Supernova Award Category

The Problem

Skincare has been identified as the most difficult category to shop, driven by confusion at shelf (in-store and online).  In the midst of confusion, consumers are seeking consultation for products at high-priced skincare counters and specialty stores. They end up spending more than they need to on products they may not end up liking. 

The Solution

Olay Skin Advisor is an algorithm-based beauty consultant rooted in Deep Learning. Based on a database of tens of thousands of selfie images, the Deep Learning algorithm behind Olay Skin Advisor has learned what a woman’s face looks like at any age to an exceptionally precise level. With each new selfie, the technology labels it with the chronological age of the person, stores its information and ultimately learns what faces of a certain age look like, literally getting smarter with every use. Furthermore, Olay Skin Advisor asks users a variety of questions to determine her skin care product preferences, and objectively recommends a personalized solution for her. 

The results

Skin Advisor transformed the way our organization operates. In addition to helping fuel Research & Development, we've unlocked a new channel to distribute skin consultations, advice and product recommendations. 

We discovered a way to use a deep learning algorithm to analyze the skin using a selfie  to detect areas that are driving her skin age and is able to analyze these aging areas with 99% precision.   The tool’s proprietary technology then determines the “best” and “improvement” areas and also looks at images of women 10 years younger to determine a highly personalized regimen, based on her preferences, that can help improve her visual skin age by 10 years. 

Metrics

Skin Advisor launched in early August of 2016. Results from early adopters are very positive. In initial studies 100% of women reported that they would like to use at least 1 of the products recommended and 94% of women believe that the products recommended will give them better results better than what they are currently using.

The Technology

Olay's approach to analyzing visual age is a patented process. We have trained a deep learning model with 50,000 facial images paired with actual ages. The output is a model that can detect a person’s key aging areas and determine how old they actually look. Calculations derived from a heat map of pixels create the aging model. With this heat map we can then determine the relative contribution of the different areas of one's face.

​

Disruptive Factor

We’ve broken the paradigm of women needing to go to the counter or specialty store where they are “sold” a “personalized” skincare solution. We’ve brought algorithms to skincare. Finally, women can have a game plan for knowing what areas of their face to prioritize with their skincare regimen. Now women can finally be confident that they are buying the right products for their skin to tackle their concerns.

Shining Moment

Women have come to rely on technology to remove the friction from their daily lives, but have yet to require or rely on technology in the beauty space. Our prediction is that having a beauty counselor in the palm of your hand could be a new behavior in the industry as this tool gains momentum and user adoption. Gone will be the days of the high pressure counters. In the future, women may turn to skin advisor as their first stop for skincare advice.