This project helped resolve a consumer pain point that LION had not fully addressed in their latest sales strategy for Aroma Rich. Our creative use of technology improved product usage and the experience of each consumer. With machine learning, we developed an image analyzing system that could identify the visitor’s color preferences and correlate that data with their preferred scents in a comprehensive database. Not only did our solution identify olfactory preferences hidden in the unconsciousness of an individual’s mind but also helped them visualize their personalized scent, which is by nature naked to the human eye.

Aroma Rich fabric softener was launched in 2010 with a mission to provide well-being and comfort to Japanese women. In 2018, LION adopted a new sales strategy that enabled women to create their own aroma by mixing the 5 products in the range. But soon discovered that the majority found it hard to find their favorite blend on their own. To help women in this quest, we created a first-of-its-kind sensory journey where women could easily discover their favorite scent. Its objective was to allow them to experience the effect of wearing a personalized scent on their clothes during summer and encourage them to use Aroma Rich at home. Those who could not attend the 4-day installation were invited to participate in the personalized experience online.

The creative idea was a first-of-its-kind sensory journey, where based on only one photo, visitors were able to effortlessly discover the Aroma Rich blend that suited them best. The target was Japanese women around their 30s dealing with the summer heat. Based on a correlation that exists between the colors and scents an individual prefers, a full body analysis integrated with machine learning determined the color ratio of each visitor’s clothes and correlated their preferred scents. The color analysis result was illustrated in real-time onto a folding fan, traditionally known in Japanese culture as a symbol of summer. In parallel, printers were set up with a special mold to print the same colored illustration including the visitor's name directly on another fan. Within 7 minutes, visitors could smell, see and feel their signature scent for the first time, as the perfect gift to beat the heat.

LION had not fully addressed a consumer pain point in their latest sales strategy for Aroma Rich: the majority of women could not find their favorite blend on their own. Our challenge was to develop an automated system that could easily identify the favorite scent of any given individual. The creative idea arose from the fact that Aroma Rich fabric softener is a product intended to be used for clothing, thus we considered that a women’s choice of color for clothing was the perfect starting point to identify her favorite scent. After visiting the event or mobile microsite, women would be able to easily make their personalized blend at home as well as be reminded of their experience and sentiment towards the brand as they glimpsed at their uniquely-designed folding fan. Our creative use of technology strongly improved product usage and the experience of each consumer. Not only did our solution identify olfactory preferences hidden in the unconsciousness of an individual’s mind but also helped them visualize their personalized scent, which is by nature naked to the human eye.

The main creative output from this campaign was that women had the chance to be gifted a tailor-made blend of Aroma Rich and a uniquely-designed folding fan to beat the summer heat. Our creative use of data & technology enhanced these outputs by creating a personalized experience through:

- The formulation of a tailor-made aroma catered to the taste of every individual by only taking one full-body shot.

- The visualization of each individual’s tailor-made aroma by illustrating the image analysis result directly onto a folding fan.

Our data journey started with the upload of one full-body shot to the server. We then used machine learning technology such as Deep Lab to identify the human and non-human elements of the image. Open Pose tracked bone structure to identify lower limbs and arms, eliminating uncovered skin areas since people wear more revealing outfits in summer. Only after that could we analyze a body cut-out image determining the exact colors of their clothes & correlate their preferred scent. As the analysis was projected in real-time, printers were set up with a special mold to illustrate a folding fan in the exact same ratio as their clothes.

We only displayed billboards for 7 days at the nearest station to the venue, but it soon became the talk of Internet & TV programs, cited 134 times. The equivalent in paid media would usually have cost us 100M JPY. Even if the venue could only host 50 people at the same time, we attracted 1,219 visitors in 4 days. In 1 month, 10,569 women had participated in the experience on the website. As a result, LION’s KPI doubled and 92.9% of women felt a positive change of attitude towards the brand.