An intelligent assistant to help young women pick outfits

Passion Project   UX Research   UX Design    UI Design


Young women at their 20s are always haunted by a question in the morning: "what to wear?” They have so much to consider: Is it going to rain today? Is this sleeveless dress appropriate for my client meeting? How can I look more professional? Picking the right outfit quickly and confidently can be a real problem for some women.

How can we make the outfit picking process less time-consuming and more enjoyable?

Date May - Jun 2017, 8 weeks

Team Joy Chen, Minrui Li

My Contribution I participated the whole design process in collaboration with Minrui and focused on the app design.

Project Passion project


Design Outcome

Mia: An intelligent assistant for picking outfits

To solve the challenge, we designed Mia, a service composed of a mobile app and a smart mirror to provide dressing suggestions and let users gain insights towards their clothes and dressing behaviour.

Note: Clothes pictures used in the prototype credit to Madewell and A&F.

Get daily dressing suggestion
Make morning routine a lot easier

Every morning, Mia suggests four outfits based on the weather and the user's schedule. Users can just simply pick one of them.

Know how the outfit looks on your body 
Save time of taking them on and off

With a 3D model embedded in the mobile app, users can "try on" multiple outfits in a second.

Know what you have and know what you love to dress, precisely
Gain dressing insights and cultivate dressing sense

By tracking the user's dressing, Mia lets users see their dressing patterns and gives personalized advice. Users can therefore explore new styles they never tried before with great confidence.

Try new arrivals and know how to match with your own clothes
Shop at home and spend money wisely

By cooperating with fashion brands, Mia extends clothes shopping from retail store to your cozy home. Users can virtually try new arrivals with the 3D model or the smart mirror and purchase them with a click.

Design Process

/ Problem space and defining goals

Women often find it hard to decide what to wear

Average lady spends at least 12 minutes of each workday morning choosing what to wear, and a further 10 minutes at weekends. Picking the right outfit can be a pressure. How can we make this picking thing less painful? At the outset of this project, we defined three high-level goals.

Our high-level goals

1. Remove uncertainty and confusion for users on outfit picking.

2. Make picking daily outfit easy and enjoyable.

3. Help users only spend money on what they really need.

/ User interview

In order to learn about women's outfit picking process, we talked to 13 ladies. The pain point got confirmed by most of them. Upon more probing, we gained the following insights specifically.

/ Competitive analysis

We looked into multiple fashion and clothes management apps, and we found that there are opportunities for a highly personalized and private service.

/ Product decisions

Based on what we learnt so far, we were able to make the following decisions for the product.

1. Target at young female professionals
2. A combination of mobile app and mirror
3. Focus on personalized suggestions
4. A daily used tool but also provide suggestions for special occasions

/ Speed dating with storyboards

Prioritize features based on user feedback

Based on the product decisions, we brainstormed several ideas and quickly made storyboards for testing. We tested four main functions: daily suggestion, clothes try-on, explore new styles for special occasions, and dressing reports.

What we learnt

1. Users like mirror for its intuitiveness.
2. Daily suggestion is needed, but the interaction needs to be as simple as possible; when users think about what to wear in the morning, they mostly think about weather and occasion first.
3. Explore function is not wanted as much as others since it's a low-frequency need.
4. Visualization of dressing behaviour only is not enough, actionable tips are more helpful.
5. Shop function has big potential if users can virtually try new arrivals at home.

/ Design decisions

An assistant, not a guide
Let users make decision themselves

Besides, we also figured that there's a big difference between dressing suggestion and general recommendation for like hotels or restaurants. People are more attached to their clothes compared to other general products or service, because clothes reflect their self-identity to some extent. They want to make decision themselves and tend to not fully trust what the app recommends to them.

We then decided to position our product as an assistant, rather than a guide. We hope the product help users better know themselves on dressing, and then enable them make decision with full information.

Finally, we identified five main functions of the app: daily suggestion, closet, report, shop, and look book. And based on the context of use, we decided to let users access three of them on mirror: daily suggestion, closet, and shop.

Flow of value

/ Information architecture & flow map

/ UI design

Design for simplicity and elegance

Since we're targeting at young women professionals, we want the app to be clean, simple and elegant. We decided to use neutral colors as primary colors and use one red color for attracting attention. We chose to use simple line icons and Avenir font to create the feeling of modern and clean.

Daily suggestion
Shop - categories expand
Look Book

/ Design iterations

Reduce cognitive load

We tested the hi-fi prototype with users and found that the cognitive load in using the app was too heavy. 

Let users be informed; assist them make decision

We kept in mind that users want to make decision themselves, so we used the following design tactics to make users feel that they are in control.

/ Mirror design

Mirror provide an intuitive and efficient way to virtually try on clothes and record user behavior. This is the closet mode and shopping mode, users finally check out on their phones.

Onboarding experience: simple popover view.

/ Reflections

This is an idea that I've been wishing to develop for a long time. We went through some zig zag turns during the process, but we also gained a lot of fun. Two lessons I learnt from it are:

Test design ideas as early as possible and keep learning from users. When we're not quite sure about a design direction, asking the users can help us get the best answer for most of the time.

Know the design process, and also know it's iterative. Designing for the real world is seldom a straight line. I need to fail fast, iterate fast, and reduce my mental attachment to a solution or prototype.

/ More projects