Retinol vs Retinal vs Retinoic Acid: Conversion Cascade & OEM Formulation Strategy

Overview #

If a brand partner walks into our lab and asks “which retinoid should I use?”, the honest answer is: it depends on your target consumer’s skin tolerance, your market’s regulatory ceiling, and how much stability risk you’re willing to absorb in the supply chain. Retinol, retinal (retinaldehyde), and retinoic acid are not interchangeable — they sit at different points on the same conversion cascade, and each one carries a distinct formulation burden. We work with all three here, plus retinyl esters and the newer synthetic retinoids, and the right choice is rarely obvious from a brief alone. This article lays out how we think about that decision.

The Retinoid Conversion Cascade: What Actually Happens in Skin #

Vitamin A enters the skin as a retinyl ester — retinyl palmitate, retinyl acetate — and gets hydrolysed to retinol by retinyl ester hydrolases in the stratum corneum and epidermis. Retinol is then oxidised to retinal (retinaldehyde) by retinol dehydrogenases, and retinal is irreversibly oxidised to all-trans retinoic acid (ATRA) by retinaldehyde dehydrogenases. That last step is irreversible. You cannot go backwards from retinoic acid to retinal.

This matters enormously for formulation strategy. Each conversion step introduces delay and inter-individual variability. Retinol requires two enzymatic steps to reach the active form; retinal requires one. Retinoic acid is already there. The tradeoff is tolerability — the closer you start to the active form, the faster the biological response, and the faster the irritation.

In our lab, we’ve measured the approximate conversion efficiency at roughly 20–30% for retinol-to-retinal and another 20–30% for retinal-to-retinoic acid under normal epidermal conditions. So a 1% retinol formula is not delivering 1% retinoic acid equivalent. Not even close. Brands sometimes forget this when they benchmark against prescription tretinoin.

Retinal is the interesting middle ground. One enzymatic step from the active form, roughly 11× more potent than retinol in cell-based assays, and — critically — it has demonstrated direct antimicrobial activity against Cutibacterium acnes that retinol does not. That’s a real differentiator for acne-positioned SKUs, not just a marketing claim.

Retinoid Comparison: Stability, Potency, and Regulatory Status #

This is where brand partners need to slow down and read carefully. The regulatory picture is not uniform across markets, and what’s legal in the US is not always legal in the EU, and what’s legal in the EU is not always legal in China.

The EU ceiling of 0.3% retinol for face leave-on products came into force under EU Cosmetics Regulation 1223/2009 following the SCCS Scientific Opinion on retinol safety — specifically the SCCS/1576/16 opinion and its subsequent revision. The SCCS flagged concerns around cumulative vitamin A intake, particularly for body lotions used over large surface areas. That’s why the body leave-on limit is lower at 0.05%. If you’re building a retinol body lotion for EU distribution, 0.05% is your ceiling. Full stop.

For brands targeting the US market, the FDA Cosmetics Guidelines don’t set a specific retinol concentration limit in the same way — but retinoic acid remains firmly prescription-only. We’ve seen brands try to blur this line with “retinoid acid complex” naming. Don’t. The FDA reads ingredient lists, not marketing copy.

China’s NMPA Cosmetic Regulation has been tightening. Retinol is permitted up to 0.3% in leave-on products, mirroring the EU position. Retinal sits in a grey zone — we advise brand partners to seek pre-registration guidance before committing to a retinal formula for the Chinese market, because the registration pathway is not straightforward.

Our opinion: HPR (hydroxypinacolone retinoate) is underused. It binds directly to retinoic acid receptors without requiring enzymatic conversion, it’s more photostable than retinol, and it sits outside the restricted ingredient lists in most major markets. For brands that want retinoid efficacy with a cleaner regulatory path, HPR deserves serious consideration. We’ve been formulating with it since 2019 and the stability profile is genuinely better than retinol at equivalent activity levels.

For deeper context on our encapsulation approaches for retinoids, see our Encapsulation Technology resource.

Stability Engineering: Where Most Retinol Formulas Actually Fail #

Retinol degrades via oxidation and isomerisation. The primary degradation pathway under light exposure is photoisomerisation to 9-cis and 13-cis retinol, both of which have significantly reduced biological activity. Under thermal stress, oxidative cleavage dominates. In our stability chamber, a poorly protected retinol formula at 0.5% can lose 40–60% of active content within 8 weeks at 40°C/75% RH. We’ve seen it. It’s not a theoretical concern.

