A product card design that helps to pick a fragrance online
The project aimed to simplify online fragrance selection, particularly for those new to Scentbird or to fragrances in general. The new design brought important information to the forefront, making it easier for users to select fragrances that matched their preferences. Here, I've recreated them as React components for you to play with:
By studying user behavior and conducting 17 interviews with power users, I gathered insights and identified patterns that informed the new design. It resulted in a 50 percent increase in the number of products added to user queues, effectively improving the user experience and our retention goals.
Organizing mobile navigation to improve UX
This project concentrated on redesigning our mobile navigation to enhance the user experience, considering 90% of our new users were accessing our website via mobile devices.
After identifying issues with our existing setup, we decided to rethink our information architecture. I gathered insights through card sorting, onsite user interviews, and tree sorting, then collaborated with a team of UI designers to explore five different prototypes. We implemented a six-month A/B testing phase on various audiences. The results showed a significant improvement, with users accomplishing their tasks 20% faster and more reliably. Additionally, the redesign led to a decrease in customer support tickets.
Modular information architecture: solving the onboarding and first-month retention problem
The new modular system helped users focus on fragrance during the first month and helped them explore other product categories later. Each module solved it’s own specific user or business problem and depended on a user's history with the service. The system also introduced elevated cards as a new design language.
Users quite often misspelled their addresses during registration. I decided to show them the delivery address so that they could change it right away. The test showed at least 1,600 users fixed their addresses after registration, saving us from shipment issues and time-consuming customer support.
New users struggled to pick a fragrance, so I illustrated for them how notes are important for a fragrance, and they started exploring based on what is popular. This section and the resulting pages quickly became one of the top-five sources for users adding fragrances to their queues.
With an expert curating each collection, it is easier to think of them as playlists or mixtapes. Each collection fits the central theme, serving as a convincing reason to try the suggested fragrances. This section quickly skyrocketed to be a top-3 source for customers adding products to their queues.
I am especially proud of this one. The UI was cluttered with upgrade promos, but they didn’t perform well. We removed 15+ banners from the UI and replaced all of them with this section, and it still performed better.
Cross-promotion was all over the place because of experiments, and I removed all other mentions and collected them in one place. All promos were combined on one page, and power users used them to get the best deals.
Each module was developed simultaneously for desktop and mobile. Some modules made significant improvements on their own, e.g., users added more products to the queue from sections like Collections, Notes, etc.
Scientific hiring
How hiring managers review portfolios, resumes, and choose whom to invite for a job interview. Research results.
I gathered the applications of 243 real designers actively looking for a job and enlisted the help of 16 hiring managers from product companies who agreed to rate the applications as suitable for an interview. The results showed that only 3% of applications received unanimous approval from all hiring managers, while just 8% were most likely to receive an invite for an interview. The study also found that it takes less than one minute for a hiring manager to decide about an application and portfolio. The most accurate hiring manager had an accuracy rate of 84%, while the least accurate had a rate of 62%.