Team
Growth and Retention team at SoundCloud
Team
Growth and Retention team @ SoundCloud
Role
Junior Product Designer
Services
Product Design, UX/UI Design, User Research
SoundCloud is a global music streaming platform that empowers creators to share their work while offering listeners a diverse library of audio content.
👩🏻💻 My role
🎯 Challenges & Objectives
During my time at SoundCloud, I worked on several key initiatives aimed at improving user engagement, retention, and subscription conversion. My role involved UI/UX design, prototyping, user research, and collaborating with cross-functional teams to enhance the user experience for both free and paid users.
SoundCloud was facing challenges related to user retention, music discovery, and conversion to paid subscriptions. Key goals included:
-
Enhancing the onboarding experience to improve long-term engagement.
-
Providing better discovery mechanisms to encourage following artists.
-
Encouraging users to convert to paid subscriptions by reducing friction and increasing perceived value.
Find People to Follow – Enhancing Engagement Through Personalized Discovery
Project 1
Context
👥 Overview
⚠️ Problem
New users often experienced an empty and uninspiring feed after signing up, leading to high churn rates. Without guidance, they struggled to discover relevant creators, making it difficult for them to engage with the platform.
Users who followed fewer than three accounts had significantly lower engagement and retention. An empty feed after onboarding resulted in a poor first experience, reducing their likelihood of staying on the platform.
Benchmark examples
👀 Key Insights
Spotify, Apple Music, and Deezer each use distinct onboarding strategies to engage new users. Spotify requires selecting at least three artists to refine its recommendation algorithm, immediately personalizing the home screen with playlists like "Made for You." Free users experience limited access to exclusive content, while Premium subscribers receive additional personalized recommendations and updates. Spotify also uses positive reinforcement, animations, and a delayed marketing email strategy to drive engagement.
Apple Music prioritizes an immersive, visually engaging onboarding experience, offering a three-month free trial but no free tier. Users select artists and genres through an interactive interface, and their choices shape home screen recommendations. Deezer follows a similar approach, requiring users to choose at least two artists to follow before proceeding. This initial selection helps tailor recommendations and playlists, ensuring immediate personalization.
While all three platforms rely on data-driven personalization, their approaches vary: Spotify emphasizes algorithmic learning from the start, Apple Music focuses on a premium-first experience, and Deezer balances between curated and algorithmic recommendations to streamline user engagement.
Design Process
🤔 Hypothesis
🧪 Exploration
🗣️ Research & Insights
If we suggested relevant and popular accounts to new users, they would engage more with SoundCloud, leading to better retention.
When we started designing Find People to Follow, we anchored ourselves in a few key principles: Start with the problem. Serve your users first. Stay agile. Iterate, then iterate again. We wanted the experience to be intuitive, understandable, and seamless, ensuring that users could easily find and follow artists relevant to their music tastes.
A challenge early on was balancing competitive benchmarking with SoundCloud’s unique identity. We didn’t want to replicate what other platforms were doing; instead, we aimed to lean into what makes SoundCloud different - its diverse and independent artist base.
We iterated five times before finalizing the design, making subtle but impactful refinements each round. Every iteration brought us closer to a simpler, more intuitive experience, removing friction while staying true to our guiding principles of consistency and usability.
We conducted remote user testing using Usertesting.com, putting our designs in front of real users to gather insights on behavior, preferences, and pain points. One of the most surprising insights was how much users cared about accuracy in related artist recommendations. If they saw suggestions that didn’t make sense - like Billie Eilish being recommended alongside Lil Tecca - they immediately disengaged. This reinforced the need for stronger algorithms and better contextual grouping in our design.
On the stakeholder side, there was a clear priority: Get the MVP out fast. This meant that while we had a vision for a more sophisticated recommendation engine, we had to first implement the simplest, most dev-friendly version that would still improve the experience meaningfully.
Outcome
💡 Solution
🏁 Impact
🧠 Key Learnings & Takeaways
After synthesizing our research, we made several major design changes to align with both user needs and technical constraints:
-
Adding a clear "Done" button - users needed explicit confirmation when completing the flow, ensuring a smooth, intentional exit.
-
We introduced a dedicated entry point on the Home page to ensure users could engage with the Find People to Follow flow beyond onboarding. While initially embedded only in onboarding, testing revealed that users wanted a way to return and discover new artists anytime. To support ongoing discovery, we also added a Home module specifically for users with fewer than three follows, encouraging continued engagement with artist recommendations.
-
Implemented a search bar and genre filters - giving users multiple ways to interact and personalize their recommendations.
Beyond UI adjustments, I worked closely with engineers to refine how we identified and surfaced related artists. We explored signal collection methods to improve recommendation accuracy and simplified how suggestions were generated, balancing feasibility with user expectations.
Our biggest quantifiable success was a 21% increase in engagement on Day 0 (likes and follows), confirming that the new flow encouraged users to interact with artists earlier in their journey.
Beyond onboarding, these changes laid the groundwork for improving playlist curation and organization on the Home page, an ongoing initiative aimed at increasing long-term retention. While we didn’t have additional usability metrics post-launch, the impact of the new Find People to Follow experience was clear in early engagement trends.
