Facts about scents
Smell is an important sense as it can alert us to danger like gas leak, fire or rotten food but also is closely linked to parts of the brain that process emotion and memory. We immediately get a feeling when we smell something. Fragrances are added to many products as it makes us feel good. If we feel good when we buy a product then we are likely to continue buying it. Sound and smell are increasingly becoming areas of huge opportunity for a brand experience. Abercrombie & Fitch, for example, is widely known for spraying its trademark scent in stores worldwide. Lush Fresh Handmade Cosmetics, soaps and cosmetics retailer, relies on exhaust systems to remove some of its signature potent aroma from stores. The aroma of warm cinnamon rolls that are baked every 30 minutes to entice you to try items at Cinnabon. For Starbucks, aroma is especially important in offering the best possible customer experience. Ikea uses the smell of food to sell more than food.
I was so excited to put together my two passions: fragrances and technology.
The initial idea was to have an app, smell recognizer, that could recognize the perfume one is wearing. For example, you are passing by someone whose smell you like, take out a smell detector and snap. The detector would analyze what the smell consists of and results would then be entered into an application. Similar to object recognition in images, the app would give you all matching perfumes and percentage of certainty.
How to accomplish?
- Aromabit, Tokyo-based company made an odor imaging sensor technology. It behaves as an e-nose device, sniffing the air, and is of a size of SD card. It generates aroma-code that can be stored and used to analyze smells.
- Fragrantica, the world’s largest perfume database. Get the data, perfume accords and tones. Use machine learning to train a model with acquired data.
The problem that occurred was the price of the sensor. It was way above our limits. We haven’t considered other smell detectors because of their size and specifications.
Let’s concentrate on producing smells.
Idea 1: Track visitors in stores. What day of the week it is. What’s the weather like? How many of visitors are men, woman, teenagers? Analyze the data and produce scent per department. If there is a campaign for the certain brand, diffuse its associate scent.
Idea 2: Create personalized scent per customer; associate fragrance with each product and create new one, for each customer, based on their shopping history.
For more ideas, please refer to following blog post.
How to accomplish?
It is connected to a computer or raspberry pi via USB cable. To communicate to the port (smell diffuser or a big fan), a message is to be sent to the COM port, in the format OUTxx1 to activate/OUTxx0 deactivate it (xx is a number of a port).
Here is an example code:
The main problem with those ideas is ventilation. If new scents are diffused frequently, space must be aired.
The diffuser we have worked with has a limitation in a number of capsules at the same time. Only 4 capsules can fit into the device. It is not convenient to remove one and insert the other capsule frequently. The other thing is that aroma of capsules is already complex, for example, aroma of leather, of a factory, of a perfume, of winter. Depending on the idea that can be either ease or a limit.
We tried to explore the world of smells and to play with digital smells. We wanted to share our thoughts and findings with you and ask you to share yours. Together, we might think of an interesting and realistic idea and make our digital world a nice smelling place.
Comment or send us an e-mail to: firstname.lastname@example.org