The Client

Curtsy is a Y Combinator graduate that pivoted from peer-to-peer renting dresses to buying and selling cute clothes.

say banner.png

The Challenge

Curtsy wasn’t getting enough users to upload posts for selling their clothes. This is their main stream of revenue as they receive 30% and a required transaction fee for each sold post. Their current process was lengthy, formulaic, and felt like work. They wanted to make the uploading experience smoother and incentivize users to create more postings.


The Goal

With not enough postings to browse and keep users loyal to habitually using the app, many were abandoning the app altogether or switching over to the competitors. According to their Mixpanel, only 9% of returning users and 11% of new users completed an upload flow, meaning they posted an item for sale. Our metric of success was to increase users completing upload flow by at least 1%, while avoiding any use of dark patterns. We also wanted to encourage users to post at least three photos (increasing the average photos per post from 2.3 to 2.7) because studies show items sell faster.

curtsy posting 2 transformation.png

Design Process

We worked directly with William, Curtsy’s CEO, to create a strategy document as well as define the scope and timeline for this project. I was the primary stakeholder for the final Hifi design deliverables and prototyping. I also created new components for their design system.


User Research

People of all different demographic and socioeconomic backgrounds used Curtsy, though its main demographic of users were women. For this reason, it was important to test usability with all users. We found new users for each round of guerrilla testing.



After prioritizing existing pain points and the synthesis of discoveries into a product audit, we began exploring solutions. This stage of the project was lean, with three rounds of experimentation, testing, and adaptation. 

Once the tests validated navigability, usability, and comprehension, we organized final design deliverables and presented a research deck to explain our decisions and discoveries. The final prototype can be viewed here.

process banner2.png


These are screens from the final prototype that solved major pain points for users (82.2% to be exact).


Users were surprised and frustrated at additional work that popped up as fields were filled. The redesigned, transparent design shows user upfront what work needs to be done. Users are encouraged to reach ideal state (complete) instead of leaving it in empty or partial state. In addition, taps and time (by at least half a minute) to complete post were reduced.


Users complained of too much back-and-forth screen switching. Design patterns, such as a pop out scroller instead of a picker, was added to minimize this.


Another example of a pop out scroller added for ease of use. In addition, emotionally sensitive copy was added as asking for size felt a bit intrusive for female users.


Users were confused on how shipping affected earnings. We clarified this through more transparent design.


Mission Accomplished

We continuously tested and prototyped until we solved over 80% of pain points, including those that arose because of new design patterns introduced for each prototype.



I've gotten better at not only communicating with design to my users, but also with my teammates. I learned organization and documentation go hand-in-hand explaining design decisions. Substantial information is best communicated first face-to-face, then in writing, which should be kept as concise as possible. Attention is always precious. Working with an established design system means design hygiene is essential: paying attention to nuances like naming symbols and color code consistency saves a lot of pain and tears.