Skincare Is About to Get a Lot Less Subjective

Skincare is about to get a lot less subjective.

For most of beauty history, the feedback loop has been remarkably weak.

Brand: “This serum improves pigmentation.”

Consumer: uses it for eight weeks.

Consumer: “I think my skin looks better?”

That ambiguity has been surprisingly useful to the industry.

It may not last.

The tools are arriving faster than the industry expected

A Y Combinator company called Lumeria recently launched the Lumoscope, a small multispectral camera that clips onto a phone. It uses visible, UV and polarized light to track things like pigmentation, redness, sebum, texture and pores over time.

Its pitch is essentially: stop remembering your skin. Measure it.

Stop remembering your skin. Measure it.

Lumeria says its first batch sold out within 24 hours. The product now sells for $299.

But Lumeria is only one signal.

L'Oréal launched Cell BioPrint with Lancôme this year. It analyses five protein biomarkers from the skin and returns a personalised assessment in about five minutes.

Perfect Corp already lets brands analyse up to 15 visible skin parameters from facial images and track them over time.

And researchers are getting surprisingly good at extracting useful skin information from ordinary photographs.

A 2026 study involving 1,099 participants found that an AI model analysing facial images got 93.3% of dermatologist graded predictions within one grade across the facial signs studied. Smartphone results generally held up well, although performance differed by parameter.

Personalised and measurable are very different things

This is where I think the beauty industry gets interesting.

We have spent the last decade making skincare more personalised.

The next decade may make skincare more measurable.

And those are very different things.

Because once consumers can measure change cheaply and repeatedly, three things happen.

“Brightens skin” sounds great on a PDP.

But imagine the customer has 12 weeks of pigmentation data sitting on her phone.

Suddenly the relevant question is no longer “Do I like this serum?” It becomes “Did anything actually change?”

If you claim “reduces redness in 28 days”, you have quietly specified three things:

• An outcome
• A metric
• A timeframe

That is much closer to an experiment than an advertisement.

Founders may eventually need to think about claim architecture much earlier in product development.

Not formula, then packaging, then campaign, then claims. But desired outcome, then measurable endpoint, then formulation, then testing, then claim.

That is a very different way of building beauty products.

This is the part I find most interesting.

You might assume better consumer measurement reduces the need for third party evidence.

I think the opposite happens.

Because a measurement is not automatically proof.

The same study found strong correlations for signs such as wrinkles, pigmentation, pores and sagging, with weaker performance for some signs such as lip dryness. Accuracy on tablet and smartphone images held up, but was lower than on high resolution DSLR images.

Another large review of dermatology AI found lower performance on darker skin tones than lighter ones, which is particularly relevant when thinking about these technologies in India.

So we are about to get more skin data and more questions about what that data actually means.

Did pigmentation improve because of the serum? Or sunscreen? Or reduced UV exposure? Or menstrual cycle changes? Or lighting? Or the algorithm?

Measurement tells you that something changed. Good evidence tells you why you should believe the product caused it.

Consumers are already moving in this direction

McKinsey's 2025 beauty research found product quality was the most commonly cited reason consumers repeatedly buy from a beauty brand, while effectiveness was also among the leading factors. It describes consumers as increasingly skeptical of hype and focused on whether products actually deliver.

Mintel's 2025 beauty and personal care research points in the same direction, highlighting science backed claims and the evidence behind them as growing drivers of consumer trust, even as many consumers say they find it hard to trust new ingredients.

The founder takeaway: keep a claim ledger

If I were building a skincare brand today, I would start maintaining a simple claim ledger for every hero product.

• What exactly are we promising?
• What measurable outcome corresponds to that promise?
• What evidence do we currently have?
• Was that evidence generated on the finished formulation or merely on an ingredient?
• What population was tested?
• Over what period?
• What would happen if a customer independently measured the same outcome?

That last question is new.

And I suspect it becomes increasingly important.

It is also one reason we have been thinking so deeply about claims and evidence at The Clean Sheet. The interesting problem is no longer simply helping consumers decode ingredients. It is connecting what a product says to what the brand can actually prove. That is the central gap The Clean Sheet is designed around.

Who controls the measurement

For years, beauty brands controlled the claim and consumers controlled the belief.

We may be entering a world where consumers increasingly control the measurement too.

That changes the economics of credibility.

And I think founders building today should prepare for it before the dashboards arrive.

Sources and further reading

1. 1Lumeria (Lumoscope), Y Combinator Summer 2026. A multispectral phone camera (RGB, UV, polarized and near infrared) that tracks skin over time; the company reports its first batch sold out in under 24 hours, with the device listed at $299. View source
2. 2L'Oréal Groupe: Cell BioPrint, unveiled at CES 2025 and piloted with Lancôme in 2026. A lab on chip device that analyses five protein biomarkers and returns a personalised skin assessment in about five minutes. View source
3. 3Perfect Corp. AI Skin Analysis, which assesses up to 15 visible skin parameters from facial images and can track them over time. View source
4. 4Lee et al. (2026), “Artificial Intelligence Based Skin Analysis Models for Predicting Visual Grades and Device Measured Physiological Values From Facial Images,” Skin Research and Technology. 1,099 participants; on DSLR images, a mean of 93.3% of predictions fell within one grade of dermatologist grading across facial signs, with high correlations for signs such as wrinkles, pigmentation, pores and sagging, and lower but robust accuracy on tablet and smartphone images. View source
5. 5Daneshjou et al. (2022), “Disparities in dermatology AI performance on a diverse, curated clinical image set,” Science Advances. Reported lower AI performance on images of darker skin tones, a well documented equity concern in dermatology AI that is particularly relevant for India. View source
6. 6McKinsey and BoF (2025), The State of Fashion: Beauty. Product quality is consumers' top cited reason for beauty purchases, with shoppers described as value conscious, skeptical of hype and focused on whether products deliver. View source
7. 7Mintel, Global Beauty and Personal Care Trends 2025. Science backed claims and the evidence behind them are highlighted as growing drivers of consumer trust, alongside difficulty trusting new ingredients. View source

This article is an opinion piece for founders and product teams. Company descriptions and study figures are cited from the sources above as reported by those companies and researchers; specific numbers and methods should be confirmed against the originals before use in any product claim.

Images, in order of appearance, courtesy of Unsplash and their photographers, including abillion, Luke Chesser and Bee Naturalles. Used under the Unsplash License.