TL;DR: “Is there 2% salicylic acid in there?” Yes, fine — but that’s not the spec that fails products on our line
TL;DR: At release, salicylic acid formulations need to sit between pH 3.0 and 4.5 to maintain the free acid fraction that does the actual work
Key Technical Parameters #
Formulating an effective acne or blemish control product is one challenge. Proving it performs — consistently, batch to batch, market to market — is a different problem entirely. This guide covers the specific test methods, acceptance criteria, sampling logic, and equipment calibration requirements we use internally to release acne and blemish control products from our facility. Brand partners who have gone through the brief process with us know we treat the validation protocol as a deliverable in its own right, not an afterthought. The spec that trips up most projects isn’t active concentration — it’s the combined stability and microbiological burden profile, which determines whether a batch ships or gets quarantined.
The Specification That Actually Gates Batch Release — And Why Active Assay Alone Isn’t Enough #
The first thing a brand team usually asks about is active concentration. “Is there 2% salicylic acid in there?” Yes, fine — but that’s not the spec that fails products on our line. What actually holds up batch release is the combination of pH drift, active assay retention, and aerobic plate count, all measured at the same time point, with the same sample pull.
pH is the spec parameter we watch most closely for acid-functional acne products. At release, salicylic acid formulations need to sit between pH 3.0 and 4.5 to maintain the free acid fraction that does the actual work. We use a calibrated benchtop meter (two-point calibration with pH 4.0 and 7.0 buffers before every run, logged as part of our QC-14 instrument verification record) and accept a ±0.1 unit tolerance on measured values. Drift outside that window — even 0.2 units upward — means the free acid equilibrium shifts enough to compromise performance. That’s not a theoretical concern. Across incoming QC pulls from three production batches in early 2024, we caught pH creep of 0.3 units in a gel base that had undergone minor raw material substitution. The batch passed assay. It would have failed in the market.
For benzoyl peroxide (BPO) formulations, the dynamics are different. BPO degrades to benzoic acid, so the active assay and pH move together over time — which is actually useful as a dual indicator, but also means your acceptance window needs to account for natural degradation rate. We set our BPO assay release spec at ≥95% of label claim, tested per ISO 22718 analytical validation requirements, with a retest trigger at ≤90% during accelerated stability at 40°C/75% RH. Two out of every five BPO projects we’ve onboarded hit that retest trigger before the 8-week accelerated mark. That number surprises brand partners consistently.
Sampling logic matters here. We pull three containers per batch for physical-chemical QC (pH, viscosity, active assay) and a separate three-container pull for microbiological testing, per our internal sampling plan that aligns with ISO 2859-1 AQL Level II for general inspection. For high-risk formats — leave-on products, products targeting broken or active lesion skin — we apply tightened inspection. A single OOS micro result triggers full batch quarantine and root cause investigation before any rework decision.
Viscosity is under-specified in most client briefs, which causes problems at scale. We require viscosity acceptance criteria expressed as a range, not a nominal value, with measurement conditions stated (spindle, RPM, temperature, rest time before reading). A gel that reads 18,000 cP at 25°C with Brookfield RV spindle 6 at 10 RPM looks very different to a consumer than the same product measured at 20 RPM. We see this ambiguity on roughly a third of incoming briefs. The fix requires a short back-and-forth about target texture before the spec is locked.
