Dr Elsa Jungman, Founder & CEO of HelloBiome
Abstract
Microbiome research in cosmetics is generating increasingly sophisticated clinical datasets, yet the real value lies not in sequencing itself but in how this data informs formulation development, personalisation strategies, and claim substantiation. As the industry moves beyond generic “microbiome-friendly” positioning, the ability to translate clinical microbiome outputs into targeted product design becomes critical.
This article examines how multi-area microbiome profiling combined with AI-driven clustering enables the identification of reproducible biological subtypes associated with sensitivity, barrier resilience, and longevity markers. These stratified signatures provide a structured foundation for ingredient selection, formulation refinement, and the development of product lines tailored to defined microbiome clusters.
Rather than relying solely on diversity metrics, actionable endpoints such as stability indices, ecosystem balance scores, and dysbiosis normalisation can directly guide formulation adjustments and support clear, defensible efficacy claims. Integrating microbiome analytics – early into R&D workflows allows brands to accelerate development cycles while building differentiated, data-driven products aligned with emerging personalisation paradigms.
By connecting clinical microbiome science to formulation decision-making and claim architecture, this approach outlines a practical pathway from data to commercially and scientifically robust innovation.
The Limits of the Bioactive Model
For decades, cosmetic innovation has largely followed a bioactive paradigm: identify a target pathway, introduce a functional ingredient, measure a biomarker, and construct a claim. This reductionist model has been effective for hydration, pigmentation, collagen stimulation, and antioxidant activity. It is built on linear cause-and-effect logic.
Microbiome biology, however, does not operate linearly. The skin microbiome is a dynamic ecosystem composed of interacting bacterial and fungal communities influenced by host factors, environmental exposures, age, lifestyle, and product use. Modulating one species rarely produces isolated downstream effects. Instead, changes occur at the network level.
This shift from single-target intervention to ecosystem modulation requires a systems framework. Without such a framework, microbiome claims risk remaining superficial, often limited to statements such as “microbiome-friendly” or “does not disrupt diversity,” without defining what balance or resilience truly means in biological terms. The transition from bioactive thinking to systems thinking marks an inflexion point for microbiome-driven formulation.
From Profiling to Stratification: Defining Reproducible Microbiome Subtypes
Traditional skin classification systems: oily, dry, combination, sensitive, were developed to simplify visible phenotypes and guide consumer-facing product selection. While useful from a communication standpoint, these categories are inherently descriptive and often subjective. Two individuals who both identify as “dry” may exhibit markedly different barrier function, immune reactivity, sebum dynamics, and microbial composition. Conversely, individuals with different self-reported skin types may share similar underlying biological profiles.
This limitation becomes particularly evident in the case of sensitive skin. Surveys consistently indicate that approximately 70% of women report having sensitive skin. If most of the population falls into a single category, the classification loses discriminatory power. A singular “sensitive” label cannot meaningfully capture the biological heterogeneity that exists within that group. Some individuals may exhibit barrier impairment, others immune hyperreactivity, others altered sebum composition, and still others distinct microbial ecosystem shifts. Microbiome science introduces an opportunity to rethink this classification.
Rather than classifying skin solely by surface characteristics or consumer perception, large-scale microbiome profiling enables the identification of reproducible ecosystem signatures. These signatures reflect not only microbial composition, but also the interaction between host physiology, age, environment, and product exposure. When analysed at scale, structured patterns emerge that cut across traditional labels.
In multi-hundred-participant datasets, unsupervised clustering approaches have revealed distinct microbiome subtypes that do not map directly onto conventional skin categories. Certain clusters characterised by a high relative abundance of Cutibacterium acnes align more consistently with sebum-driven profiles, while other clusters enriched in taxa associated with lower sebum production and higher reactivity correlate more frequently with dryness and perceived sensitivity. Importantly, individuals who all describe themselves as “sensitive” may distribute across multiple microbiome-defined clusters, each with different biological drivers.
Age-related hormonal changes further influence the structure of the skin microbiome. Declining sebum production and shifts in lipid composition, particularly during menopause, are associated with measurable reductions in Cutibacterium abundance and relative increases in taxa adapted to drier microenvironments, including certain Corynebacterium species. Several studies (1, 2, 3) have also reported an overall increase in microbial diversity with age, suggesting that ageing reflects an ecological restructuring rather than a simple loss of microbial presence. These transitions reinforce the concept that skin ageing is accompanied by ecosystem-level shifts that may require distinct formulation strategies compared to younger, sebaceous-dominant profiles.
