Introduction
Launched in Q4 2022, Sourcerie was born from the collaboration of founders Alex and Kristen, whom I connected with through an investment partner at LYST, a fashion meta-search app I helped launch. My role as Principal Product Designer allowed me to leverage my expertise from working on LYST, where I significantly increased mobile app revenue contributions.
< 1 % to > 12 %
% revenue contribution 2020/21 — LYST mobile app
Having previously worked on this marketplace product with applied machine learning. This was an opportunity to consult with them and develop their product for launch — as their Principal Product Designer.
The System
Sourcerie, backed by investors from Farfetch, Vestiaire Collective, Trouva, and Amazon, is a machine learning-driven marketplace for personal care. The platform focuses on under-represented verticals such as condition groups, ethnicities, and life-stages. It generates personalised care routines by analysing unstructured customer reviews and structured product data.
To ensure seamless integration, we developed a comprehensive library of usability-tested UI components that align with both the product patterns and the engineering codebase. Early on, we prioritised accessibility standards, collaborating closely with the CTO, Head of Product, and Engineering Lead to utilize existing front-end libraries and UI frameworks for initial implementation.
We created a robust design system for the store and web app using Tailwind.css, emphasising componentisation and establishing an accessible colour library to enhance user experience.
UX Design
Mapping the system architecture and database schema was crucial in aligning design system naming conventions with engineering. We created detailed site maps, journey maps, and notification mappings, along with defining email templates for various scenarios in payments, checkout, marketing communications, and profile completion.
We built a backlog of post-launch A/B and multivariate tests to optimise content hierarchy on ‘Learn/Buy’ pages against our paid ad spend. A unique challenge we tackled was developing a visual language for a customer’s personalised care profile.
Research
Leveraging well-established e-commerce patterns such as search, browse, product cards, product detail pages (PDPs), add to bag/cart/basket, checkout, payments, and address entry was a starting point. However, disruptive products like Sourcerie require innovative patterns to break conventions and enable new behaviours.
In our case, the challenge was presenting machine learning-driven recommendations effectively. We conducted multiple user tests, both unmoderated via Maze and moderated with a panel of early adopters, to refine the UX/UI variants and naming conventions for ‘Collections’.
Branding
Our branding efforts encompassed creative direction, social content templates, and the refactoring of brand guidelines into Figma. We meticulously defined colour and logo usage, content guidelines, and developed illustration and animation assets to ensure a cohesive and engaging brand identity.
Conclusion
My role in developing Sourcerie involved a multifaceted approach combining advanced machine learning with intuitive design, robust UX, and strong branding. The collaboration and rigorous testing processes ensured a highly personalised and accessible user experience.
If you’re looking for a product designer who can drive innovation and deliver impactful results, I’m ready to bring my expertise to your next project.