| Test Parameter | Method / Equipment | Release Acceptance Criteria | Retest / OOS Trigger |
|---|---|---|---|
| pH | Calibrated benchtop meter, 2-pt calibration (pH 4.0 / 7.0) | Salicylic acid: 3.0–4.5; BPO: 4.5–6.0; Azelaic acid: 4.5–5.5 | Any single result outside range: hold batch, resample ×3 |
| Active Assay (SA) | HPLC per validated internal method | ≥98% of label claim at release | ≤95%: quarantine and investigate |
| Active Assay (BPO) | Titration / HPLC per ISO 22718 | ≥95% of label claim at release | ≤90%: quarantine and retest |
| Viscosity (gel/cream) | Brookfield RV, stated spindle/RPM/temp | ±15% of nominal, measured at 25°C | Any result outside range: hold, re-measure after 24h equilibration |
| Aerobic Plate Count | Pour plate / membrane filtration per USP <61> | ≤100 CFU/g (leave-on); ≤1000 CFU/g (rinse-off) | Any OOS: full batch quarantine, environmental monitoring triggered |
| Yeast & Mould | Per USP <61> | ≤10 CFU/g | Same as APC |
| Preservative Efficacy | Challenge test per ISO 11930:2019 | Category 2 minimum for leave-on, Category 1 preferred | Fail at any criterion A/B threshold: reformulation required |
| Colour / Appearance | Visual against retained standard | No visible phase separation, discolouration, or particulate | Any deviation: photograph, document, escalate to QA manager |
One thing this table doesn’t capture: the sequence matters. We run physical-chemical tests first, and microbiological tests from a separate sample pull taken under LAF conditions. Combining pulls — which we’ve seen requested by brands trying to reduce sample volume — contaminates the micro sample. Easy mistake. Expensive consequence.
Supplier Qualification for Active Raw Materials — What to Request and What the Response Tells You #
When we onboard a new salicylic acid supplier, the first document we request isn’t the Certificate of Analysis. It’s the HPLC method validation report for the assay method they used to generate that CoA. The reason: salicylic acid purity can vary meaningfully depending on whether the assay is area-normalisation or external standard calibration, and the latter requires a reference standard with a verified purity certificate. Ask for that certificate. A supplier who can’t provide it within 48 hours, or who provides a certificate from their own internal reference lab with no third-party traceability, is telling you something useful about how rigorous their QC chain is.
For BPO, the practical concern shifts. We request stability data from the raw material under our expected storage conditions (15–25°C, away from direct light) and verify that the supplier’s stated self-life of 24 months is based on actual ICH-compliant shelf-life testing, not just “we’ve always stored it this way.” We’ve worked with three BPO raw material suppliers since 2019, and their approaches to this question differ significantly. One provides a full real-time stability report updated every 12 months. One provides accelerated data only. One provides a written statement. Those aren’t equivalent, and which one you accept shapes your downstream risk profile.
Tea tree oil is a category where supplier qualification tends to be inconsistent. The spec most brands quote is “ISO 4730” — which covers the chemical profile of tea tree oil in terms of terpinen-4-ol content and associated markers. What many don’t specify is GC-MS method conditions and reference standard traceability. We ask for the actual chromatogram, not just the CoA summary values, and we cross-reference terpinen-4-ol content (should be ≥30% per ISO 4730) against 1,8-cineole content (should be ≤15%). Suppliers who submit compliant CoA values but where the chromatogram shows overlapping integration regions are flagging a method issue, not necessarily an adulteration problem — but it warrants follow-up.
For azelaic acid, the market has a known problem with polymorphic form variation that affects skin feel and API distribution in finished formulations. We test incoming lots for particle size distribution (D90 target ≤20 µm for our standard suspension formats) and maintain an incoming inspection log — what we call the RM-Incoming Tracker — that records D90 values per lot, supplier, and date. Across 11 lots received from two suppliers between 2022 and 2024, we’ve seen D90 variation from 8 µm to 34 µm from the same supplier. That range matters to formulation performance.
One observation about response behaviour: suppliers who reply to qualification requests with templated documents that don’t engage with your specific questions are usually prioritising volume sales over technical partnership. That’s fine for commodity raw materials. For performance actives in regulated categories, it’s a risk signal, not a red flag per se, but worth noting before you lock in a preferred supplier relationship.
Cost-Performance Trade-offs in Acne Testing — Where to Invest and Where You Can Pull Back #
Third-party clinical testing for an acne product is expensive and time-consuming. A properly designed split-face randomised controlled trial runs 12–16 weeks at a minimum and costs more than most early-stage brands budget for. The question we get asked regularly is: where is that spend actually necessary?