These findings suggest that what has historically been described as a “skin type” may in fact represent multiple distinct biological states. A microbiome-informed framework, therefore, does not eliminate traditional terminology, but refines it, introducing a stratified, data-driven layer that allows for greater granularity in profile development and formulation targeting.
In this sense, clustering is not merely a statistical exercise. It represents a structural evolution in how the industry defines target populations. By shifting from surface description to biologically anchored ecosystem states, microbiome analytics provides a pathway toward more precise formulation strategies, more nuanced personalisation, and more defensible claims.
Translating Microbiome Clusters into Formulation Strategy
The practical question for R&D teams is how to convert microbiome data into formulation decisions.
The first step is to identify a target population represented within a defined cluster. The metadata and microbiome profile of that cluster are examined to understand its biological characteristics. From there, ingredients and active compounds are selected to rebalance identified ecosystem shifts while simultaneously addressing associated skin concerns through complementary biological pathways.
For example, once a subtype associated with dryness and sensitivity is characterised by specific abundance patterns and diversity profiles, ingredient screening can focus on actives that modulate that ecosystem signature. Candidate ingredients are then incorporated into formulations tested in controlled clinical settings, with microbiome endpoints assessed alongside traditional measures such as transepidermal water loss, erythema, or sebum output. This structured loop of cluster identification, ingredient selection, formulation development, and clinical validation shifts microbiome science from descriptive profiling to applied formulation architecture.
A critical evolution in this approach is moving away from the concept of a single hero being active. Every ingredient in a formulation contributes to the ecosystem impact. Fragrances, essential oils, preservatives, and emulsifiers can influence microbial balance as significantly as the featured active blend. Minimising unnecessary components and selecting ingredients with minimal negative impact on microbial equilibrium is essential. The entire formulation must be considered as a biological system, not only the highlighted actives.
Importantly, this approach moves beyond the simplistic objective of increasing diversity. In certain contexts, higher diversity may correlate with disease states, while in others, stability within a defined ecological range may be more desirable. Formulation strategies must therefore be guided by cluster-specific interpretation rather than generalised assumptions.
Within this framework, clusters become the new biological profiles for product development, replacing traditional segmentation based solely on descriptive skin type categories.
Personalisation as a Structural Outcome
Personalisation in cosmetics has often relied on questionnaires, consumer perception, or lifestyle segmentation. Microbiome clustering introduces a biologically grounded foundation.
When reproducible microbiome subtypes are identified across large datasets, product portfolios can be structured around defined ecosystem signatures. Instead of a single universal formulation marketed as suitable for all microbiomes, modular strategies become possible:
- Formulations optimised for high sebum, Cutibacterium dominant clusters
- Strategies tailored to low sebum, higher diversity, sensitivity associated clusters
- Age-associated ecosystem modulation approaches
This does not require fully bespoke manufacturing. Cluster-informed product lines can offer stratified solutions grounded in measurable biological profiles.
Such an approach also refines claim language. Rather than broad positioning, claims can reference support of defined microbiome balance parameters within a characterised subgroup, provided substantiation aligns with regulatory expectations.
Clusters can also serve as a structured basis for diagnostic interpretation. By anchoring reports to known and reproducible biological patterns, interpretation becomes more robust. An emerging challenge in the field is the increasing use of general artificial intelligence tools to interpret microbiome reports, which can generate biologically inaccurate or hallucinatory conclusions. The next step is the development of domain-specific AI tools trained on validated datasets to provide a deeper layer of interpretation and recommendation beyond clustering alone, while maintaining scientific rigour.
Integrating Microbiome Endpoints into Claims Architecture
As microbiome claims proliferate, regulatory scrutiny increases. To remain defensible, claims must be linked to measurable endpoints and a clear study design.
Embedding microbiome analysis into clinical protocols from the outset strengthens substantiation. Baseline sampling establishes ecosystem subtype. Post intervention sampling assesses directional change relative to predefined endpoints. Correlation analyses may link microbial shifts with improvements in clinical parameters.
For example, a claim framework might demonstrate that a formulation:
- Maintains bacterial to fungal balance within a defined range
- Reduces the abundance of species associated with sensitivity-prone clusters
- Rebalances a defined microbiome profile while improving clinical parameters
Such endpoints provide greater specificity than generic statements of compatibility. However, the rigour of claims depends not only on endpoints but on how participants are recruited, how testing is conducted, and how data are analysed.
