a style recommendation system for online and in-store shopping
Clothes shopping is poised for disruption. Discovering new clothing online is difficult with the huge number of available retailers, and real world discovery is hampered by boutiques having typically poor exposure. In response to the growing chaos in this industry, Tailored Fit is an algorithmic style recommendation system that bridges the experience between online and in-store shopping and intelligently mitigates the overwhelming number of options for the customer.
Tailored Fit was born from the idea to create a "Pandora for clothes shopping". We received overwhelmingly positive reception for the idea of a tool that aggregates stores/brands and individually tailored filters that increase the discovery of new clothing that the user actually likes.
During concept validation, many young women shared their thoughts on the frustration of having to go to a ton of different stores and websites.
Tailored Fit's online and mobile applications uses over 1,200 clothing attributes to understand consumer preferences.
As users like, dislike, and purchase clothing, we create a detailed profile of their style and make suggestions based on that profile from a wide collection of online retailers.
Since launching our website, we've maintained a 15% landing page conversion rate, 10% conversion from organic search, and 60% retention rate from first interaction.
Our website experience reflects our goals for the users, allowing them to save "racks" of clothes they like for different occasions.
The mobile product was built to be intuitive for the user. I was the primary designer for the mobile interface. I used competitive analysis, extensive mental model mapping, and guerrilla user testing methods to determine an interface that could allow the user to be engaged with our product in primary shopping contexts.
The dashboard contains personalized recommendations for the user. It's the landing page that users will see after they log in. Tapping on the product will show users an expanded view of the product as well as options to see more info, buy, share, and like.
Active discovery and curation
When a user wants to start a discovery session, they can preemptively adjust filters. The layout of the products allow the user to see more of the product in depth while still allowing them to scan/like quickly. They can also view their list of racks of products that they have saved before.
Delivering a seamless experience across all retail touch points remains both a key challenge and prime opportunity for retailers today.
Users can carry their Tailored Fit account with them as they shop. Our "local" feature, which was still in progress as we launched the app initially, extended the Tailored Fit shopping experience to physical contexts.
Retailers like the ability to bring in new shoppers to their physical location,
which is something we can do through recommendations based on shoppers' purchases at other stores.
We started testing this experience at e.b.Pepper, a local boutique store in Pittsburgh. This experience is enabled by iBeacons containing inventory data and can communicate to portable electronic devices via Bluetooth low energy transmission.
If a user walked by a store that contained clothing that fit into their preferences, then they would recieve a notification that directed them to review the item in-app as well as location of the store.
The Clothing Genome Project is a key component to the algorithmic recommendations Tailored Fit provides.
Analogous to the Human Genome Project, the "DNA" of an article of clothing comprises binary values for clothing attributes.
We have a set of over 800 tags (or “genes”) for articles of clothing that we base our recommendation system on. The recommender system is able to calculate the similarity and difference between any two products. As users like and dislike different items, we can see which clusters of ratings they fit into best and recommend new items that share those properties. We currently have 65,000 pieces of clothing on our site with 54,535,000 tags applied to them through an automated tagging process.
This system, from tagging to recommending, has the potential to be patented once it’s been sufficiently developed. It also has the potential to be used in other industries such as health, food, consumer goods, hiring, and more.
As of now, no other shopping startup is focusing as heavily on algorithmic recommendations, and nor publicly working on location-based shopping recommendations based on a user’s style profile. If we can make significant moves into this space, we'd be the first to integrate with retailers for real-world localized shopping recommendations, a relationship that makes switching services difficult.
When forming our business strategy, we positioned Tailored Fit in ways that would engage the market - and grow beyond it.
We also created a strong focus on user-centered design and performative feedback gathering, while our main competitors in this field do a poor job engaging directly with users to build the product to their needs. We built a referral and social sharing system built into the product, based on the ability to share your curated “racks” through social media, especially Pinterest. Every Tailored Fit user is contacted for feedback, with 40% of them responding.
When we launched our public beta to early test users, we grew 10-15% week over week by word of mouth alone.
From our market research, we decided that the visual identity should emulate a sense of being confident and comfortable in one's clothing. The aesthetics were carefully composed to resonate with our target market.
When thinking about the impression of the logo, we wanted it to represent the individualized aspect of style. Previous iterations that attempted to be more minimal were considered to be too "tech-oriented" or impersonal. We decided on a handwritten cursive logo that characterized Tailored Fit with a strong sense of personality.
I used a Canon EOS 5D Mark II and a Canon Rebel XT with 18-55mm lens. This is a sample of photos used in marketing materials and within our apps.
Achievements & What I found really cool
The idea started as the "Pandora for clothes shopping" at Startup Weekend Pittsburgh on October 18th, 2013. We won first place after a 48-hour build-a-thon. We were incredibly fortunate as a group of 8 individuals, all previously strangers to each other, to be able to work together so strongly towards this project. At the event, even after we were announced winners, we continued working for hours in order to see the last pieces through.
It was the first time I felt like I was truly utilizing my education in business administration and passion for design. I could not have been able to communicate as well or iterate my designs as quickly as I did had I not been able to actively flesh out the business structure while formulating the user experience and marketing.
This was start-up life. For months, after I finished classes for the day and Sol, the other designer, finished work, we worked for hours together creating design assets and making wireframes. The demanding schedule we lived in took us to winning pitch competitions, gaining connections to investors, and becoming incorporated. I relished wearing many hats, from photographer to content strategist to design specs slave (no job is too low).
For a year, we were an incorporated seed company funded by AlphaLab (a startup accelerator in Pittsburgh), and released our app in the App Store in 2014. I made the decision to leave the company two months before the company dispersed, as I was starting my masters degree in Human-Computer Interaction. I learned incalculable lessons through the chaos of creating a startup, both professionally and personally. The experience made me both intimately appreciate and critically evaluate business structures and processes.