Virtual wardrobe and personal stylist
Two innovative entrepreneurs approached us with the idea to disrupt the online fashion market.
Problem / Opportunity
With style bloggers and fashionistas getting more and more common place, society as a whole is getting more and more fashion conscious. Based on knowing your personal wardrobe, the entrepreneurs stipulated they can know your taste and style and are able to best recommend suitable, new fashion on offer around the web. All the while you can now organise your wardrobe and create your own looks without standing in your closet.
We paired them up with our innovation specialists, UX experts, creative director and senior architects and this innovation team went straight into ideation. In discovery workshops we sketched out the ideal story and UX for the envisioned iOS app, a branded marketing website and various marketing channels.
We proceeded with the development of the looks curation portal and iPhone app iteration by iteration. After the first iteration, we had a prototype ready to test with real users and the fully worked out UX and designs. This first version encapsulated the end to end transaction, from admin looks curation, over user signup, wardrobe upload and in-app look creation.
One key feature of Stylishly is the look curation with automated item replacement. The Admin creates new generic looks and our machine learning based algorithm analyses a user’s wardrobe, picks one or 2 suitable items that the user has a high propensity to use, and puts them into the generic look, there by personalising the look. This ensures each user regularly gets highly personalised, relevant new looks suggested, without a human curator having to do all the legwork.
To train our ML model, we stipulated a few exclusion criteria for which items could not replace others and proceeded to generate a large training data set by human fashionistas. They would compile personalised looks and create many variations on the look by replacing 1 or 2 items. The ML model was trained on identifying item similarities and select user’s wardrobe items based on a close visual match, according to style, pattern and colours.
Today Stylishly is used by a large community of fashionistas in their daily morning routine and new look discovery. A variety of celebrity look curators bring the latest and greatest to Stylishly users. For some users, this has taken all stress out of the daily “what am I going to wear” conundrum and for others now Stylishly informs them on which fashion item to buy next.