K-Beauty Manufacturer Q&A: Custom Formula, Stability Testing, Claim Support (2026)

By the ALTA MEET editorial team | K-beauty ODM consulting

Most first-time indie founders sit down with a Korean ODM and walk out of the first briefing without asking three of the questions that decide whether the product ever ships on schedule. This is a founder-side Q&A you can print, mark up, and bring into that meeting. Fourteen questions across five topics: what "custom formula" really means, which stability protocol the ODM is running, when preservation efficacy tests become mandatory, how claim substantiation files need to look for MoCRA and EU 655/2013, and what paperwork you need at first sample and at ship-out. Answers reference public regulatory sources so you can verify the language yourself.

1. Custom formula scope: what "custom" actually means at a Korean ODM

The word "custom" covers three very different things at a Korean ODM. Being specific about which one you are buying prevents the most common founder-side surprise: paying for a customization tier the ODM did not include in the quote.

Q1. How custom is "custom"?

ODM (Original Design Manufacturer) generally spans three levels of formula work: (a) selecting an existing house formula and re-labeling it; (b) modifying a house formula (swap an active, adjust a texture, add a fragrance, tune preservative load); (c) developing a formula from a written brief. According to a public overview of Korean ODM models, most manufacturers provide end-to-end services from formulation development to packaging, regulatory documentation, and export preparation, but the depth of the "development" step varies dramatically by tier (Gravel AI: Top Korean Beauty Contract Manufacturers 2026). Ask which tier your quote reflects, and ask how many rounds of reformulation are included at that price before change-orders start.

Q2. Who owns the formula after production?

Ownership sits in the contract, not in intuition. In a standard ODM engagement where the ODM developed a house formula and you re-labeled it, the ODM usually retains rights to sell that formula to other brands. In a bespoke development where you commissioned the R&D and paid a development fee, the contract can grant you exclusivity for a defined window, defined territory, or defined product category. The structural question is not "do we own it" but "for what scope and for how long is anyone else prevented from selling something identical." Read the exclusivity clause line by line before you commit to volume. For a broader founder-side checklist on the pre-PO stage, see our pre-PO due diligence guide.

Q3. What should you bring to the first briefing?

The briefing goes faster and cheaper if you show up with (i) the target INCI of two or three benchmark competitor products, (ii) a texture reference (a jar of something you like), (iii) the retail price band you are aiming at, (iv) the market you will sell in first (this drives which regulatory regime governs the formula), and (v) the packaging spec if it is already decided. Missing (iv) is the most expensive omission; the market determines allowable preservatives, active concentration ceilings, and whether the product will need functional cosmetic notification in Korea, MoCRA-aligned safety substantiation in the United States, or a Product Information File under EU Regulation 1223/2009.

2. Stability testing: what to ask before you commit to the run

Cosmetics are not drugs, but Korean ODMs frequently benchmark stability protocols against the pharmaceutical guideline ICH Q1A(R2), because the ICH time-point structure gives the ODM and the founder a shared vocabulary for reading results. Understanding that vocabulary before the study starts saves a reformulation cycle later.

Q4. Which stability protocol should the ODM run?

Ask for the study design in writing, referencing the applicable guideline. The pharmaceutical benchmark ICH Q1A(R2) is published on the FDA guidance page and mirrored by the EMA. The guideline defines three parallel storage arms: a long-term arm at 25 °C (relative humidity 60 %), an intermediate arm at 30 °C (RH 65 %), and an accelerated arm at 40 °C (RH 75 %), each with the standard tolerances published in the ICH document (FDA Q1A(R2)). Time points on the long-term arm run at 0, 3, 6, 9, 12, 18, and 24 months, and on the accelerated arm at 0, 3, and 6 months. Cosmetic stability programs commonly adapt these arms with the same accelerated condition but shorter reads. If the ODM proposes something different, ask what the deviation is and why.

Q5. What does the accelerated study actually predict?

