To answer the question of "is this cosmetic a good choice for me?", we have developed a concept of a "skincare match" and underlying skincare matching technology able to deconstruct skincare products. Our complex system, powered by Artificial Intelligence, works by assessing product effectiveness for different skin types and conditions in light of scientific evidence and by emulating and embedding the knowledge and expertise of skincare professionals.
By analyzing tens of thousands of products and comparing them to each other, we have derived a scoring method by our product matcher. This scoring method is inherently relative, which means that a product score has to be interpreted in a relation to other products within the same category. For example, if a moisturizer scores 30 on a dimension of "dry skin type", it means that 70% of other moisturizers in our catalogue are a better choice for dry skin than this one. It may still be that this moisturizer is a good product on its own, but it simply scores poorly in comparison to other, better-matched products.
A well-matched product goes beyond its effectiveness and includes lifestyle and personal choices. Our final product match is determined by taking into account all individual scores and users' personal preferences.
The problem of finding a well-matched skincare product that will be able to help your skin and meet your standards is not an easy task. We are surrounded by conflicting information, unreasonable marketing claims, and word-of-mouth advice that is never really accurate.
At Skin Bliss we believe that a task that is difficult for a human might be easy for a machine. More precisely, for an intelligent computer program. This is the core innovation behind Skin Bliss and what separates us from any other app or website. We have been developing an Artificial Intelligence engine designed to understand, deconstruct, assess and recommend skincare. Our system is based on approximating and assessing product formulation in light of scientific evidence, ingredient data, and regulatory laws on the one hand, and detailed textbook and expert-knowledge information on skin types and skin concerns on the other hand.
We have been working with skin(care) experts to help teach our algorithms best practices and guide us in deconstructing the complexity of skincare science. While we already use this system in our app, please note that it is under continuous development and improvement.
Every person and every person's skin profile is unique. Not only do we all have a slightly different biological predisposition, but our lifestyle and personal beliefs are different. The perfect skincare match goes way beyond the product's efficacy. The perfect skincare match respects your life choices.
I'm searching for a halal cruelty-free cleanser that is suitable for dry, acne-prone skin with eczema...
You cannot just google it. There are literally tens of thousands of possible cleansers to choose from. Not all products have been formulated equal, and most importantly not all products are good for everybody. Because we all have unique personal skin profiles only a narrow selection of all possible cleansers will be suitable for us. You can use Skin Bliss to help you find such products. Not only do we filter the products for you, but we also evaluate each of them on different dimensions and levels of being your match.
A general concept of a match is rather complex, and it cannot be captured by a single arbitrary number. Product match needs to be evaluated on different levels. Skin Bliss can distinguish the following levels:
All of these dimensions are taken into account to estimate your match. Please note, all of these single scores are relative, we always show how far a given product is from the "perfect product" that we could identify on a given dimension from our database (for more explanation see below). Currently, the overall match (total score) is a formula set by us, but in the future, we might allow more flexibility and users will be able to set relevance value to their overall score.
Let's assume you are looking for a cleanser. Assuming that you have dry acne-prone skin with eczema, we can distinguish 3 main subgroups of cleansers that make up your basic skin profile:
We can also see that there is a certain overlap between these products. We have cleaners that are:
And then, lastly, there is a subset of these cleansers that are suitable for dry acne-prone skin with eczema.
Skin Bliss looks at each of the dimensions and ranks products according to how good they are on a given dimension compared to all the other products from this category.
For example for category cleansers and the dimension dry skin, the algorithm goes through all cleansers and asks:
"is this cleanser more or less suitable for dry skin compared to the previously checked cleanser?"
The outcome of this process is an order of cleansers based on how suitable they are for dry skin. By nature of ordering, there will always be the "worst" and the "best" option. Even if the "worst" one is not bad on its own, it is still worse than all of the other options. The worst/best labels are only in relation to all the other product:
Here we see a simplified example of 6 cleansers. We order them depending on how good (suitable) they are for dry skin (again we do it by asking for each of the cleansers whether is more or less suitable for dry skin compared to a previously checked cleanser). And from this we derive our numeric scores. If all other cleansers are better than cleanser A, this cleanser gets a score 0. If 4 out of 6 cleansers are worse than cleanser B, it gets a score of 67.
What does it mean for a product to be more / less suitable on a given dimension?
Different products have different formulations. And not all formulations are suitable for a given skin type or skin concern. Some ingredients can either worsen a given concern or simply not have a good influence on a skin type. In this case, products with the strongest potential negative effect are ordered on the left (red) side of the suitability scale and products with the strongest expected positive effect are ordered to the right (green) side of the scale.
There can also be cases where even if there is nothing red indicated for a given product, you still find it on the left (red) side of the suitability scale. How come? Simply because such products contain less beneficial ingredients than products on the green side. It doesn't mean they are bad or wrong, just very basic and score poorly compared to the other richer products.
The product in the green bottle contains only three types of ingredients. The product definitely does its job. It's cleansing, but that's it. You will not get any extra benefit or protection from using it. It still can be a good product for you, but we simply tell you that there are better options, like for example the product in the dark blue bottle. It contains extra ingredients that have a beneficial effect on your skin. But it is your choice to decide which of these products to use.
Here we look at the example of actual output from our algorithm for different cleansers for dry skin. We can see how progressively the ratio of positive to negative ingredients increases when we look at examples of different products with different scores. To emphasize it again: these are relative scores, so 1% (or score 1) means that 99% of other cleansers available in our product catalogue are more suitable and beneficial for dry skin than this one. The highest-rated of those 4 products has the additional remark **Highly focused formulation. That means that almost all of the ingredients used in the product individually have a positive effect on dry skin.
Creating a skincare recommendation system that can work for anybody in the world is not an easy task. Skincare science is complex and there are many parameters that can, in principle, influence how well our skin will react to a given product, and it's not always clear which of these parameters are the relevant ones.
Currently, we are two people working on Skin Bliss full time, and we are both well equipped to carry out the task of developing such a system from a technical perspective. We both have PhD degrees in the fields related to Artificial Intelligence, industry experience in data science, and most importantly, we are huge skincare enthusiasts. We are also very lucky to have found skincare experts who are guiding us in this journey and making sure our algorithms show reasonable outcomes.
We believe that diversity is the key to great progress, the diversity of opinions, points of view, cultures, and experiences. This is why we welcome all comments and criticism from your side. Especially if you are a skincare professional, we value your input and your opinion and we'd love to be challenged. If you would like to comment on our methodology, criticize it or simply share some observations or insights, please email us at firstname.lastname@example.org
- Maria & Marvin, the founders