatibility are different tests and passing one does not imply the other. A formula can be chemically stable in a glass jar and still discolour, leak or lose fragrance in the PET bottle you actually intend to sell.
- Claims create test obligations. “SPF 30”, “dermatologically tested”, “suitable for sensitive skin” and “non-comedogenic” each require specific evidence on file before the words appear on a label — and the panel study is usually the most expensive line item in the whole programme.
- Ask who owns the reports. If the test certificates sit only in your manufacturer’s file, moving production to a second factory later means repeating and re-paying for work you already funded once.
Which tests does a cosmetic product actually need before its first production run?
Every cosmetic product needs, at minimum, evidence that it stays the same over its intended life, that it survives its own packaging, that it resists microbial growth, and that it complies with the market it will be sold into — four questions answered by four different test families. Nothing on that list is optional in a regulated market; what varies is who performs the tests, whether the cost is bundled into your unit price, and how much of the paperwork you ever see.
The reason this catches first-time brand owners is structural rather than dishonest. A contract manufacturer quoting a stock or lightly modified formula has usually already tested the base formula, so it prices the job as filling plus components. The moment you change the fragrance, the colour, the active percentage or the container, some of that inherited data stops applying to your product. If you are working from a bespoke formula rather than an off-the-shelf base, read the development lifecycle in our explainer on custom formulation in cosmetics manufacturing first — the test panel below sits inside that lifecycle, not after it.
The core test panel
The table below is an original planning aid built from typical private-label programmes. Durations are the elapsed time the test itself occupies, not lab queue time, and costs vary widely by country, laboratory and how many variants you submit. Treat it as a list of questions to put to your manufacturer, not as a quotation.
| Test | Question it answers | Typical elapsed time | Usually arranged by | What a pass looks like |
|---|---|---|---|---|
| Accelerated stability (40 °C, plus 4 °C and cycling) | Will the formula separate, discolour, change pH or lose viscosity over its shelf life? | About 8–12 weeks | Manufacturer or formulator | Signed report with time-point data, not a one-line “stable” statement |
| Real-time stability | Does the accelerated prediction hold at ambient conditions? | Runs the full claimed shelf life; starts now, reports later | Manufacturer | Ongoing study opened before launch, with scheduled pull points |
| Packaging compatibility | Does the formula attack, stain, swell or leak from the chosen container and closure? | Runs alongside stability, in the final component | Manufacturer, with component supplier input | Weight-loss data, no deformation, closure torque retained, no fragrance scalping |
| Preservative efficacy (challenge) test | Does the preservative system control contamination introduced in normal use? | About 28 days | External microbiology laboratory | Log-reduction criteria met for the specified organisms |
| Microbiological content of the finished batch | Is the produced batch itself within total viable count and pathogen limits? | About 3–7 days per batch | Manufacturer QC, per batch | Result on the batch certificate of analysis, not a one-off study |
| Heavy metals and trace contaminants | Are lead, arsenic, mercury, cadmium and similar impurities within the destination market’s limits? | About 1–3 weeks | External analytical laboratory | Certificate naming the method and the limit applied |
| Safety assessment or safety report | Has a qualified assessor concluded the finished product is safe as formulated and packaged? | About 2–6 weeks once the data pack is complete | Safety assessor or regulatory consultant | Signed assessment referencing your actual formula, packaging and use pattern |
Two of those rows are frequently confused, so it is worth separating them: a challenge test is a one-off study proving the formula can defend itself, while microbiological testing of each production batch proves the factory did not contaminate that particular run. Passing the first does not remove the need for the second. Where those batch results should appear is covered in our walk-through on reading a cosmetic spec sheet and certificate of analysis.
The safety assessment row is the one that ties everything together. In the European Union and Great Britain a signed Cosmetic Product Safety Report is a legal precondition for placing a product on the market, and the assessor draws on the stability, challenge and contaminant data listed above. If any of those inputs is missing, the assessor cannot sign — which is why a brand that skipped testing in month two often stalls in month six. Our article on what a Cosmetic Product Safety Report contains sets out exactly which line items the assessor expects the manufacturer to supply.
What is the difference between stability, compatibility and challenge testing?
