FAQ

No. An AI skin analysis describes cosmetic skin attributes such as moisture, redness, pigmentation and pore size. It does not diagnose, and it is not a medical device. Thea Care's analysis is built with dermatological expertise and trained on dermatologist-annotated images, which is why the cosmetic assessment holds up, but a consumer noticing a changing mole, a persistent rash or a lesion should see a dermatologist rather than a shop.

No, and brands that run both tend to describe it the other way round. The gap an AI skin analysis fills is the shop, where nobody is standing next to the customer to explain why one serum suits her and another does not. That conversation already happens in store; online it does not happen at all. Where staff are present, the analysis gives them a starting point for the conversation rather than a script to read out. Thea Care is white-label, so the brand decides what the tool is called and how prominently it sits next to human advice.

This is the right question to ask any vendor, and most cannot answer it. Thea Care separates two things instead of merging them. Direct revenue is an order containing a recommended product, tied to the session in which the analysis was completed. Indirect revenue is a later purchase by the same consumer, reported as its own line rather than folded into the headline. Both are then held against the shop's own baseline: the conversion rate, average order value and revenue per user from before the analysis went live, or measured alongside it during the validation phase. That is a comparison against your own numbers rather than a controlled experiment, and we say so.

Partly, and the part that changes is not all error. Skin genuinely moves with sleep, cycle, season and hydration, so a moisture or redness score that shifts between Monday and Friday may be reporting something real. What a vendor can control is the capture: Thea Care runs an automatic image quality check that rejects blurry or badly framed uploads before analysis, and the camera guidance handles distance and lighting. The same photo always returns the same scores. Across different photos, individual parameter values move more than the recommended routine does, because the routine is derived from the overall picture rather than from a single score.

Three cases, and they are worth naming plainly. A brand without its own online shop has nowhere to place the analysis and no range to recommend from, so there is no revenue case to measure. A reseller without an own range can run an analysis but cannot control what it recommends, which is usually the opposite of what the retailer wants. And a shop below roughly 10,000 monthly visits will not produce the analysis volume at which the investment returns itself; between 10,000 and 25,000 visits it depends on how prominently the entry points sit. Thea Care would rather say this in the first call than after a validation phase.

Switching AI skin analysis vendors is mostly a catalogue and rules migration rather than a rebuild. The product data moves across as a Shopify export, CSV or JSON feed, the links behind each recommendation are adjusted to the new shop, and the result is tested with several people with different skin before launch. Thea Care does this during the validation phase, so the new analysis already runs and is measured before the old contract has to end.

Performance across skin tones depends on the training data, and it is a fair question to put to any vendor in writing. Thea Care classifies skin tone with a machine-learning model that maps each face onto the Fitzpatrick scale, and measures the same eight parameters regardless of tone. Ask vendors how their models were evaluated across tones rather than accepting a general statement about inclusivity.

Published studies put agreement between AI skin analysis and dermatological assessment at roughly 69 percent to over 90 percent, depending on the vendor, the method and the concern measured. The lower figure describes per-class agreement across many fine-grained categories; simply detecting whether a feature is present scores higher. No AI skin analysis is a diagnosis. Ask vendors which study, which method and which concerns a quoted number refers to.

An AI skin analysis is software that scores a person's skin from a selfie and turns the result into product recommendations. An AI computer vision model measures individual parameters such as moisture, redness, wrinkles or pigmentation, and a recommendation layer maps those scores onto a product catalogue. In e-commerce it is normally white-labelled, so consumers see the brand rather than the vendor. Thea Care measures eight parameters and returns a routine from the brand's own range.

No. Thea Care GmbH is a B2B company based in Berlin, Germany, registered as HRB 252846 B at Amtsgericht Charlottenburg, selling white-label AI skin analysis to skincare, cosmetics and pharma brands including Weleda and sebamed. It is not affiliated with the consumer app "Thea" by Thea Technology Inc., nor with Laboratoires Théa. Thea Care has no consumer app of its own: consumers meet the technology inside a brand's own shop, under that brand's name.

Yes. Skin condition scores inferred from a facial photo count as health data and fall under Art. 9 GDPR as a special category. That means a cookie banner or a click on a start button is not a sufficient legal basis: explicit consent under Art. 9(2)(a) is required before the image is processed. Thea Care collects it on a dedicated consent screen and runs the analysis itself under Art. 6(1)(b).

Four things decide it: explicit consent under Art. 9(2)(a) collected before any image is transmitted, a documented legal basis for the analysis itself, defined retention periods with automatic deletion, and an Art. 28 data processing agreement between brand and vendor. Server location matters as a fifth. Thea Care shows a consent screen before the analysis starts, keeps the result for the session only, processes in Frankfurt, and signs an AVV with every brand during onboarding.

In a GDPR-compliant AI skin analysis the photo is processed during the session and not kept afterwards. Thea Care pseudonymises the image, analyses it on servers in Frankfurt, Germany, and deletes it immediately once the analysis is complete. Storing images to train the models is possible, but only where the brand enables it and the consumer gives separate explicit consent under Art. 9(2)(a). Where a brand does not enable it, no image is retained at all.

An AI skin analysis pays for itself once it produces around 500 completed analyses a month, because that is the point where the additional orders it drives outweigh the licence. How much traffic that takes depends on where the entry points sit: with entries on collection and product pages, roughly 2 to 5 percent of sessions end in a completed analysis, so the threshold falls somewhere between about 10,000 and 25,000 monthly sessions depending on placement. Use the session count from your own shop backend rather than a third-party traffic estimate, because the two often differ by more than the decision itself turns on. Average order value shifts the threshold as well, since a higher basket needs fewer analyses to produce the same revenue. Below roughly 10,000 monthly sessions the investment does not return itself regardless of placement.

