INTRO
Everyone deserves to feel beautiful and comfortable in what they wear.
Lelu [lay-loo] is a style discovery engine that personalizes suggestions that fit you perfectly, based on your measurements, color season, body type, and personal preferences.
No more second-guessing, inconvenient returns, and feeling uncomfortable in your clothes.
TIMELINE
July 2024 (3 weeks)
TEAM
MY ROLE
UX researcher, UI/UX designer
PROBLEM
Online shopping for clothing is frustrating due to inconsistent sizing, indiscernible quality, and a lengthy discovery process.
You're casually browsing online for clothes, maybe something for the new season or an upcoming event. Hours later, you're still adding and unadding things to your cart. Suddenly, uncertainty creeps in. Will this even fit me? Is it good quality? What's the return policy? You close the tabs without buying a single thing.
A little industry secret…
The clothes don’t fit the models either. Usually, they are pinned to perfection on set, creating the illusion of a tailored fit in product photos. Despite knowing this, I struggle to visualize how clothes will fit me when shopping online and hear this echoed by many others. This led me to question how the online shopping experience could be reimagined.
SO I QUESTIONED…
HMW help online shoppers efficiently find clothes that fit their body and style values well?
GOALS
User
Through research, the following goals emerged:
Focus on finding fit through measurements, not sizing labels
Reduce returns (and waste) through increasing consumer confidence and encouraging mindful consumption
Help users find style inspiration easily
Business
Because this was a self-led project, I gave myself the following (hypothetical) KPIs to design for as a constraint:
Customer satisfaction: positive user feedback (simulated via usability tests) → indicative of potential demand and retention
Conversion rate: how easily users can find a desired item and complete checkout → ensures design guides users to key actions
SOLUTION
Receive personalized recommendations for clothing tailored to you in an all-in-one marketplace.
Onboarding: learning what you like
From the get-go, Lelu begins with a quick onboarding quiz to learn who the user is and what they like, making them feel engaged and understood.
Complete Profile: ensure the best fit
Users can enter their measurements, complete an AR color analysis, and an AR body type scan so their recommendations are grounded in more precise data.
Checkout: tailored just for you
Users receive recommendations personalized to the information they’ve shared, and can either save the item to collections, check out, or discover new brands and items.
But wait, how did we get here? 🤔 Let's rewind a bit to the process…
RESEARCH
White-paper research: Sizing is outdated, leading to costly and frustrating returns.
I first conducted white-paper research to validate the necessity of the product and inform my user interview questions.
Vanity sizing
Purposely labeling clothing with smaller or larger measurements (Hoegg et al., 2014). This makes it difficult to know what size you actually are, potentially warping self-esteem and body-image.
24.4%
The online return rate for apparel in 2023. It was the most commonly returned online purchase, mostly due to sizing issues (Coresight 2023).
$38B
The total cost of online returns for apparel. Total online returns also generated 5.8 billion lbs of landfill waste in 2020 alone (Coresight 2023, Optoro 2020).
User Research
I conducted user interviews with 6 fashion-conscious online shoppers. This included male and female users, aged 20-60, with varying incomes and geographic locations in the United States. Between iterations, I also conducted multiple usability tests.
My research produced key insights that drove 4 major design decisions.
INSIGHT-INFORMED DECISIONS
1
"When shopping for clothes, I highly value comfortability/fit and quality."
5 out of 6 participants explicitly stated that “comfortability” is a key determinant for purchase
4 out of 6 participants value high-quality clothing, stating that material is important to them
2
"I’m hesitant when I shop online, mostly due to inconsistent sizing."
Virtually every participant also expressed frustrations with online shopping, due to difficulty visualizing the clothing, uncertain sizing, and a waste-generating return process
3
"I like to use social media for outfit and style inspiration."
4 out of 6 participants take inspiration from social media, namely Pinterest and TikTok
4
"I think body/color analysis can be helpful, but I’m not super familiar with it. In fact, I’m concerned about privacy and accuracy when it comes to biometric data collection."
4 out of 6 participants expressed that color analysis can be a helpful tool, but most participants were not super familiar with the concept
3 out of 6 participants expressed concerns about the use of their sensitive data in AR technologies
ACCESSIBILITY
Accessibility for the benefit of all users was an inherent design consideration, including contrast ratios that meet WCAG standards, multiple navigation options, and clear visual hierarchy for screen readers/AT.
WRAPPING THINGS UP
Reflections
1
Never be married to a solution
After conducting user research, I restructured the entire app and integrated the AR portion as a feature that aids the algorithm, rather than a leading feature. I learned that it’s not about what I think is good, it’s about what the users tell me.
2
Embrace a lack of linearity
Although I followed a design thinking framework, I constantly revisited and repeated certain stages, scraping previous work and iterating. I learned to lean into the process and respond dynamically to ensure every step was purposefully aligned with the end goal.
3
Managing scope creep
My first ideation list was about a mile long. I wanted to give users every feature they could possibly use. However, to be practical and realistic, I focused on feasibility and learned to prioritize only the most essential features based on the goals.
4
Make every element purposeful
I tried to be thoughtful in the placement and function of every element, ensuring that the app was as simple as possible—no more, no less.
Next Steps
Given more time or development opportunity, I would explore the following features:
Search by image: since many consumers get fashion inspiration from Pinterest and other social media, it may be helpful to allow users to search by images like screenshots
Prioritize UGC from consumers: during user interviews, participants mentioned that reviews and photos on real people are extremely helpful
Add multiple profiles: for cases such as people adding their partner, parents adding children, etc.