Stability asks whether the formula changes over time, compatibility asks whether the formula and its container damage each other, and challenge testing asks whether the preservative system survives repeated contamination — three separate failure modes that need three separate studies. Running only one of them and assuming the others is the most common technical gap in first-time private-label launches.
Stability testing places samples at controlled temperatures — commonly 40 °C for acceleration, 4 °C for cold stress, plus freeze–thaw cycling and sometimes light exposure — and records appearance, odour, colour, pH, viscosity and phase behaviour at fixed pull points. Three months at 40 °C is conventionally read as a rough proxy for a longer ambient life, but that is an industry rule of thumb, not a validated conversion, which is why a real-time study should be opened in parallel rather than skipped. The mechanics are covered in more depth in our explainer on stability testing for cosmetic products.
Compatibility testing is stability run in the actual sales pack rather than a laboratory jar, and it exists because the pack is a chemical participant, not a neutral box. The failures it catches are physical and unglamorous: an essential oil softening or crazing a polystyrene cap, a low-pH serum corroding a metallised pump spring, weight loss through a permeable LDPE tube, fragrance being absorbed into the plastic wall so the product smells weaker at month nine, or a hot-filled balm deforming a thin-walled jar. None of these show up in a glass-jar stability study, and all of them show up in customer complaints.
Preservative efficacy testing, usually run to ISO 11930, inoculates the product with specified bacteria, yeast and mould and measures how far the counts fall over 28 days. It is required for any water-containing product a consumer will dip a finger into, and it must be repeated whenever the formula, the pH or the pack format meaningfully changes — moving from a pump bottle to an open jar changes the contamination risk even if the formula is identical. Some anhydrous products, high-alcohol formats and extreme-pH products may qualify as microbiologically low-risk under ISO 29621, but that determination is a documented risk assessment by a competent person, not a decision a brand owner can make to save money.
What changes force a re-test?
The practical trigger list is shorter than most brands expect and longer than most factories volunteer:
- Any change to the pack — different resin, different supplier for the same resin, a new closure, a switch from opaque to clear — restarts compatibility, and clear packaging may add a light-stability requirement.
- Fragrance or colour changes restart stability, because both can shift pH, colour drift and preservative efficacy.
- A change of preservative system or supplier restarts the challenge test. So does a pH change of more than a few tenths.
- A different fill volume or a scale-up in batch size can change heat history, mixing shear and cooling rate, which is why the first commercial-scale batch deserves its own retained sample even when the pilot passed.
The cheapest place to discover all of this is a sample batch, not a production run. Our guide on ordering a sample batch from a private-label factory explains how to specify that batch so it is actually usable as a test article rather than a marketing sample.
Which tests are only required because of the claims you want to make?
Claim-driven tests are the ones you can still avoid at this stage, because they are triggered by the words you choose to print rather than by the product itself. Every performance, tolerance or efficacy claim on your label and in your advertising needs evidence on file before publication — and in the European Union that principle is a legal requirement, not a best practice, under the common criteria for cosmetic claims.
This is where launch budgets break. A brand plans a serum, prices the fill and the bottle, then decides late in the process that the pack should say “clinically proven to hydrate for 24 hours” — a sentence that costs more to substantiate than the entire first production run. Deciding the claim set before the artwork brief goes out is the single most effective cost control available.
