It is harder for women of color to buy cosmetics. They often leave retailers without buying anything because they can’t find products that suit their skin. Even conscientious retailers can face difficult tradeoffs in catering to women of colour, as any kind of minority product can divert shelf-space away from higher-volume items.

Finding long-tail products online may be the best option for those that struggle finding products in store. Unfortunately, buying new makeup online is invariably a gamble. Screens render colors differently, so it’s impossible to tell what shade you will actually get. To make good decisions about skin care It often feels like you need a masters in biochemistry. It’s hard to feel confident in your purchases unless you are able to try it first.

How to bridge the online cosmetics gap?

There are companies that specialize in facial recognition systems to help customers find the right product for them. Most popular of this are virtual try on systems that allow users to “try on products' ' using their webcam. These services use facial recognition technologies to locate various key regions of the face, the eyes, the lips, the cheeks and superimposes a makeup layer on the user’s face. in theory this would be a great way for people with darker skin tones to buy makeup. However historically technology has been created for white people. Film couldn’t render dark skin. Foundation only worked on lighter skin tones. Facial recognition software doesn’t work as reliably for people with darker skin. 

Virtual try-on systems fail women of color. Assuming the technology even recognizes faces with dark skin tones, the next problem is that try-on services may alter the user’s face to appear more “white”. Faces may become edited. Face shape changes, noses made thin and skin tone lightened, resulting in an unnatural look that reinforces caucasian ideals of beauty.Promoting an ideal standard of beauty that is euro-centric is not going to be effective in western markets for very long. Increasingly diverse demographics will lead to increasingly diverse standards of beauty. Brands need to offer solutions that work for people of all skin tones and genders, and pick suppliers and technology partners that represent those values.

The next hurdle is user-experience design. An intelligent recommendation system ought to be able to select intelligent defaults. When a POC goes to try on a lipstick shade, the first shade is one that looks good most users but it’s often too light for dark skin and looks bad on them. Good IT systems should be able to adjust the first product a user sees so that it is tailored to that user. Online retail is the one place where it should be incredibly easy to find the right products. Unfortunately what we see time and time again is that tech companies over look women (and the industries that serve them) and minorities.

Brands who have these problems with their technology run a deeper risk. It doesn’t look good when a system is biased.

There is a solution. 

How AI tools can help Diversity in cosmetics

Buying cosmetics online should be even easier then looking in store. Online search engines can instantly search though a catalog of a million products, technology can assess a user’s skin with incredible accuracy and can build custom experiences for users. Quizzes and facial recognition services should make online cosmetic purchases easier, not harder.

I firmly believe that AI tools, especially recommendation tools can help women with dark skin tones find products that suit their skin, if these services are designed well.