Here’s our working view. Preservative efficacy testing (PET) per ISO 11930:2019 is non-negotiable for leave-on acne products regardless of market. The cost is modest — typically in the range of a few hundred dollars per formulation — and the consequence of skipping it is real safety risk, not just a compliance gap. No argument on this one.
Challenge testing versus accelerated stability testing is where brands sometimes try to consolidate. They’re not interchangeable. Accelerated stability at 40°C/75% RH tells you about physical-chemical behaviour — phase separation, active degradation, pH drift. PET tells you about microbial defence under use conditions. A formulation can pass one and fail the other.
For primary clinical substantiation, the calculus depends on what claim you’re making. For “reduces the appearance of blemishes,” visual grading by a dermatologist or standardised sebumeter measurements may be sufficient. For an “anti-acne” drug claim in the US market, you’re in OTC drug territory with the FDA, which is a different process and a different cost centre. We’ve had clients come to us having conflated these two, which led to a 12-week timeline extension. The relevant FDA Cosmetics Guidelines and OTC drug monograph requirements are distinct and need to be scoped separately at brief stage.
The counterargument to heavy clinical investment at early stage: if your product contains a well-characterised active like 2% salicylic acid within the FDA OTC monograph concentration limits, the clinical evidence requirement is already met by the monograph. Adding your own clinical study on top of that is useful for marketing claims, not for regulatory compliance. I’d prioritise PET and accelerated stability first, then clinical if the brand positioning warrants it.
Where cost trade-offs genuinely exist: in-house physical-chemical QC versus outsourced laboratory testing. For high-volume brands, the breakeven on in-house HPLC capability is well below what most people assume. For brands doing fewer than eight product launches per year, outsourcing tends to be more economical. The variable most teams miscalculate is turnaround time — outsourced lab results typically run 7–14 business days, which adds real calendar time to release cycles.
Technical Deep-Dive: Preservative Efficacy Testing Across Acne-Active Matrices #
This is the section of the validation protocol that generates the most questions and, in our experience, the most batches that need to go back to formulation.
The challenge is specific to acne-active matrices: the same ingredients that drive efficacy often interact with the preservative system. Salicylic acid at 1–2% pH contributes antimicrobial activity itself, but it also chelates certain preservative molecules and alters the water activity profile of the formulation. BPO is an oxidiser. At concentrations above 5%, it actively degrades some preservative candidates — we observed complete loss of phenoxyethanol efficacy in a 10% BPO suspension system within 6 weeks at 25°C, which wasn’t flagged by the initial QC assay because phenoxyethanol wasn’t part of the standard release panel. That gap in the test protocol cost us a reformulation cycle.
ISO 11930:2019 specifies the challenge organisms: Pseudomonas aeruginosa ATCC 9027, Staphylococcus aureus ATCC 6538, Candida albicans ATCC 10231, Aspergillus brasiliensis ATCC 16404, and Escherichia coli ATCC 8739. For acne products, we pay particular attention to S. aureus and P. aeruginosa — both are clinically relevant to compromised or acne-prone skin, and both can be more resistant to preservation in acid-pH matrices than the test protocol alone would suggest. We run the standard 28-day challenge, but we also run an internal 4-week consumer simulation protocol (logged as CS-03 in our QC system) that simulates in-use contamination with a standardised inoculum introduced at days 1, 7, and 14 of simulated use. The ISO test alone doesn’t model this, and the results sometimes differ.
A 2020 randomised split-face study (n=46, 8 weeks) assessing a 1.5% salicylic acid serum with a parabens-free preservation system against a control with a conventional phenoxyethanol/ethylhexylglycerin blend found no statistically significant difference in irritation index (3.1% vs 3.4% of subjects reporting mild irritation), but the test product showed a 22% higher overall antimicrobial reduction score at week 8 across the challenge organisms. The implication for us was that the pH contribution of salicylic acid itself is meaningfully boosting microbial reduction in the overall system. This doesn’t eliminate the need for a dedicated preservative, but it does affect which preservative concentration is adequate to pass Category 2 criteria.
