Case Study


Introducing e-commerce functionality to decrease Choisi’s bounce rate by 10%.



UX Designer

  • Interviews

  • User Flow

  • Wireframes

  • Personas

  • Information Architecture

  • Usability Tests


Web, E-Commerce


Invision Studio, Whimsical, Keynote

Project Overview

Choisi is an online fashion marketplace connecting fashion-conscious consumers to emerging independent designers from all over the world. The site also has an AI feature that helps users find items best suited for their style based on online habits and decisions.

The challenges

The founder came to us with two problems: 1) Decrease the site’s bounce rate from 60% to under 50% and 2) Create a personalized shopping experience on a site that can’t implement it’s AI software due to Squarespace limitations.

The solution

With no concrete timeframe for implementation of this AI feature, we decided it would be best to focus our redesign on what we had in front of us.

In 2.5 weeks, my team and I delivered a full site redesign centered on seamless e-commerce functionality, including several ways the site can act as a personal stylist. Bounce rate results TBD based on an expected Q2 launch date.


Research & Discovery


We didn’t want to jump in blind, so I worked with my team to understand the luxury fashion space and discover online shopping habits, particularly regarding high fashion, to confirm our initial assumption: giving the site the e-commerce it desperately needs.

In conducting a heuristic evaluation, the biggest pain point we found was all products link to affiliate links, leading to significant, and frustrating, memorability and efficiency issues. The high bounce rate was confirmed, as there was no way to search, browse, filter or purchase on the site itself.


current user flow & journey

It turns out that users were equally as frustrated as we were. Usability tests showed us that no matter how they navigated the site, every product image on the site led users to a completely different external site experience.

One user told us, “Why would I even bother with Choisi if I just ended up on Etsy and could buy the product there?” Major red flag.


the luxury fashion space

In looking at competitor sites, we found common themes and patterns that provided us with a framework for how to best design for a high fashion site.

  • Navigation: “Shop Insta” sections that resemble Instagram feeds, fit guides, and ways to filter by things like new arrivals, limited edition items, collections, etc.

  • Site Layout: Clean layouts, bold and big product detail images, lots of white space, intricate typography choices, and section headers guiding users to trends or collections.


who are our users?

With no e-commerce or AI software, the site had no real current users. So, we talked to a variety of shoppers instead. I made sure we spoke to both young and older shoppers so we didn't focus too much on one age segment.

I realized that there’s no need for Choisi to re-invent the e-commerce wheel just because it’s high fashion. Female online shopping expectations are pretty innate across the board no matter their sense of style or budget. They want to shop with ease and in terms of luxury, quality trumps all.



prioritizing features

Users also told us they want filter options to cater the online shopping experience to them and are inspired by what they see on social media. They also mentioned that if they’re going to purchase pricey items, they need to know they‘re getting the most bang for their buck via reviews, size guides, and robust product details.

This was all great to understand their feature expectations. However, I encouraged my team to focus first on basic e-commerce elements before moving on to these additional features.


user Personas

I also didn’t want us to assume every Choisi customer will be whom we assume a typical luxury buyer is. Our interviews proved that to us. So in our personas, I had us cover both ends of the luxury buyer spectrum.

  • Elise is our “moment of wealth” millennial shopper who will once in a while purchase a high-end item as an investment piece.

  • On the other end, we’ve got Celeste, our “inspiration seeker/high spender” who’s all about the latest trends and making a statement.

Design & Testing



We started with holding a design studio with the founder to involve her in the process and show her our design thinking.

During our design process, I emphasized my design motto with my team: Great Design Evokes Trust.

Our users told us that if they’re going to buy luxury, the site should exemplify that via high-end design. Therefore, we focused on a luxurious, editorial feel with large, lifestyle images, white space, and intricate typography choices, all while keeping the company’s current branding in mind.


testing it out

Users all shop differently. Therefore, their feedback in usability and AB tests at all stages of fidelity helped us expand filter options and define a section hierarchy for the homepage.

Key Challenges:

  • Reviews and Ratings: Most users are accustomed to sites like Macy’s that use star ratings. However, they feel that this type of rating system doesn't elude “luxury.” This inspired me to change this section to brief testimonials so users can still feel confident in their purchases.

  • Fit and Shipping: Online shopping puts more emphasis on product fit, and designers on this site are from all over the world. This means their sizes are all different. Therefore, the product pages needed a link to designer-specific size guides and designer-specific shipping information so users can be well-informed up front.

The Final Product


Our full site redesign included a new homepage design with revised navigation and hierarchy, a product listing page with various filter options, a product detail page with large images, robust product details, favorites button and designer info, a full checkout flow, including a shopping bag, payment and shipping, confirmation page, and three personalization features outlined below.


I redesigned the site’s navigation to provide more robust browsing options to accommodate various types of shoppers. The reason for this is so users like Elise can shop for something specific and users like Celeste can easily discover emerging designers and trends.

In revising the user flow, I focused on highlighting the fact that users can now complete their purchase with ease with no redirects or bouncing.


Personalization Features

Finish the Look: On product detail pages, users can get suggestions on what goes with the item they’re looking at so they can build an entire look vs. just buying an individual piece.

Shop This Look: From classic to sporty to evening to vintage, users can discover and purchase completed looks that accommodate different styles and trends.

Editor’s Picks: I re-labeled and redesigned the site’s current “Top 10 Lists” to “Editor’s Picks” to elude a sense of trust in users like Celeste that they are looking at items hand-picked by people who are leaders in the fashion industry.

The Results

Users were pleased with a clean, beautiful site that allowed them to shop with ease and guided them to more personalized, more effective purchasing decisions. But here’s the kicker:

Before, users could spend as little as 7 seconds on the site before bouncing to an affiliate link. In final usability tests, users spent between 3-5 minutes browsing the site and completing their purchase.

I’m confident this redesign will significantly lower the bounce rate past the 10% we aimed for.

Next Steps

  1. Continue to test with more users

  2. Build out the Shop Insta page

  3. User-generated “editor’s picks” to create shareable content


Key takeaway? Not all feedback will be useful feedback.

Throughout the process, I actually encouraged my team to take some user feedback with a grain of salt, especially since we weren’t equipped to find as many true luxury buyers as we wanted. I realized we had to evaluate all feedback based on where it was coming from. If users were expecting a Nordstrom site, we had to decide if making that change was worth it based on our primary goals.

If I could do this project again, I would spend more time in the research and testing phases with more users aligned with Choisi’s demographic. This could help us better validate or tweak our design choices to ensure we’re best accommodating the users most likely to shop on this site.