MY FIT STUDIO

AI-powered virtual try-on feature for the ASOS mobile app to improve fit accuracy and enhance user satisfaction

PROBLEM STATEMENT

The current fashion recommendation systems often lack personalization, failing to fully utilise user data such as purchase history, browsing behaviour, and style preferences. This results in a generic user experience that doesn't resonate on a personal level. To create a more engaging and tailored shopping experience, the new feature will provide personalised fashion recommendations, thereby increasing user satisfaction and driving sales.

PROJECT BACKGROUND

ASOS plc is a British online fast-fashion and cosmetic retailer, primarily aimed at young adults. They sell over 850 brands as well as its own range of clothing and accessories, and ships globally from fulfilment centres in the United Kingdom, the United States, and Europe.

HYPOTHESIS

Integrating an AI-powered Personal Stylist feature into the ASOS mobile app will provide personalized fashion recommendations based on user preferences, style, and browsing history. This will create a more engaging and tailored shopping experience, leading to increased user satisfaction and driving higher sales.

USER RESEARCH

RESEARCH METHOD

To guide the design process, I conducted a competitive analysis of AI-powered fashion apps and did secondary research, including digital listening, to understand user pain points and industry trends. I also carried out in-depth user interviews, which helped me create user personas and a user journey map.

COMPETETIVE ANALYSIS

IN-DEPTH USER INTERVIEWS

KEY QUOTES BY PARTICIPANTS

OBSERVATIONS FROM AFFINITY MAPPING

  1. Users need personalised styling & pairing recommendation based on the product they want to buy. This recommendation should be based on their own style preferences

  2. Users want to get a good sense of how a product looks on them before buying it

  3. Users frequently use features like search, filter & sort to narrow down their options based on brand preference and budget

  4. Users go through reviews and want to know both positive and negative aspects before purchasing a product.

  5. Users spend most time in determining the right size and fit for a product (they narrowed down to) and factor in body shape, height, material & color before selecting

DEFINE

USER PERSONA

USER JOURNEY MAP

DISCOVERING THE REAL PROBLEM

PIVOT TO AI VIRTUAL TRY-ON

During the research phase, it became clear that the primary pain point was not the lack of personalized fashion recommendations but rather issues related to sizing. Users frequently encountered problems with finding the right size, leading to a high number of returns, which negatively impacted both user satisfaction and company costs.

HOW MIGHT WE…

IDEATION

FEATURE ROADMAP

  1. The design aims to solve sizing discrepancies in online shopping by integrating several key features.

  2. Users can create realistic AI lookalike avatars and use the Virtual Try-On feature to see how clothes fit and look before purchasing.

  3. A comprehensive Sizing Guide helps users choose the right size, while profile management features allow them to update their information and add multiple profiles for others.

  4. Additionally, the app could offer personalised recommendations based on body profiling, ensuring a more accurate and satisfying shopping experience.

INFORMATION ARCHITECTURE

SITEMAP

This feature provides two key access points for users: the first is on the homepage and account page, allowing users to create their virtual model; the second is on the product detail page, where users can generate their model wearing the selected garment for a more personalized fitting experience.

USER FLOW

TASK: CREATE A TRY-ON AVATAR

LO-FI WIREFRAMES

SCAN BODY MEASUREMENTS

PRODUCT DETAIL PAGE

USER INTERFACE DESIGN

HI-FI WIREFRAMES

ENTRY POINT

To introduce users to the new feature, access to creating an AI lookalike model is prominently placed on the homepage. Additionally, users can easily access this feature from their account page for convenience.

CREATE A TRY-ON AVATAR

The "Create a Try-On Model" feature lets users generate a realistic AI avatar using their body measurements. This virtual model allows users to try on clothes digitally, helping them see how different styles and sizes will fit.

PRODUCT DETAIL PAGE INTEGRATION

To introduce users to the new feature, access to creating an AI lookalike model is prominently placed on the homepage. Additionally, users can easily access this feature from their account page for convenience.

HOW TO GET STARTED

To make this feature beginner-friendly, a comprehensive guide is included to assist users step-by-step before they begin using the tool. This guide ensures that users can easily navigate and fully utilize the AI Virtual Try-On feature, enhancing their overall experience.

USER TEST

RESEARCH GOAL

To conduct a usability test to understand end users’ preferences, with the aim to evaluate how users navigate the website, find information and accomplish tasks.

RESEARCH METHOD

  1. Conducted tests remotely using zoom and took notes during test

  2. Used affinity mapping method to identify common themes and recommendations Sorted the feedback using feedback grid

  3. Sorted recommendation through frequency to severity mapping to prioritise results

TASKS

  1. Find the virtual try-on feature

  2. Create your virtual model

  3. Try on clothes

FEEDBACK GRID

FINAL ITERATIONS

CREATE A TRY-ON MODEL

PRODUCT DETAIL PAGE

KEY TAKEAWAYS

The AI Virtual Try-On feature in the ASOS app can effectively address sizing discrepancies, reducing return rates and enhancing user satisfaction by allowing customers to create realistic avatars for virtual fitting.

NEXT STEPS

USER RESEARCH

Future research will focus on improving the AI Virtual Try-On feature's accuracy and user experience. By continuously gathering and analysing user feedback we can identify areas for improvement and new feature requests

FUTURE USE CASE

  1. Social media integration: Enable users to share their virtual try-ons with friends and family for feedback

  2. Add multiple user profile: Users can create and save multiple AI models as separate profiles, allowing them to shop for their family members

  3. Enhanced AI algorithms: Continuously refine AI algorithms for improved accuracy and personalisation.

  4. Accessory and footwear try-ons: Extend the virtual try-on feature to include accessories and footwear.

  5. Personalised recommendation: Provide personalised product recommendations tailored to users' body profiles for a better fit and style