My Role
Senior Product Designer
What I did
Product design | Stakeholder management | Interaction design | User research & testing | Front-end Implementation
Timeline & Status
2024, 6 months, Shipped
Overview
As a Senior Product Designer, I led the e2e re-design of the content discovery mechanism in Brix. I owned the design process, from idea generation to implementation, working closely with users, developers, product managers and senior stakeholders (end-users and VP's of Product). I also balanced business goals with product goals and planned the release schedule for new features.
Context
Brix is an asset sharing platform for machine learning code assets that reduces time-to-value for AI innovations by enabling the discovery and reuse of code snippets, visualisations, notebooks, and pipelines.
As the number of active users grew to 1,500 and the amount of code assets reached over 500 items, finding information on Brix became increasingly difficult. The filtering tools on different pages didn't work consistently, showing that the system was disjointed and unable to manage the growing amount of content. This contributed towards higher bounce rates, overwhelm and frustration.
Recognising this, we saw an opportunity to increase engagement and conversion by developing a robust and scalable filtering and search experience.
Impact
In the 12 months since shipping new features and implementing a new design system.
53%
Increase in consumption of assets on Brix.
25%
Increase in monthly active users (MAU).
48%
Increase in asset contributions on Brix from 500 to 741 items.
Scoping deliverables
Before diving into detailed design development understanding user intentions for interacting with the Brix was crucial, which helped focus the scope of deliverables. Initial interviews with users and product owners revealed two key archetypes:
Users who are focused on quickly narrowing down results to match their specific interests or needs
Users who are directly seeking specific information or content entering keywords or phrases
This distinction allowed us to tailor our approach, ensuring the platform meets the unique needs of both user types and enhances overall user satisfaction and engagement.
Filters
Optimising filters with key content tags and types.
Mapping out the content and data structure helped cement our understanding on what the filtering parameters and selection inputs should be. While it might be easy to mirror all potential properties of the data, they don’t all share the same usability value in the eyes of the users. For Brix, “subject-matter tags” and “content type” were the fields which tend to be more looked at and ultimately, they differentiate the content.
Streamlining asset discovery with real-time filters.
Unlike typical e-commerce and media discovery platforms that encourage casual browsing and exploration, user interviews and concept testing revealed that Brix focuses on efficiency: users need to quickly find assets to solve specific problems. Based on this insight, implementing a sidebar pop-up filter provided a compelling alternative to top bars and traditional fixed filter panels, allowing users to concentrate on one task at a time and preventing analysis paralysis.
Each filter input sends a request to the server, updating and displaying the number of available results in real-time. This provides users with immediate feedback on how far or close they are towards their comfortable range.
Search
Top navigation and autocomplete for better query formulation
An analysis of design patterns from platforms such as Amazon, LinkedIn, and Pinterest revealed that the top navigation is a common and effective location for search, which provided further rationale behind our design decision. This ensures the search bar is easily accessible on every page, providing a consistent user experience across Brix.
A key consideration in developing the search experience was 'how do we optimally guide users in constructing their search queries'. Users often struggle with query formulation; if they don’t get good results initially, subsequent attempts rarely improve. Therefore, autocomplete suggestions are a crucial aspect of the Brix search, helping users articulate more effective search queries.
Enhancing search efficiency with highlighted keywords
Highlighting text in search results allows users to quickly scan and identify the most relevant information. By visually emphasising keywords or phrases that match the search query, users can easily spot the connections between their query and the content within the results. This reduces the time spent reading through irrelevant text, making it faster and easier to find what they’re looking for