A major methodological risk in microbiome trials is excessive baseline variability. If participants with highly divergent microbiome profiles are grouped, meaningful rebalancing effects may appear statistically insignificant at the cohort level because different species shift in different individuals. Leveraging cluster knowledge during recruitment allows for more homogeneous cohorts, reducing variability and enabling clearer statistical demonstration of microbiome modulation when it occurs.
Extending Systems Thinking to At-Home Testing and CRO Models
At-home sampling supports testing in real-world conditions. Participants do not need to travel to a laboratory or be available on specific clinic days. This approach expands access and enables recruitment across broader geographic and demographic groups. It also facilitates the inclusion of additional metadata parameters and supports recruitment of cluster-specific cohorts.
At home sampling has emerged as a viable extension of traditional clinical workflows, particularly with the rise of digital infrastructure and AI-supported data systems. Participants register their kits online, complete structured questionnaires, and collect standardised skin swabs before and after product use, with centralised sequencing and analysis. Images and video documentation can also be collected to complement microbiome endpoints. When aggregated across large cohorts, this enables longitudinal tracking of microbiome dynamics under everyday conditions.
Publicly available examples demonstrate that such models can assess ecosystem stability over multi-week usage periods, evaluating diversity metrics and abundance patterns with measurable rebalancing. While these studies do not replace controlled clinical trials, they complement them by capturing ecological responses in real use environments, particularly relevant for consumer and beauty applications.
For contract research organisations, microbiome endpoints can be integrated into standard protocols. Baseline subtype identification, longitudinal sampling, and correlation with clinical markers allow microbiome modulation to become a structured dimension of product testing rather than an exploratory addition. In many cases, sampling can be integrated without complex freezing logistics, and results can be generated within days. In this context, at-home sampling functions as a scalable data layer within a broader systems framework, linking laboratory validation, clinical efficacy, and real-world evidence.
From Data Generation to Data Integration
The maturation of microbiome science in cosmetics will not be defined by the volume of sequencing performed, but by how effectively that data informs formulation strategy and claim substantiation.
The transition from bioactive thinking to systems thinking requires:
- Large-scale reproducible clustering
- Clearly defined ecosystem endpoints
- Integration of microbiome analytics into early R&D stages
- Structured validation loops combining in vitro, clinical, and real-world data
When microbiome profiling is treated as infrastructure rather than marketing language, it enables a disciplined pathway toward personalisation and innovation.
The future of microbiome-driven cosmetics lies not in broad declarations of compatibility, but in measurable ecosystem modulation anchored in structured, scalable, and defensible scientific frameworks.
References and notes:
- Sun, C., Hu, G., Yi, L. et al.Integrated analysis of facial microbiome and skin physio-optical properties unveils cutotype-dependent aging effects. Microbiome 12, 163 (2024). https://doi.org/10.1186/s40168-024-01891-0
- Howard B, Bascom C, Hu P et al. Aging-Associated Changes in the Adult Human Skin Microbiome and the Host Factors that Affect Skin Microbiome Composition. Journal of Investigative Dermatology, 2021; 142, 1934-1946.e21 https://linkinghub.elsevier.com/retrieve/pii/S0022202X21026038
- Larson PJ, Zhou W, Santiago A, Driscoll S, Fleming E, Voigt AY, Chun OK, Grady JJ, Kuchel GA, Robison JT, Oh J. Associations of the skin, oral and gut microbiome with aging, frailty and infection risk reservoirs in older adults. Nat Aging. 2022 Oct;2(10):941-955. doi: 10.1038/s43587-022-00287-9. https://pmc.ncbi.nlm.nih.gov/articles/PMC9667708/
Dr Elsa Jungman will present ‘Turning Clinical Microbiome Data Into Actionable Claims & Product Innovation’ at in-cosmetics Global 2026 on Tuesday 14 April, from 13:15–13:45 in Technical Seminar Theatre 1, Hall 7, Level 3 (Booth 3A130).
About the author: Dr Jungman holds a PhD in skin barrier function and is a recognised leader in skin and microbiome innovation. After launching a pioneering microbiome-friendly skincare line and consumer test kit, she founded HelloBiome, a B2B AI platform advancing clinical research and product innovation and edited the Handbook of Cosmetic Science and Technology.
