Anti-Aging — Industry Case Study

Key Technical Parameters #

A North American prestige brand came to us in Q1 2023 with a clear brief: reformulate their flagship 30ml anti-aging serum, which had a 3.4-star average review driven by consumer complaints about skin sensitivity and pilling on application. The core challenge wasn’t finding better actives — it was rebuilding tolerance without sacrificing the 12-week efficacy data the brand had already used in retail sell-in. Brands in the prestige DTC segment face a specific pressure here: they can’t quietly swap formulas without triggering FTC-adjacent scrutiny if existing clinical claims are retained. This case study walks through how we rebuilt the formula, what we measured, and where the project nearly came off the rails.

The Diagnostic Phase — Why the Original Formula Failed at Scale #

The brand’s existing serum had a clean efficacy story. Wrinkle depth reduction was validated by a dermatologist-commissioned split-face study. On paper, the formulation looked reasonable: 0.5% encapsulated retinol, 5% niacinamide, a peptide blend at 3%, pH 5.2, standard carbomer gel base. What wasn’t visible in the bench data was how the formula behaved after six months in a warm warehouse and two weeks in a consumer’s bathroom.

We ran the original formula through our incoming re-analysis protocol (what we call IRT-04, our re-test procedure for reformulation briefs) and found three problems immediately.

First, the carbomer network was interacting with the cationic species in the preservative system — a benzalkonium chloride variant — causing viscosity drift. By month 4 at 40°C, viscosity had dropped from 18,000 cP to roughly 9,000 cP. That explains the pilling: a serum that’s too thin doesn’t form a proper film before the next layer goes on.

Second, the encapsulated retinol was releasing prematurely. The brand’s supplier had provided encapsulation data based on a simple water dispersion at pH 6.5. Our formula sat at pH 5.2. Below pH 5.5, the PLGA-type shell the supplier was using starts to hydrolyze faster — not dramatically, but enough that by the time the product was 60 days old, roughly 20–25% of the retinol was already free in the base. Free retinol at that loading drives the sensitivity complaints.

Third, and this took us longest to figure out, the silicone slip agent used to improve skin feel was creating a barrier that slowed the release of niacinamide into the skin. The ingredient was doing its job aesthetically — but it was partially defeating the brightening stack. Not every brand partner would even notice this. They did, because their consumer panel data showed niacinamide-related evenness scores underperforming versus competitors at week 6.

These three failures are interconnected but they have different root causes. You can’t fix them with one reformulation decision.

Reformulation Strategy — Choices, Trade-offs, and What We Rejected #

We presented three reformulation routes in a structured brief-back session. The table below summarises the key differences across the approaches we considered.

The brand chose Route B. Not Route A, which we’d anticipated — and not Route C, which we’d quietly hoped for.

Their reasoning was commercially sound: they didn’t want free retinol in the formula at all anymore, and they weren’t willing to wait 22 weeks. Route B let them retain a retinoid narrative on pack by framing retinyl propionate as a “next-generation vitamin A derivative,” which holds up under FDA Cosmetics Guidelines as long as no drug claims are attached.

Honestly, retinyl propionate was the right call for their consumer. It converts to retinol in the epidermis via lipase activity, but the irritation profile is gentler — and the pH tolerance is much wider, stable up to pH 6.5. We reformulate at pH 5.7. At that pH, the polyglyceryl-based emollient matrix we use as the new base behaves predictably, and we could drop the benzalkonium chloride preservative entirely, switching to a phenoxyethanol/ethylhexylglycerin system at 0.8% total that doesn’t interact with the thickener.

What we rejected: the brand initially wanted to keep the peptide blend from their original formula supplier. We pushed back. The peptide supplier was offering a pre-blended concentrate at pH 7.0–7.2, and buffering that down to pH 5.7 without precipitating the blend requires a re-qualification process. After one failed attempt to make it work (the concentrate went cloudy at pH 5.8 in two out of three batches during pilot), we switched to sourcing the peptides separately and building the blend in-house.

