Collagen has become a central idea in anti-ageing skincare, and the framing around it has shifted. The industry now talks about “collagen banking”: acting before collagen decline becomes visible rather than correcting it afterwards. In practice that means products aimed at consumers in their twenties and thirties, positioned around maintaining what’s already there rather than restoring what’s been lost.
That shift has pulled a lot of formulation activity toward one question: can this product make the skin produce more of its own collagen? We’ve looked at what’s driving the trend, what “boosts collagen” actually measures, and how far in vitro data can take a claim before you commit to a clinical study.
What’s driving it
Peptides are the ingredient story of 2026, and specifically peptides as collagen signals. Industry commentary describes a move away from basic moisturising peptides toward targeted blends designed to mimic the skin’s own signalling pathways and make collagen stimulation more efficient [1]. Retinol and peptides work through different routes: retinol accelerates cell turnover, while peptides signal fibroblasts to produce more collagen and elastin.
The prevention mindset is now a distinct and commercially significant consumer segment. YouGov’s 2026 anti-ageing research found that 35% of US consumers identify as “aging preventers”, and that spending concentrates sharply in that group: 19% of them spend $50 or more a month on skincare, supplements or anti-ageing products, against 2% of those indifferent to ageing. Skincare is the entry point, with traditional anti-ageing skincare showing the highest future consideration of any option at 32%. Prevention is a bigger priority still in several Asian markets, at 77% in Indonesia and 64% in India, against 35% in the US [2]. Asian markets frequently lead on beauty trends, so the gap is more likely to narrow than hold.
The direction is consistent. Consumers are being sold on mechanism, not just outcome, which means the mechanism needs evidence behind it.
“Collagen” on a label means at least three different things
The distinctions below will be familiar to anyone formulating in this space. They’re worth setting out because consumers largely don’t make them, and that gap is where claims get scrutinised.
Three quite different things travel under the same word.
Ingestible collagen is a supplement category, taken orally. It sits alongside topical skincare in a lot of marketing but it’s a separate business with separate evidence requirements.
Collagen as a topical ingredient is a different thing again. The 500 Dalton rule holds that molecules above roughly 500 Da show markedly reduced penetration through the stratum corneum [3]. Native collagen is around 300,000 Da, and hydrolysed collagen, while considerably smaller, typically still sits in the thousands. Where collagen appears in a formulation it is generally working at the surface, as a humectant and film former. That’s a legitimate function. It isn’t collagen boosting.
Collagen boosting is the claim that the product stimulates the skin’s own fibroblasts to synthesise more collagen. This is the biological efficacy claim, and it’s the one that needs data.
A consumer reading an INCI list will often treat all three as broadly the same thing. That’s an environment where a well-evidenced claim earns something and a loose one is exposed.
What “boosts collagen” actually measures
There are two main routes. Gene expression by qPCR shows that the fibroblasts have been instructed to produce more collagen. The Sircol method measures the collagen actually produced, using a Sirius red dye that binds specifically to collagen and is quantified colorimetrically.
Both support a collagen claim. Which one suits depends on what you want the data to show.
One practical advantage of the cell-based work is that the culture medium can be retained for later analysis of other biomarkers, elastin for instance, so the same experiment can answer more than one question.
A tiered approach, and why the second tier matters here
We run claims support in tiers, which is simply a way of matching the depth of the evidence to what the claim needs to carry, and answering the simpler questions before the more costly ones.
The first tier uses human dermal fibroblasts, the dermal cells responsible for collagen production. Test items are dosed across a range of concentrations and collagen production is quantified by Sircol or qPCR. This suits ingredients and water-soluble formulations, and it’s where you find out whether there’s an effect at all and how it varies with dose and time.
The second tier uses reconstructed human skin, and for collagen specifically it answers a question the fibroblast work cannot. Fibroblasts sit in the dermis. A topically applied active has to get through the epidermis to reach them. A fibroblast assay bypasses that entirely by applying the active directly to the cells.
Testing in reconstructed skin incorporates absorption and metabolism, which makes it the closest in vitro simulation of real-life exposure. It is suitable for more formulation types rather than being limited by aqueous solubility, and allows picrosirius red staining and histology alongside the same molecular endpoints.
That second tier matters more for collagen than for most endpoints. An ingredient can perform well on fibroblasts and still not deliver in a finished product if it never reaches the dermis, and the first tier alone won’t tell you which situation you’re in.
Where mechanistic data takes you
Evidence that an ingredient increases collagen synthesis in a fibroblast model is a claim about collagen synthesis. An appearance claim, “reduces the appearance of fine lines” and similar, is a different endpoint clinically measured on people.
The useful thing is that the first tells you a great deal about whether the second is worth pursuing. Clinical studies are expensive and slow, and going into one without knowing whether your active does anything at a cellular level is an expensive way to find out. In vitro collagen data answers that question first, on a small sample, in weeks rather than months.
It also tells you which formulation to take forward. If you have three candidates and the mechanistic data separates them clearly, you’re running the clinical work on the strongest one rather than on all three or on a guess.
So the mechanistic evidence does two jobs. It supports the mechanistic claim in its own right, and it de-risks the appearance claim before you commit to proving it.
If you’re working on a collagen claim
We run collagen testing in animal-product-free conditions, across fibroblast and reconstructed skin models, using Sircol and gene expression analysis depending on what the claim needs to carry. If you’re developing an active or a formulation and want to work out which endpoint suits, get in touch.
Email info@x-cellr8.com or call +44 (0)1925 607 134.
References
[1] Croda Beauty. How peptides are shaping the US cosmetics industry.
[2] YouGov (2026). Forever Young? Anti-Aging Report 2026. Fieldwork 16 December 2025 to 12 January 2026, 17 markets.
[3] Bos, J.D. and Meinardi, M.M.H.M. (2000). The 500 Dalton rule for the skin penetration of chemical compounds and drugs. Experimental Dermatology, 9(3), 165-169.
———
Contact us if you’d like to learn how you can support our mission, become a corporate sponsor or donate to our crowdfunder.
Email info@x-cellr8.com
Call +44 (0)1925 607 134.
Follow Us!