FragranceFit

In the last couple of months I've worked with some awesome ML engineers to develop a Computer Vision based solution for analyzing the facial expressions a customer has after testing a fragrance in a store. We were able to get 2nd place in an Innovation Contest at Europe level with this idea! In this article I am going to explain the technical solution and the challenges we faced.

But first let me start with a short introduction...

In today's fast paced world where there are millions of options and fragrance stores all over the world, targeting the customer with the best possible product recommendation is essential to selling a product/closing a deal. Fragrances are a hard sell for online stores because the clients prefer to smell it before actually purchasing it. There is a real need for a solution that attracts the potential customers to the stores and encourages them to experiment and try out their next favourite fragrance.

At the same time, when a potential client enters a fragrance shop, they might not be open to talk with a consultant or there might not be enough consultants available to recommend a fragrance that would be a good fit for the customer's taste. The aforementioned solution should seamlessly integrate in the store's environment and enable consultants to intervene with perfume recommendations that have a high chance to make the visitor say "Thanks, I'll buy this one", instead of "No thanks, I'm just looking".

Our solution is based on a ML Computer Vision system built around an array of smart cameras placed all over the store that would capture the facial expressions of the potential clients while they are testing a fragrance, understand their emotional response to the fragrances, construct a profile of preferences and recommend the best matching perfume for their particular tastes.

All this data would be captured in a database and be available in real-time to the store consultants on a mobile app so that they could come and directly recommend a particular fragrance based on it.

Click on the image below to see the presentation video

We've also done a nice social media campaign to raise awareness and see how would such a solution be accepted within regular perfume buyers. Check the picture below too see the results.