We also rejected adding ceramides to Route B at this stage. The brief asked for it. But ceramide loading above 0.5% in a water-based gel base creates long-term phase separation risk unless you’re using a lamellar gel network system, which adds another 6 weeks of development and moves you closer to Route C economics anyway. We recommended ceramides as a Phase 2 product extension instead.

Stability Data and the Real-Time Confirmation Problem #

Reformulation completed in week 11. We had 12-week accelerated stability data by week 23. The results were solid: viscosity maintained at 16,500–17,200 cP at 40°C over 12 weeks (compared to the original’s collapse to 9,000 cP), retinyl propionate assay retained at 97.3% at week 12, and pH drift contained to ±0.15 units.

But here’s what we flagged to the brand and what caused real commercial hesitation: accelerated stability at 40°C for 12 weeks does not map cleanly to a 24-month room-temperature shelf life for all formula types. Per ICH Stability Guidelines — specifically Q1A(R2), which most cosmetic-adjacent brands apply voluntarily — the correlation between accelerated and real-time data is an approximation. For retinol systems especially, we’ve seen real-time 18-month results underperform the accelerated prediction.

We initiated real-time 25°C/60% RH stability at the same time as accelerated, per standard practice. We told the brand clearly: we won’t have 24-month confirmation until mid-2025. They launched in month 9 of the project. That’s their risk to take, and they took it with full information.

Our own dataset on retinyl propionate formulations is smaller than we’d like — based on 11 projects over three years, not the decades of data we have on retinol. We’re still building confidence in the long-term oxidation rate at different packaging formats. The brand is using frosted glass with an airless pump, which reduces oxygen exposure materially. That helps. Whether it’s enough to guarantee a 24-month assay of 95%+ is something we’ll know in mid-2025.

We haven’t fully resolved how packaging material variance affects the retinyl propionate oxidation rate across different glass suppliers. It’s on our internal tracking list.

Clinical Revalidation — Running a New Study Against a Moving Brief #

Route B required a new clinical study if the brand wanted to retain specific numeric claims. They did. We helped them commission a dermatologist-run split-face, double-blind, placebo-controlled study (n=44, 12 weeks) measuring wrinkle depth via optical profilometry and transepidermal water loss (TEWL) as a tolerance proxy.

Results at 12 weeks: 28% reduction in wrinkle depth versus baseline (vs. 31% in the original formula’s study, which used a different measurement methodology — the brand’s statistician confirmed the two numbers aren’t directly comparable). TEWL improved by 18% versus baseline in the new formula group; no measurable change in the placebo group. Sensitivity adverse events were reported by 3 out of 44 participants in the treated group, compared to 9 out of a comparable 44-participant cohort in the original formula’s historical data.

That last number is the one the brand’s marketing team cared about most. Three versus nine. That’s the commercial story: same efficacy category, meaningfully better tolerance profile. Under EU Cosmetics Regulation 1223/2009 Article 20, which governs claim substantiation, the new study is a standalone document — it doesn’t require comparison to the old formula. But the brand chose to use the contrast in internal sell-in decks and retailer presentations, which is their prerogative.

The study also gave us something useful for our own formulation library: a validated reference point for a retinyl propionate system at 1% in a polyglyceryl base at pH 5.7. That’s a data point we’ll carry into future briefs in a similar category. See also our anti-aging formulation framework, where we track validated actives against their clinical evidence base across multiple projects.

ROI and Scalability — What the Numbers Actually Looked Like #

The brand’s original serum was priced at $68 retail. Their blended COGS target was $7.20/unit at 10,000 units/month. The original formula landed at $6.95/unit at that volume.

Route B came in at $7.22/unit at 10,000 units/month — $0.27 above target. After two rounds of negotiation on the polyglyceryl emollient sourcing (we have a tiered volume agreement with our primary supplier, activated at 500kg batches), we brought it to $7.08/unit. Still slightly above the original, but within the brand’s revised tolerance after factoring in reduced return rates.

