FRAGRANT MEMORIES FLOWER GIFT

TitleFRAGRANT MEMORIES FLOWER GIFT
BrandLES MILLE FEUILLES DE LIBERTE
Product / ServiceLES MILLE FEUILLES DE LIBERTE
CategoryE02. Innovative Use of Technology
EntrantDENTSU ISOBAR Tokyo, JAPAN
Idea Creation DENTSU ISOBAR Tokyo, JAPAN

Credits

Name Company Position
Isamu Tokiwa Dentsu Isobar Executive Producer
Takafumi Aihara Dentsu Isobar Creative Director
Takuya Nakanishi Dentsu Isobar Integrated Producer
Yuki Nishihata Dentsu Isobar Copywriter
Yosuke Kawasaki Dentsu Isobar Account Director
Hidefumi Miyahara ideaLump Technical Producer
Hiroaki Matsuoka REVORN Tech Lab Manager
Noriaki Koide REVORN Data Scientist
Masashi Imamura un-T system! Technical Producer
Tomoyoshi Kato un-T system! Art Director
Tomohiro Yamanaka un-T system! Front-End Engineer
Tsuyoshi Yoshida un-T creative media lab Interactive Director
Wataru Amano SIGNAL PR Producer
Shuhie Suzuki SIGNAL PR Director
Akiyo Kiyota SIGNAL PR Assistant Director
Miki Isogai SIGNAL Research Producer
Yoshiro Oku SIGNAL Research Producer
Daisuke Tomomatsu SIGNAL Researcher
Kazuhiro Narisawa Studio Gear Event Craft Producer
Hideto Sakai GrowEgg Event Producer
Toshiyuki Nakabachi alive promotion Event Producer
Yoichi Kasugai frasco Photographer/Film Director
Shidaka Gotou ideaLump Film Editor

The Campaign

Giving flowers is one of the oldest forms of expressing love and affection. Their power is transporting and can bring us closer to those moments. We found a way to ignite memories using flowers as the conductive agent. We delivered the world’s first memory data mining to select flower bouquets that could trigger precious memories. Using sensors and creating AI to emulate human's sense of smell, we analysed memorabilia and photos associated with such memories. We created Fragrance2Vec, the world’s first AI that understands scents and converts their data into vectors. Combining language and images, these were collated with a vast database of elements including flower scents, image data as indicators for scents of memorable places, and language data to supplement the scent’s image. By Maximising the Proust effect – the phenomenon of scent triggering memories – we delivered data mining of memories in a whole new way.

Creative Execution

In pursuing the expansion of flower sales in the market, Fragrant Memories partnered with the Tokyo Wholesale Market, the largest of its kind in Japan. Together, they entered a new territory driven by technology. Inspired by the connection between flowers and emotional memories, this campaign leveraged sensors, artificial intelligence and science to transform the experience of gifting flowers in the market. Unlike our other four senses, smell and its associated memories do not easily fade. These directly stimulate our bodies’ limbic system and trigger instinctive behaviors and emotions. Combining the characteristics of scents – a phenomenon known as the Proust effect – with a database of vectorized scents, flower bouquets were personalized using memorabilia to make recipients vividly revisit their memories. The campaign encouraged consumers to show affection to their loved ones by going live on November 17 and culminating on November 22, a day known in Japan as “Good Couples Day”, a day where married couples express their gratitude to one another. Liberte set the experience in the upscale Marunouchi neighbourhood where customers were invited to try the experience in-store by bringing their mementos and placing them inside the memory data mining device to generate and purchase an entirely bespoke flower bouquet that could trigger those memories for their significant other. The campaign generated outstanding results: Sales increased a whopping 300% on an year-to-year basis and hit the highest advertising equivalency value of 1 million US dollars, a historical record within the market segment. Fragrant Memories turned flower gifting into a deeply personalised experience, bringing together AI technology, data and human emotion to create a bond between loved ones and evolve a business model that has remained relatively unchanged over time.

Links

Supporting Webpage