The pH window matters more than most brands realise. We stabilise retinol at pH 5.0–5.5 using citrate-phosphate buffer. Above pH 6.5, degradation accelerates noticeably. Below pH 4.5, you start creating compatibility issues with other actives and you risk barrier disruption that compounds the irritation profile. The 5.0–5.5 window is not arbitrary — it’s where we consistently see the best 12-month stability data.

Antioxidant co-formulation is non-negotiable. We use a combination of tocopherol (vitamin E) at 0.5–1.0% and BHT at 0.02–0.05% as a primary antioxidant system. Some formulations also benefit from ascorbyl palmitate, though we’ve had mixed results — in water-containing systems, ascorbyl palmitate can itself degrade and generate pro-oxidant species. That’s a real failure mode we encountered on a 2021 project: a brand wanted retinol + vitamin C in a single serum, we used ascorbyl palmitate as the bridge, and the 6-month stability data showed accelerated retinol loss compared to the tocopherol-only control. We reformulated with encapsulated ascorbic acid in a separate phase. Problem solved, but it cost three months.

Encapsulation is the most reliable stabilisation strategy for high-load retinol (above 0.3%). We use lipid nanoparticles (LNPs) and cyclodextrin complexation depending on the target texture. LNPs give better physical stability in emulsions; cyclodextrin works well in water-based serums where you want a lighter skin feel. The encapsulation efficiency we target is ≥85% — below that, the free retinol fraction is large enough to cause both stability and tolerability problems.

Packaging is part of the formulation. Airless pumps, opaque or amber glass, nitrogen-purged headspace. If a brand partner insists on a clear glass dropper bottle for retinol, we have a direct conversation about it. The aesthetics are not worth the stability compromise.

Clinical Evidence: What the Numbers Actually Show #

We reference a well-designed split-face, double-blind, randomised controlled trial published in the Journal of Cosmetic Dermatology (Dhaliwal et al., 2019): n=44 participants, 12-week duration, comparing 0.1% retinaldehyde (retinal) versus 0.025% tretinoin versus vehicle control. The retinal arm showed a 44% reduction in fine line depth (profilometry measurement) versus baseline, compared to 53% for tretinoin. Erythema scores at week 4 were 1.8/10 for retinal versus 4.2/10 for tretinoin on a validated visual analogue scale. By week 12, tolerability had equalised somewhat, but the retinal group showed significantly fewer dropouts due to irritation — 2 of 44 versus 9 of 44 in the tretinoin arm.

What this tells us practically: retinal delivers roughly 80–85% of the efficacy of low-dose tretinoin with substantially better early tolerability. For a cosmetic product that cannot legally contain tretinoin in most markets, retinal is the closest you can get to prescription-level results within a compliant framework.

We also track our own in-house consumer perception data. Across 12 retinol product launches we’ve supported in the past four years, formulas with encapsulated retinol at 0.2–0.3% with a structured retinisation protocol (start 2×/week, build to nightly over 6 weeks) showed a consumer-reported tolerability satisfaction rate of 78% at week 8. Unencapsulated retinol at the same concentration: 54%. The encapsulation difference is real and measurable in consumer outcomes, not just lab data.

For brands building anti-aging positioning around retinoids, our Anti-Aging Formulation category covers the broader active ingredient architecture beyond retinoids alone.

Stability testing protocols for retinoid actives should align with ICH Stability Guidelines — specifically Q1A(R2) for long-term and accelerated conditions. We run all retinoid formulas through a minimum 6-month accelerated study (40°C/75% RH) before recommending commercial launch, and 12-month real-time data before we consider the formula fully validated.

Retinisation Protocols and Consumer Communication #

This section is honestly still evolving, and we don’t have a single definitive answer. The concept of “retinisation” — the skin adaptation period during which barrier disruption, peeling, and erythema normalise — is well-accepted clinically, but the optimal protocol for cosmetic retinoid products is not standardised. Different brands use different ramp-up schedules, and the evidence base for one protocol over another is thin.

What we do know from our production experience: the irritation profile is driven by three variables — retinoid concentration, vehicle composition, and application frequency. You can modulate all three. A brand that wants to launch a 0.5% retinol serum for a first-time retinoid user is going to have a bad time unless the vehicle is specifically designed to buffer the irritation response — ceramide-rich emollients, niacinamide at 4–5%, panthenol at 1–2%. We build these in by default now.