Accelerated stability is a compressed rehearsal, not a shelf-life guarantee. The rule of thumb Korean ODMs cite is that stability for six months at 40 °C generally supports a claim of approximately 24 months at room temperature, provided nothing anomalous appears in appearance, assay, pH, or microbiological attributes at the intermediate reads (FDA Q1A(R2) guidance). If your quote lists a shelf-life claim on the label before the six-month accelerated read is complete, ask what data supports the claim. A quiet caveat: pigmented color cosmetics and emulsions with novel actives frequently fail the accelerated arm without failing long-term; do not assume a delayed launch is the ODM's fault when accelerated data comes back off-spec.

Q6. What time points do you need before ship-out?

Ask for the minimum data package required to release the first production lot. A defensible package typically includes: T0 baseline, one-month accelerated read, three-month accelerated read, and any freeze-thaw or light-exposure sub-studies specific to the format. Ampoules and oil-heavy serums often carry additional oxidative reads. For a walk-through of how to read the ICH and ISO reports the ODM will hand you, see our guide to stability testing reports.

3. Preservation and microbiology: when ISO 11930 becomes non-negotiable

Water-based cosmetics require a preservation system that will hold up under real consumer use, not just in the lab. ISO 11930 is the standard test method for evaluating that system. It is not optional in most launch scenarios.

Q7. When is ISO 11930 mandatory?

ISO 11930 preservative efficacy testing (also called challenge testing) is expected for high-water-content formulas, products packaged in wide-mouth jars or droppers with repetitive consumer exposure, reformulations that alter pH, surfactant system, or preservative load, and products carrying "preservative-free" or "natural" claims. It is also a regulatory expectation in the EU, the UK, and several APAC markets under national implementations of good-manufacturing-practice requirements. Ask the ODM to confirm in writing that ISO 11930 is in the study plan, and ask which day-90 acceptance criterion the formula is being evaluated against.

Q8. What organisms does the challenge test cover?

The ISO 11930 method inoculates the product with a defined panel: Staphylococcus aureus, Escherichia coli, Pseudomonas aeruginosa, Candida albicans, and Aspergillus brasiliensis. The method also allows the addition of other microorganisms when the product's intended use requires it (for example, marine-water-derived actives may warrant additional environmental strains). Ask whether the ODM plans to add any strains beyond the standard five and, if so, why.

Q9. Which pass/fail criteria should you look for on the report?

Reads are taken at day 7, day 14, day 28, and, where the protocol calls for it, day 90. Log-reduction thresholds differ by organism and by whether the product qualifies as Criterion A (standard, no additional protective measures needed) or Criterion B (formulation with weaker intrinsic protection, additional risk-management measures required). Ask the ODM report author which criterion your formula meets, and ask what would move it from A to B if that changed on reformulation. A useful founder-side habit is to file the day-28 log-reduction table alongside your Product Information File so the number is retrievable when a regulator asks.

I'm Liz, I run altameet from Manhattan, NYC. Before your first Korean ODM briefing lands, the fastest thing I can do is read your target INCI, tell you which of the three "custom" tiers you are actually buying, and flag the two clauses in a standard ODM MSA that indie founders miss. If that would help you avoid a change-order in month two, I'll give you fifteen minutes free. Email liz@altameet.com.

4. Claim substantiation: what "adequate" looks like on paper for MoCRA and EU 655/2013

The claim on the label is a promise to the regulator as much as to the consumer. Two frameworks now govern what evidence has to sit behind that promise: MoCRA in the United States and EU Commission Regulation 655/2013 in Europe. They overlap in principle and diverge in mechanics.

Q10. Which claims need what evidence under MoCRA?