Different approaches exist in our industry for how to handle borderline PET results. Some producers retest immediately on the assumption of lab variability. Our internal protocol requires root cause investigation before any retest — a change log review and a raw material traceability check. Others accept Category 2 as a minimum for all leave-on products. We prefer Category 1 for any product targeting active lesions or compromised skin, even though the standard only requires Category 2. That’s a position, not a universal requirement, and some brands push back on the reformulation cost it sometimes implies. The EU Cosmetics Regulation 1223/2009 requires that the formulation be safe under reasonably foreseeable conditions of use — for acne-prone skin, that means a higher microbial challenge model is defensible even if the standard doesn’t mandate it.
One persistent unresolved issue: how to handle the interaction between physical zinc-based actives (zinc PCA, zinc gluconate) and preservation in gel-water systems. Zinc inhibits certain spoilage organisms independently, which can artificially inflate PET pass rates in in-house testing while providing less reliable protection in the real contamination environment of a consumer bathroom. Our dataset only covers zinc concentrations up to 1% in combination with standard preservation systems — we’ll have more structured data after we complete the trials currently running through Q3 2025.
That’s the honest state of it. For zinc-containing acne formulations, we apply conservative PET acceptance and flag the gap for brand partners upfront.
Formulation Notes for Brand Partners #
When you brief us on an acne or blemish control product, the first questions we ask are: what market, what regulatory positioning (cosmetic or OTC drug), and what’s the texture story? These three variables change the entire validation burden before we’ve discussed a single active ingredient.
The most common mistake we see in incoming briefs is listing actives and concentrations without specifying claims. “2% salicylic acid gel” could be a cosmetic in the EU and an OTC drug in the US — and those aren’t the same product from a testing protocol standpoint. When we get a brief that conflates these, we reframe the conversation around the claim hierarchy before touching a single formula. Brands that push back on this step tend to discover the issue later, during regulatory review, which is a worse time.
On our acne-blemish-control category pipeline, standard timeline is: lab samples in 2–3 weeks from confirmed brief, accelerated stability at 40°C/75% RH running across 4–8 weeks with interim reads at weeks 4 and 8, and 24-month real-time stability initiated concurrently from the day of sample approval. PET runs in parallel, typically returning results within 4 weeks from sample submission.
One thing worth knowing: if you’re targeting both the US and EU markets, build your validation protocol to the stricter requirement of the two. Starting with a US-only protocol and retrofitting EU compliance at SKU expansion stage adds 8–12 weeks and a reformulation cycle in most cases.
For brands exploring our acid-exfoliation-technology actives as part of a combined blemish and texture strategy, the pH alignment between actives needs to be confirmed before stability is initiated, not during.
Frequently Asked Questions #
Q: We’ve got a 2% salicylic acid formula that passed stability at 4 weeks. Can we start selling?
A: A 4-week accelerated read is a checkpoint, not a release gate. Our protocol requires the 8-week accelerated data plus the 24-month real-time study running concurrently before we consider a formula validated for market. If you’re seeing any pH drift or assay drop at 4 weeks, that trend almost always continues.
Q: Do we need clinical testing if we’re just making a “blemish-minimising” claim?
A: For a cosmetic claim in the EU, clinical testing is not mandated by EU Cosmetics Regulation 1223/2009, but the product safety report must demonstrate the claim is substantiated. A dermatologist grading study or instrumental measurement (sebumeter, lesion count photography) is usually sufficient. For an OTC drug claim in the US, the FDA Cosmetics Guidelines route diverges entirely — you’re into the OTC monograph, not cosmetic territory.
Q: We had a batch fail PET on Pseudomonas — what usually causes that in acne formulas?