| Claim you want to print | Evidence usually expected | Test type | Indicative elapsed time |
|---|---|---|---|
| “SPF 30” or any numeric sun protection factor | In-vivo SPF determination on a human panel, plus a UVA method | Human panel study | About 6–12 weeks |
| “Dermatologically tested” | A dermatologist-supervised skin tolerance study, typically patch testing | Human panel study | About 4–8 weeks |
| “Hypoallergenic” or “suitable for sensitive skin” | Repeat-insult patch testing (HRIPT) plus a documented formulation rationale | Human panel study | About 8–12 weeks |
| “Non-comedogenic” | A comedogenicity study, or a defensible ingredient-level rationale where accepted | Human panel or documented rationale | About 4–10 weeks |
| “Ophthalmologically tested” (eye area) | Ocular tolerance study under supervision | Human panel study | About 4–8 weeks |
| “Reduces the appearance of fine lines” and similar efficacy claims | Instrumental measurement or a controlled consumer-perception study on the finished product | Clinical or consumer study | About 8–16 weeks |
| “95% natural origin” or a certified organic mark | Calculation to a recognised standard, or a certification audit | Documentation or certification | Weeks to months, depending on scheme |
| “Vegan” or “not tested on animals” | Supplier declarations across the full ingredient list, held on file | Documentation | Days to weeks |
Three practical rules fall out of that table. First, a study run on the base formula by your manufacturer is not automatically a study on your product — ask whether the report names your formula code and your pack. Second, an ingredient-level claim does not carry to the finished product: a supplier’s data showing an active performs at 5% does not entitle you to claim the result if your formula contains 0.5%. Third, marketing copy and social captions count as claims in most jurisdictions, so a claim you deliberately kept off the carton can still create an obligation when it appears in an advertisement. Our sourcing piece on sensitive-skin claims in cosmetics OEM sourcing works through what a factory can and cannot support in this area.
How do you sequence testing so it does not delay your launch?
Start the long-duration tests as early as the formula and pack are frozen, and treat everything else as work that fits around them — because challenge testing, accelerated stability and human panel studies run on fixed calendars that cannot be shortened by paying a rush fee. The sequence below is the one that keeps most launches on schedule.
| Stage | What happens | What starts here | Who owns it | Common failure |
|---|---|---|---|---|
| 1. Formula sign-off | Bench sample approved on texture, scent, colour and performance | Nothing yet — testing on an unfrozen formula is wasted money | Brand | Approving “almost final” and tweaking later |
| 2. Pack sign-off | Bottle, closure, liner and decoration method locked with the actual supplier | Compatibility study in the real component | Brand and manufacturer | Testing in a lookalike sample pack, not the production part |
| 3. Pilot batch | Small batch produced on representative equipment | Stability, challenge test, contaminant screen, retained samples | Manufacturer | Filling from a beaker rather than the real line |
| 4. Claim studies | Panel or instrumental studies commissioned | Only the claims already committed to artwork | Brand, via the laboratory | Commissioning after artwork is printed |
| 5. Data pack assembly | All reports collated for the safety assessor | Safety assessment or safety report | Assessor or consultant | Discovering a missing certificate at this point |
| 6. Notification or registration | Product notified to the relevant authority before sale | Market entry paperwork | Responsible person or brand | Booking freight before the notification clears |
| 7. Production batch | Full run, released against agreed specifications | Per-batch QC and certificate of analysis | Manufacturer QC | No written release criteria, so “pass” is undefined |
Stage 6 deserves emphasis because it is a hard gate rather than a soft one. Most markets require notification or registration before the product is offered for sale, and the submission draws on the same documents the testing programme produced; our walk-through on notifying a cosmetic product covers what that submission asks for. Similarly, the release criteria in stage 7 should be written down before the first batch runs, not negotiated after it — the practices in our guide to quality control in cosmetics manufacturing describe what a defined release specification looks like.
Questions to ask your manufacturer before you sign
- Which of these tests are included in the quoted unit price, and which are billed separately? Ask for the answer per test, not as a single yes.
- Is the stability data for my exact formula and my exact pack, or for the base formula? Request the formula code and pack description printed on the report.
- Who holds the original reports, and will I receive copies I can give to a safety assessor or a second factory? Ownership of test data belongs in the supply agreement.
- Which external laboratories do you use, and are they accredited for these methods? A named laboratory and method reference is a stronger answer than a certificate with a logo.
- What are the written release criteria for each batch, and what happens when a batch fails them? Silence here becomes an argument later.
- Are retained samples kept, for how long, and can I request one? Retained samples are what settle a customer complaint eighteen months from now.
If a factory cannot answer those six questions in writing, that is itself a finding. The broader verification routine — certifications, documents and factory evidence — is set out in our guide to cosmetic GMP and ISO 22716 for brand owners, and you can compare how different manufacturers present this information in the OEM manufacturer directory before shortlisting.