We take data privacy seriously and operate in full compliance with GDPR regulations from our servers located in Frankfurt, Germany. For enterprise customers, you maintain complete control over your data with the ability to delete information at any time, and we only store what you explicitly choose to save. We have Standard Contractual Clauses (SCCs) and Technical Implementation Agreements (TIAs) in place with all third-party providers to ensure comprehensive data protection throughout our system.

An AI skin analysis runs in the mobile browser, so consumers need no app download and no special hardware. Any smartphone camera of recent years is enough. Thea Care opens from a link in the shop, on the brand's own domain, and the same analysis works on tablet and desktop; most consumers use it on a phone, which is where they already shop.

Four criteria separate AI skin analysis vendors: whether the tool can attribute revenue at order level or only reports add-to-cart, where the analysis runs and under which data-processing agreement, whether the recommendation logic is transparent and brand-controllable, and how deep the shop integration goes. Thea Care attributes orders individually, runs analysis on servers in Frankfurt under an Art. 28 GDPR agreement, and lets brands set their own recommendation rules.

Yes, our platform supports up to 25 languages, enabling you to serve customers across different markets with localized experiences that resonate with their preferences and expectations.

AI skin analysis software is normally sold on an annual contract. A typical engagement opens with a validation phase of one to three months, in which the analysis runs live in the shop and its numbers are measured, followed by a yearly rollout term. Some contracts run over two years, and monthly terms are possible at a higher cost. Thea Care follows exactly that structure: one to three months of validation, then an annual rollout contract.

Depending on your chosen plan, we provide comprehensive support including weekly success calls, thorough onboarding, and additional training sessions as needed. Our goal is to ensure your team feels confident using the platform and maximizing its value for your business.

Yes, though it is normally structured as a validation phase rather than a free trial: a defined period of one to three months in which the analysis runs live in the shop and the resulting conversion and revenue figures are measured. Thea Care uses that validation phase before any annual rollout contract, so the decision is made on the brand's own numbers rather than on reference cases.

An AI skin analysis scores the skin from a selfie. Thea Care measures eight parameters, among them moisture, oiliness, redness, wrinkles, pigmentation and pore size, then matches those scores against the attributes assigned to each product in the brand's catalogue. Age and gender estimated from the same photo can feed the rules as additional inputs, from the Standard tier upwards. The logic is transparent and the brand sets its own rules.

An AI skin analysis returns a routine of at least three steps: cleansing, a treatment step and moisturising. The practical floor is therefore a catalogue covering those three steps for normal, dry, oily and combination skin, in practice about 10 products. Below that, different analyses start returning the same routine. There is no upper limit: Thea Care also runs catalogues of 300 SKUs and more, where the extra range buys longer routines and recommendations targeted at specific skin conditions or ingredients.

AI skin analysis software is billed in three ways, frequently as a combination of them: a fixed monthly or annual licence, a fee per completed analysis, and a one-off setup fee. Which mix applies depends on the vendor and on how permanently the analysis runs in the shop. Thea Care bills annually, on one of three feature tiers: Basic, Standard or Premium. Monthly terms are possible and cost more than the annual equivalent.

Yes, we've built integrations with all major e-commerce platforms to ensure smooth implementation regardless of your current technical setup. Our platform is designed to work seamlessly with your existing infrastructure without requiring significant technical changes.

Absolutely. Maintaining your unique brand identity is crucial to us, and we ensure each brand receives a fully customized experience. You can personalize everything from colors and logos to email templates and the entire customer journey flow. This ensures your customers experience a seamless extension of your brand rather than a generic third-party tool.

Results vary by traffic, catalogue and placement, so we publish the individual numbers rather than one headline average. Each comes from order-level tracking in the brand's own shop: Physiogel +258% conversion, +34% AOV and +380% ARPU. NKM Naturkosmetik +91% conversion, +71% AOV and +243% ARPU. Judith Williams Cosmetics 3.5x conversion, +107% AOV and +72% ARPU. How each figure is measured is set out in the matching case study.

An AI skin analysis is usually deployed on its own subdomain, typically under the brand's own domain, rather than embedded in the shop template; iframe and API integrations are alternatives. Setup runs in three steps: styling the interface to the brand, importing the product catalogue, and configuring the recommendation rules. Thea Care integrates with Shopify, WooCommerce, Shopware, Spryker and Channable, gives brands an admin dashboard for every setting, and handles the configuration itself from the Standard tier upwards.

Published figures for AI skin analysis cluster around €200–600 per month for self-serve tools and €0.10–0.25 per completed analysis where billing is per scan; white-label deployments with order-level revenue tracking sit above that band. Beyond that, cost depends on the feature tier and on how many shops and markets the analysis runs in. Thea Care sells three tiers: Basic covers one market with core analysis, Standard adds all eight skin parameters, age-aware recommendations, Klaviyo and Shopify connections and A/B testing, and Premium adds special views, API access and custom integrations.

You can provide your product data through multiple convenient methods: Shopify integration, CSV upload, or JSON feed. The essential information includes product IDs, names, categories, images, descriptions, and usage instructions. Our AI then analyzes this data along with skin type and skin condition attributes to create intelligent recommendations, though you can also leave the recommendation logic entirely to our AI if you prefer. Our flexible mapping system allows you to easily connect your existing data structure with our platform requirements.

For medium to large clients, our typical setup process takes approximately two weeks from start to finish. During this time, we handle the technical integration, customize the solution to match your brand, and ensure all product data is properly configured for optimal performance.