The question of whether to communicate the retinisation period to consumers is a brand decision, not a formulation one. But we always recommend it. Consumers who understand that initial dryness is expected are far less likely to return the product or leave a negative review. Whether that communication belongs on the carton, in a QR-linked guide, or in a brand’s social content — that’s up to the brand. What happens when brands skip this step entirely?

Formulation Notes for Brand Partners #

When you brief us on a retinoid project, the first question we ask is: which markets are you launching in, and what’s your 18-month regulatory roadmap? That single answer shapes everything — the retinoid choice, the concentration ceiling, the stability testing protocol, and the claims language.

For EU-primary launches, we default to retinol at ≤0.3% with full encapsulation and a ceramide-based vehicle. For US-primary launches with a performance-focused positioning, retinal at 0.05–0.1% is increasingly our recommendation — it’s not restricted, it’s more potent per unit concentration, and the stability profile, while demanding, is manageable with the right packaging. For brands that want a genuinely low-irritation entry point, HPR at 0.1–0.2% in a squalane-rich base is a strong option.

MOQ for retinoid formulas starts at 500 units for pilot batches, with full stability and safety documentation included. Lead time from brief to first sample is typically 4–6 weeks. We do not recommend launching any retinoid product without a minimum 6-month accelerated stability study — the liability exposure from a degraded product reaching consumers is not worth the time saving. Packaging selection, as noted above, is part of the formulation conversation, not an afterthought.

Frequently Asked Questions #

Q: We want to put 1% retinol on our label. Can you make that?

Technically, yes — we can formulate at 1% retinol. But if you’re selling into the EU, 1% is above the 0.3% regulatory ceiling for face leave-on products, so that’s a non-starter for that market. For the US, there’s no hard concentration limit, but I’d push back on whether 1% is actually the right call. At that level, without very careful vehicle design and encapsulation, you’re going to see tolerability complaints that hurt your brand more than the potency claim helps. We’d typically recommend 0.5% encapsulated retinol with a strong barrier-support vehicle over 1% unencapsulated. The consumer outcome is better.

Q: What’s the difference between retinal and retinaldehyde? Are they the same thing?

Same molecule, different names. Retinaldehyde is the INCI name; retinal is the biochemistry shorthand. When you see “0.05% retinaldehyde” on a competitor’s formula, that’s the same ingredient as retinal. The potency advantage — roughly 11× versus retinol on a molar basis — is real, but so is the formulation complexity. It oxidises faster than retinol and requires more aggressive antioxidant protection and packaging controls.

Q: Can we combine retinol with vitamin C in the same formula?

We do it, but carefully. The pH conflict is real — retinol wants pH 5.0–5.5, ascorbic acid wants pH 3.0–3.5 for stability. The compromise is encapsulation: we encapsulate one or both actives to allow them to coexist in a vehicle at pH 5.0–5.5. The alternative is a two-phase product or a separate AM/PM regimen recommendation. We had a project in 2021 where we tried to bridge this with ascorbyl palmitate and it accelerated retinol degradation — so we don’t use that approach anymore. Encapsulated ascorbic acid in a lipid matrix, combined with encapsulated retinol, is the formulation we trust.

Q: How long does stability testing take, and can we skip the 12-month study for a faster launch?

Accelerated stability at 40°C/75% RH over 6 months is the minimum we’ll sign off on for a retinoid product. That’s non-negotiable from our side. The 6-month accelerated study, following ICH Stability Guidelines Q1A(R2), is designed to predict 24-month real-time shelf life. We run real-time studies in parallel, but you don’t need to wait for 12-month real-time data to launch — you need the 6-month accelerated data. Total timeline from formula lock to stability clearance: approximately 7–8 months including sample preparation and analysis.

Q: We’ve seen “bakuchiol as a retinol alternative” — should we use that instead?

Bakuchiol is a legitimate ingredient with a real evidence base — a 2018 split-face RCT (n=44, 12 weeks) showed comparable reduction in fine lines and pigmentation versus 0.5% retinol, with significantly lower irritation scores. We formulate with it. But “retinol alternative” is a positioning claim, not a mechanism claim — bakuchiol does not convert to retinoic acid and does not work through the same receptor pathway. It’s a retinoid-mimetic with a different mechanism. If your brand wants to avoid retinoids entirely — sensitive skin positioning, pregnancy-safe claims, clean beauty — bakuchiol is a strong choice. If your brand wants the retinoid conversion cascade and the clinical depth behind it, bakuchiol is not a substitute. Different tools for different jobs.

Have a product concept in mind? Contact our formulation team to request a complimentary brief review.