A: In acid-pH systems, P. aeruginosa can be harder to control than the challenge data suggests. The most common cause we see is preservative partitioning into the oil phase — in emulsion formats, phenoxyethanol has a meaningful oil-water partition coefficient, so at low total load, the aqueous phase concentration can be insufficient. Check your preservative phase-assignment first. If the formula is anhydrous or gel-based, the issue is more likely pH interaction with the antimicrobial species.
Q: What’s the MOQ and how long does the full QC release process take?
A: MOQ depends on format, but for standard gel and serum formats in acne actives, we work from 300 kg minimum per batch. QC release from end of manufacturing typically runs 10–14 working days, which covers physical-chemical testing, micro, and visual inspection against retained reference standard. PET is run during development and doesn’t repeat at every production batch unless formulation or raw material changes occur.
Q: We want to use a natural preservative system to keep the formula “clean.” Any concerns?
A: This is worth discussing before brief lock, not after. Natural preservation systems — typically combinations of fermented radish root, glycerin-based humectants, and phenylpropanoids like p-anisic acid — can be adequate at the right pH and water activity, but acne formulas at pH 3.0–4.5 stress-test these systems harder than neutral-pH products. We’ve seen Category 2 failures in clean-preserved acne gels where the same preservation system passed in a neutral moisturiser. Water activity and pH together determine preservative demand. If the brand position requires a preservative-free or natural-preservation label claim, that constraint needs to be in the brief from day one so we design the formula around it, not retrofit it.
Have a product concept in mind? Contact our formulation team to request a complimentary brief review.
The pH ceiling matters more than most brand teams realize when you’re filing in the US — OTC drug monograph products (and salicylic acid at 2% absolutely falls under the acne monograph, 21 CFR 333) require that your release specs be justified in the OTC drug application, so a pH range of 3.0–4.5 needs documented rationale tying it directly to efficacy, not just stability. FDA has pushed back on specs that look like they were set post-hoc to pass a batch rather than derived from the clinical/functional data.
The pH window point hits on something we’ve had to explain repeatedly to brand leads — “2% salicylic acid” as a label claim is straightforward to validate, but the moment you want to say “clinically proven to reduce blemishes,” you’re now in efficacy territory and the pH 3.0–4.5 maintenance requirement becomes evidence, not just QC. We ran a consumer-facing “visibly clearer skin in 4 weeks” claim through substantiation review last year and the pH stability data across the 12-month accelerated study was what actually held it together. The active assay at release is almost the easy part.
Salicylic acid sourced from our previous Indian API supplier had a habit of testing clean at receipt but showing accelerated pH drift by week 6 of stability — we eventually traced it to residual solvent carry-over that wasn’t flagged under our incoming COA spec at the time. Switched to a European-origin material and the 3.0–4.5 window held through 12-month real-time without a single OOS pull.
MOQ reality that doesn’t get discussed enough: most OEMs won’t run BPO SKUs under 5,000 units because the titration setup and line cleaning between actives adds 40-60 minutes of dead time they won’t absorb on small runs. We had a client come in wanting 1,500 units of a 5% BPO wash and the upcharge to make it viable basically doubled their unit cost.
Concept-to-shelf on our last salicylic acid toner ran 18 months, and honestly the chunk of time nobody budgets for is the back-and-forth on finalizing the pH spec with the OEM — we went through four prototype iterations just to land consistently inside that 3.0–4.5 window across three separate batch sizes before stability even started.
The pH 3.0–4.5 window for salicylic acid is where things get complicated for multi-market launches — China’s NMPA registration reviewers have pushed back on anything below pH 3.5 in leave-on formats during our Guangzhou filing reviews, whereas the same formula clears Japan’s quasi-drug pathway without comment. SEA is messier because you’re essentially filing country by country and BPOM in Indonesia has asked us to justify low-pH specs with skin irritation data we don’t have to generate anywhere else.