The brand estimated — and this is their own model, not ours — that sensitivity-driven returns and refunds were running at approximately 4.2% of units sold. Post-reformulation, six months in, that figure had dropped to under 1.5%. At their sales volume, that’s a meaningful cost recovery. Their calculation showed full payback on reformulation development costs (which they put at $38,000 total including clinical study) within approximately 8 months.

Scalability had one real constraint we hadn’t fully anticipated: the airless pump component. At 50,000 units/month — the brand’s 18-month target — their existing pump supplier couldn’t guarantee lead times below 14 weeks. At 10,000 units, 8-week lead time was fine. This is a packaging issue, not a formulation issue, but it blocked scale-up more than anything in the formula. We flagged alternative pump suppliers with verified fill compatibility for the new base viscosity range.

For our face serum category more broadly, this project reinforced something we see repeatedly: scale-up failures in prestige skincare are more often a packaging supply chain problem than a formulation problem.

Formulation Notes for Brand Partners #

When you brief us on a reformulation project — especially one involving a formula already in-market — the first questions we ask are about the existing failure mode, not your target formula. What’s the complaint data? What’s the return rate? What did the consumer actually say? Without that, we’re guessing at root causes, and that costs time.

The most common mistake we see in reformulation briefs is brands treating this as an ingredient swap exercise. “Replace retinol with something gentler” sounds like a simple brief. In practice, switching the retinoid system changes the pH window, which changes the preservative selection, which may change the texture and the interaction with every other active in the formula. A brief that feels like a 6-week project often becomes 14 weeks because the interactions weren’t mapped upfront.

What we need from you: target market (EU claims requirements differ materially from US and NMPA), format and packaging spec, any clinical claims you plan to retain, and your COGS ceiling at your forecast volume. The COGS number is the variable most brands tell us last. Tell us first.

Timeline for a reformulation project at this complexity: lab samples in 2–3 weeks, accelerated stability 4–8 weeks, real-time stability initiated concurrently. If a new clinical study is required, budget 14–18 weeks for study completion from protocol sign-off, typically running in parallel with stability.

Frequently Asked Questions #

We want to reformulate but keep our existing clinical claims — is that possible?
A: It depends on how material the formulation change is and what the claim says. A preservative swap with no change to actives usually stays within variance. Changing the retinoid type entirely — as in this case — typically requires a new study if you want to retain specific numeric claims. Some brands accept the risk of relying on existing data with updated INCI; we advise against it if the claim is on-pack.

Does the EU treat retinyl propionate differently from retinol under Regulation 1223/2009?
A: Currently, EU Cosmetics Regulation 1223/2009 concentration restrictions apply specifically to retinol and retinaldehyde, not retinyl esters as a class. Retinyl propionate sits in a different regulatory position than retinol for now — but watch the SCCS Scientific Opinion updates. The SCCS has been reviewing the entire retinoid family, and the current boundaries may not hold for the full 2025–2026 cycle.

What’s the risk of running a reformulation with accelerated stability only and no real-time data at launch?
A: We’ve seen two situations where accelerated data looked clean and real-time 18-month results showed unexpected active degradation — both in systems with air-permeable packaging. The risk is real, and it’s brand-specific depending on packaging format and distribution conditions. Airless glass reduces it. Jar packaging with a wide mouth increases it noticeably.

What’s your minimum order quantity for a reformulation project like this?
A: For pilot production, we run at 500–1,000 units minimum to generate meaningful stability samples and QC data. Commercial production for a serum at this complexity typically starts at 5,000 units. Full cost structure doesn’t optimise until you’re past 10,000 units per run — that’s where the bulk active sourcing agreements we hold actually activate.

What’s a question we should be asking but probably aren’t?
A: Whether your packaging supplier has tested fill compatibility with your new base. Pump dispensers calibrated for your old formula’s viscosity will misfire at a different cP range — and the misfiring shows up as dosing inconsistency, not a visible defect, so it doesn’t get caught in standard QC. We include fill-rate testing in our pre-production checklist for any reformulation that changes base system or viscosity by more than 15%. Not every manufacturer does.

---

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