MoCRA (the Modernization of Cosmetics Regulation Act of 2022) explicitly requires that a responsible person for a cosmetic product ensure, and maintain records supporting, that there is adequate substantiation of safety of the product. Adequate substantiation is defined as tests, studies, research, analyses, or other evidence or information considered, among experts qualified by scientific training and experience, sufficient to support a reasonable certainty that the cosmetic product is safe (FDA: MoCRA overview). In practical terms the file typically covers ingredient-level toxicology dossiers or read-across, finished-product stability data, preservative efficacy data, packaging compatibility notes, and, for any efficacy-adjacent claim, methodology and results of a study conducted under controlled conditions.

Q11. How does EU 655/2013 differ?

EU Regulation 655/2013 sets six common criteria that every cosmetic claim, explicit or implicit, must meet: legal compliance, truthfulness, evidential support, honesty, fairness, and informed decision-making by the consumer. The evidential-support criterion states that claims for cosmetic products, whether explicit or implicit, shall be supported by adequate and verifiable evidence (Commission Regulation (EU) No 655/2013). Under Article 11 of the parent Cosmetics Regulation 1223/2009, every cosmetic product placed on the EU market must have a Product Information File containing that evidence. Where MoCRA leans on the words "reasonable certainty of safety," 655/2013 leans on "adequate and verifiable evidence" for the specific claim wording. A founder targeting both markets typically maintains one master evidence file structured to answer both questions.

Q12. What does "adequate substantiation" look like on paper?

Regulators generally look for evidence proportional to the claim. A moisturizing claim needs less than an anti-wrinkle claim, and both need less than a claim that references a specific mechanism ("increases dermal density," for example). The EU guideline associated with 655/2013 notes that ex-vivo and in-vitro tests must be conducted under standardized conditions with protocols referencing published or in-house validated methods, with clear documentation of methodology and statistical analysis, and, where used as evidence of an in-vivo effect, validated by studies on humans. When you receive an ODM's proposed claim list, cross-check whether the underlying evidence in the file actually supports the strongest words used. It usually does not, and the founder is the last line of defense before packaging goes to print.

5. Documentation flow: what paperwork you need at first sample and at ship-out

Two documentation moments matter more than any other: the file that arrives with the first sample, and the file that arrives with the first shipment. Missing either creates cascade problems at Customs, at the retailer, and at the FDA facility registration desk.

Q13. What paperwork should the ODM hand you with the first sample?

A defensible first-sample package includes: the INCI declaration in descending order, the master batch record (or its summary), the specifications sheet with pH, viscosity, and appearance targets, a preliminary stability plan, the preservation strategy including which preservative system will be run through ISO 11930, and the master claim list the ODM believes the formula supports. According to a public compliance overview, supplier obligations when evaluating Korean cosmetic manufacturers include provision of ingredient lists (INCI), safety data, and certificates as part of the standard compliance package (Source of Asia: Cosmetic Manufacturers in South Korea). If any of these are missing at the first-sample stage, ask when they will land. If they will not land until production, you are buying a sample without a spec.

Q14. What paperwork should the ODM hand you at ship-out?

The ship-out package includes the batch-specific Certificate of Analysis, the stability data supporting the printed shelf life, the microbiology report (ISO 11930 pass, plus in-process microbial counts against ISO 17516 finished-product limits), the packaging compatibility report if novel packaging was used, the batch traceability record for any regulated actives, and the safety data sheet for hazardous constituents. For MoCRA-listed products the file also needs to slot into the FDA product-listing record (Form FDA 5067) and the facility-registration record; the responsible person must list each marketed cosmetic product and update annually (FDA: Registration and Listing). Adverse events that meet the FDA definition of "serious" must be reported to FDA within fifteen business days of the responsible person becoming aware, per updated MoCRA instructions (FDA: Updated Instructions for Serious Adverse Event Reporting). None of that reporting is possible without the ship-out documentation intact, so the ship-out package is where regulator-facing risk actually crystallizes.

