GDPR | costumer segmentation | Sep 2019

Experience design

Targeted shopping experience through swiping

With the invention of digital technology, we have been given new ways to express ourselves. New terms and meanings as well: such as scrolling, chatting, commenting, liking, re-posting/tweeting and finally swiping. These are all perceived as relational expressions between the body/mind and the digital environment. Interaction is a crucial tool of design, that stimulates people and brings out their sense of collaboration and participation. This project was about unfolding the information potential inherit in these gestures.

 

Online shopping has revolutionized the way we buy things. Millions of items made available with the click of a button. However, unlimited choice is not always desirable. It can lead the user on a journey of untargeted and unsatisfying scrolling through a bottomless page that continues to load in more and more results.

 

The goal of the program was to introduce a new, intuitive and fun way of browsing an online shopping catalogue using a technology that everyone is familiar with: swiping. The application will show you a stack of products and it is then for you to decide whether to swipe right for “like” or left for “dislike”. Swiping upwards (a.k.a. “superlike”) directly adds a product to your shopping cart.

 

The algorithm will first assemble a stack of items to swipe on, based on a first guess what you might like. This stack of suggestions is then continuously updated as you proceed to swipe items. The backend algorithm analyses not only what you swipe left or right but also HOW you swipe it. We developed several meaningful features that characterises this “how” through simple metrics like; swiping speed, acceleration and reaction time. But also, second-order features like the length of the swiping trajectory and a hesitation coefficient.

 

In testing the application, we did a preliminary analysis, in which 60 test subjects swiped images on an online platform. This showed that users indeed expose different general swiping behaviours. Some swipe faster and react quicker than others. Some hesitate before swiping, others make up their mind right away. This enables two key mechanisms: First, users can be matched to a user group and a behavioural pattern in a more meaningful way. Second, swipes can be characterised in relation to how a user normally swipes. Imagine someone that usually takes long to decide on whether he likes a product or not. Then comes an item which he swipes right fast and without hesitation. We believe that this can tell us a lot about the user’s attitude towards a product, and we wanted to leverage exactly that.

 

The application is also a new way of navigating users straight to their purchase through swiping analysis. We give meaningful and targeted suggestions where user information previously seemed to be unavailable. This narrows down the vast choice in the online shopping world, leading to buying decisions that are quicker and more convenient for the user.

 

The project team consisted of myself, Vera Fristed [SWE], Johannes Leonhard Rúther [GER] and Valentin Viennot [FRA].

IFrisbæk

StudioI

I

6x Final presentation board, original  14000x8000px

Swiping analytics on browser version

Mapping of physical watch store visit

credits

I hope that you find my work inspiring and I encourage you to use it as much as you like. I do however demand that you credit my work.

 

© 2019 Mikkel Frisbæk Sørensen

Website

This website is created with the intend to showcase undergraduate projects of my studies at Aarhus architecture school as well as personal projects. It is shared as an online work folio – and maybe an inspiration for others.

GDPR | costumer segmentation | Sep 2019

Experience design

Targeted shopping experience through swiping

With the invention of digital technology, we have been given new ways to express ourselves. New terms and meanings as well: such as scrolling, chatting, commenting, liking, re- (posting/tweeting) and finally swiping. These are all perceived as relational expressions between the body/mind and the (digital) environment. Interaction is a crucial tool of design, that stimulates people and brings out their sense of collaboration and participation. This project was about unfolding the information potential inherit in these gestures.

 

Online shopping has revolutionized the way we buy things. Millions of items made available with the click of a button. However, unlimited choice is not always desirable. It can lead the user on a journey of untargeted and unsatisfying scrolling through a bottomless page that continues to load in more and more results.

 

The goal of the program was to introduce a new, intuitive and fun way of browsing an online shopping catalogue using a technology that everyone is familiar with: swiping. The application will show you a stack of products and it is then for you to decide whether to swipe right for “like” or left for “dislike”. Swiping upwards (a.k.a. “superlike”) directly adds a product to your shopping cart.

 

The algorithm will first assemble a stack of items to swipe on, based on a first guess what you might like. This stack of suggestions is then continuously updated as you proceed to swipe items. The backend algorithm analyses not only what you swipe left or right but also HOW you swipe it. We developed several meaningful features that characterises this “how” through simple metrics like; swiping speed, acceleration and reaction time. But also, second-order features like the length of the swiping trajectory and a hesitation coefficient.

In testing the application, we did a preliminary analysis, in which 60 test subjects swiped images on an online platform. This showed that users indeed expose different general swiping behaviours. Some swipe faster and react quicker than others. Some hesitate before swiping, others make up their mind right away. This enables two key mechanisms: First, users can be matched to a user group and a behavioural pattern in a more meaningful way. Second, swipes can be characterised in relation to how a user normally swipes. Imagine someone that usually takes long to decide on whether he likes a product or not. Then comes an item which he swipes right fast and without hesitation. We believe that this can tell us a lot about the user’s attitude towards a product, and we wanted to leverage exactly that.

 

The application is also a new way of navigating users straight to their purchase through swiping analysis. We give meaningful and targeted suggestions where user information previously seemed to be unavailable. This narrows down the vast choice in the online shopping world, leading to buying decisions that are quicker and more convenient for the user.

 

The project team consisted of myself, Vera Fristed [SWE], Johannes Leonhard Rúther [GER] and Valentin Viennot [FRA].