Red flags in a testing conversation
Four answers should slow you down. A quotation that promises delivery in a timeframe shorter than the tests it implies means testing was either omitted or inherited from another product. A verbal assurance that a formula is “already tested” without a report that names your pack means compatibility was probably never run. A refusal to share test reports on confidentiality grounds is workable for the underlying formula but not for the safety and compliance data you are legally responsible for. And an offer to skip the challenge test because the product is “natural” inverts the risk — low-preservative and botanical-rich formulas usually need more microbiological scrutiny, not less. Once the reports are in hand, work through our pre-launch checklist for a private-label product and the packaging and label requirements checklist before releasing artwork to print.
Frequently asked questions
Does my OEM manufacturer automatically do all of this testing?
No, and assuming so is the most common gap we see in first launches. Most contract manufacturers include batch-level quality control and will have stability data on their standard base formulas, but claim studies, contaminant screening and the finished-product safety assessment are frequently quoted separately or left to the brand entirely. Ask for a written list showing each test as included, quoted separately, or not offered — and get it before you sign, because the negotiating position changes once a deposit is paid.
How much should I budget for testing a first product?
There is no honest universal figure, because the range is driven almost entirely by your claims: a single unscented moisturiser with no performance claims sits at the bottom of the range, while an SPF product with a numeric factor and a sensitive-skin claim can cost several times the value of the first production run. The reliable planning method is to list your intended claims first, price the studies each one triggers, and only then decide which claims are worth keeping. Budget separately for a re-test allowance, since one changed component can restart a study.
Can I use the accelerated stability result as my shelf life?
It is the normal basis for setting a provisional shelf life or period-after-opening at launch, but it is a prediction rather than a measurement, and it should be confirmed by a real-time study running in the background. Open the real-time study at the same time as the accelerated one so that confirmation arrives before your first restock decision. If a real-time result later contradicts the prediction, the shelf life or period-after-opening on the pack has to change — which is a labelling and stock problem worth anticipating rather than discovering.
Do I need to repeat testing if I switch manufacturers?
Usually yes for the process-dependent parts, even when the formula is identical on paper. A new factory means different equipment, different mixing shear, different raw-material suppliers and a different water system, so batch microbiology, stability in the new pack and a fresh contaminant screen should be repeated. Formula-level data such as a challenge test may transfer if the formula, preservative system and pack are genuinely unchanged and you hold the original reports — which is precisely why report ownership belongs in your first contract, not your second.
Sources and limitations
Standards and regulatory references used in this article: ISO 11930:2019 for evaluation of the antimicrobial protection of a cosmetic product; ISO 29621:2017 for identifying microbiologically low-risk products; ISO 22716:2007 for cosmetics Good Manufacturing Practices; ISO 24444:2019 for in-vivo determination of the sun protection factor; the European Commission’s cosmetics legislation portal for Regulation (EC) No 1223/2009 and the common criteria for cosmetic claims; the CosIng database for ingredient nomenclature and restrictions; the Scientific Committee on Consumer Safety for safety-assessment guidance; the US FDA’s pages on product testing of cosmetics, microbiological safety and heavy-metal contaminant testing, plus MoCRA for safety substantiation and adverse-event obligations; the FDA Cosmetics Labeling Guide for expiry and durability declarations; and HSA Singapore for the ASEAN Cosmetic Directive. All links checked on 14 August 2026.
Limitations: the durations and sequences in the tables are illustrative planning ranges drawn from typical private-label programmes, not surveyed data, and they vary substantially by laboratory, country, product category and how many variants are submitted together. We did not audit any laboratory or factory for this article, and we have not priced any specific test. Which tests are legally mandatory, and to which limits, depends on the destination market and the product type — a rinse-off body wash, a leave-on eye cream and a sunscreen are treated differently in most jurisdictions. Confirm the current requirements with the relevant authority for every market you sell into, and have a qualified safety assessor review your finished data pack.
This article is general industry guidance for brand owners working with contract manufacturers and does not constitute regulatory, legal, medical or toxicological advice. Testing requirements change; verify current rules with the relevant authority and confirm the exact test panel, ownership of reports and batch release criteria directly with your manufacturer before committing to production.