Key takeaways for the founder side of the briefing

Bring the fourteen questions above to the first ODM meeting and you will leave with an aligned scope, a study design, a claim list bounded by evidence, and a documentation plan. Skip them and you will discover the gaps in month two or month five when a change-order arrives or a retailer asks for a Certificate of Analysis you do not have. The regulatory frameworks you need to reconcile in 2026 are MoCRA in the United States, Regulation 1223/2009 and Regulation 655/2013 in the EU, and the MFDS Cosmetics Act in Korea (MFDS Cosmetics). Each has its own vocabulary; the ODM has heard them all before, so ask which one is driving each element of the study plan.

Reading the ODM contract clause by clause is the second habit that separates founders who ship on schedule from founders who do not. For a walkthrough of the typical Korean OEM contract structure, see our inside a Korean OEM contract explainer, and for the primer that clarifies whether you are actually in an OEM, ODM, or private-label engagement, see the OEM vs ODM vs private label primer.

Founder FAQ

Does the ODM own my formula if I paid for the samples?

Paying for samples typically covers sample production cost only, not formula ownership. Formula ownership is negotiated separately, usually as a development fee plus an exclusivity window written into the ODM master service agreement. Read the exclusivity clause and the field-of-use clause together; either one can quietly limit your rights.

Can I skip stability testing to launch faster?

Not defensibly. Shelf-life claims on the label are regulatory statements, and the file that supports them is what a regulator asks for after a complaint. Accelerated stability can compress the timeline, but it does not eliminate it. Founders who launch without at least a three-month accelerated read carry all of the recall risk personally.

Do I need ISO 11930 if my product is anhydrous?

Anhydrous products (solid balms, oil serums with no water phase) typically fall outside the scope of ISO 11930 challenge testing because there is no aqueous phase for microorganisms to proliferate in. Ask the ODM to confirm the anhydrous determination in writing, and ask what alternative microbial control the specification uses (usually tighter raw-material microbial limits and packaging integrity).

How much of the claim substantiation file does the ODM produce vs the brand?

The ODM typically produces the ingredient-level dossier, the stability data, the preservation data, and packaging compatibility. The brand typically commissions any efficacy study that supports the specific marketing claim (a moisturization panel, a wrinkle-appearance panel, a soothing panel). Some ODMs will manage the efficacy study as an add-on line on the quote; the study itself is usually run by a specialist CRO.

What happens if a serious adverse event is reported?

Under MoCRA the responsible person, defined as the manufacturer, packer, or distributor whose name appears on the product label, must file the report with FDA within fifteen business days of first receiving the information. If additional information about the event surfaces within one year of the initial report, the responsible person must submit that new information within another fifteen business days (FDA). The definition of "serious" is narrow: death, life-threatening experience, inpatient hospitalization, persistent disability, congenital anomaly, infection, significant disfigurement, or requiring medical or surgical intervention to prevent one of those.

Does Korea require any pre-market approval for functional cosmetics?

Yes. Under the MFDS Cosmetics Act, products marketed with functional claims (whitening, anti-wrinkle, sun protection, hair loss relief, and several others) require either a submission of a functional-cosmetics report or a full evaluation, depending on whether the functional ingredient, efficacy, and dosage match ingredients already notified by MFDS (MFDS Cosmetics). If they match, the report route can complete within approximately seven business days after MFDS system verification; if the ingredient or formulation is new, the full evaluation applies.

How is the ingredient list order different in Korea vs the US and EU?

Order-of-ingredient rules differ. According to a public compliance summary, K-beauty ingredient lists sometimes list "good sounding" ingredients earlier than the EU or US rules would require, because the Korean ordering convention differs (Gravel AI). If you are re-labeling a Korean-market product for US or EU sale, you almost always have to re-sort the INCI declaration to match the destination-market order-by-percentage rule.

Next step for indie founders

If you are heading into a Korean ODM briefing this quarter, work through these fourteen questions before the meeting rather than during it. Ask ALTA MEET for a fifteen-minute call at liz@altameet.com if you want a second pair of eyes on the ta