This project was a defining experience in my growth as a designer, particularly in three key areas:
-
Working with large, cross-functional teams. I learned how to align design, engineering, and product goals while advocating for the best user experience.
-
User testing & agile iteration. I saw firsthand how small changes - like adding a "Done" CTA - could make a significant difference in usability.
-
Confidence in decision-making. Since this project, I’ve become much more confident in my design choices and ability to push for impactful solutions.
Looking back, if I could do it all over again, I would think bigger. While the MVP approach was necessary, I now understand the importance of pushing for long-term vision alongside quick wins.
This experience shaped how I now approach product design as a Senior Product Designer - balancing speed with strategy, advocating for users while working within constraints, and always iterating towards better, bolder, and smarter solutions.
Import Playlist – Reducing Barriers to Switching to SoundCloud
Project 2
Context
👥 Overview
⚠️ Problem
For users switching from other streaming platforms, rebuilding their music library from scratch was a major barrier to fully adopting SoundCloud. Many had spent years curating playlists and were hesitant to leave that behind.
New SoundCloud users often cited playlist migration as a major friction point in switching services. Without an easy way to bring their existing collections, users felt less invested in the platform, making retention and engagement more difficult.
Design Process
🤔 Hypothesis
🧪 Exploration
🗣️ Research
👀 Key Insights
If we provided an easy way to import playlists from other streaming services, users would engage with SoundCloud faster and retain better, as they wouldn’t need to manually recreate their music library.
From the start, we leaned into User-Centered Design, Simplicity, and solving the real problem first. Our goal was not only to minimize effort for new users but also to showcase SoundCloud’s unique value proposition—discovery and curation beyond just playlist migration.
However, we faced two major design challenges:
-
Balancing MVP vs. future vision – We needed a fast solution to remove switching friction but also wanted to provide long-term differentiation beyond a simple playlist transfer.
-
Ensuring ease of use – Any additional complexity would deter users rather than encourage them to migrate.
We explored multiple designs, each optimizing for different levels of effort and user needs, though I don’t recall exactly how many iterations we went through before finalizing the MVP.
To deeply understand user needs and pain points, we relied on three key sources of research:
-
Churn analysis – We reviewed existing user data on why people left SoundCloud.
-
Surveys – We sent out additional surveys to understand what mattered most to users when switching services.
-
Remote interviews – The Senior User Researcher and I conducted in-depth interviews to explore how users find, listen to, and organize their music.
Users wanted to import the playlists they spent time creating from other platforms so they could bring their valuable collection with them when joining SoundCloud. Without this feature, switching services felt like starting over.
Outcome
💡 Solution
🏁 Impact
🧠 Learnings & Takeaways
We explored three MVP designs, each with different approaches to onboarding users into playlist importing:
1. Home Entry Point (Multiple Playlists) – Users could import multiple playlists directly from their home screen.
2. Library Entry Point (One Playlist) – Users imported playlists one at a time from their Library, prioritizing a focused experience.
3. Upfront Data Fetching – We experimented with fetching playlist metadata before user confirmation to minimize steps.
Beyond the MVP, I also created a future vision design that would go beyond simple playlist transfers. This concept aimed to:
-
Recommend related tracks based on what was imported.
-
Fill in gaps where SoundCloud didn’t have the exact same tracks.
-
Enhance discovery by surfacing unique SoundCloud content users may not have found otherwise.
However, technical feasibility, engineering effort, and timeline constraints influenced the decision to move forward with an MVP using a third-party service (TuneMyMusic) instead of building a fully native solution.
Throughout the process, I worked closely with engineers to understand what would be required for an in-house playlist importer, ensuring feasibility while advocating for an optimal user experience.
Unfortunately, I don’t believe this feature was ever fully launched, as it was expected that users would use the third-party integration instead. I left SoundCloud before seeing the final results, so I’m not sure whether it had a measurable impact on paid subscriptions or retention.
That said, this project set the groundwork for how SoundCloud approached reducing switching friction, and the vision of personalized music discovery still stands as a key opportunity.
What I Learned as a Designer:
-
“Sometimes you need to kill your darlings.” While I strongly believed in the future vision, I had to prioritize what was feasible and align with stakeholder goals.
-
Balancing MVP vs. long-term strategy is a critical skill in product design.
If I Had More Time…
I still believe that using a third-party service was the right call for development efficiency, but I would have pushed harder for our vision—a solution that truly showcased SoundCloud’s value by integrating discovery and personalization into the import experience.
This project reinforced my ability to navigate technical constraints, advocate for users, and think strategically about long-term product impact—skills I now bring to my work as a Senior Product Designer.
Geo-Recommendations – Personalizing Music Discovery Through Location Data
Project 3
Context
👥 Overview
⚠️ Problem
I facilitated a workshop on leveraging third-party data to improve music discovery on SoundCloud. This workshop brought together stakeholders from product, engineering, and design to brainstorm ways to enhance content recommendations. One of the key outcomes of this workshop was the idea of using location data and event feeds to surface relevant music recommendations, which became the focus of a Hackathon Week experiment.
As a user of SoundCloud, content discovery is not always easy or relevant. Listeners often struggled to find new music that resonated with their interests, and there was no way to explore what was trending locally. This lack of personalization resulted in a disconnected discovery experience.
Design Process
🤔 Hypothesis
🧪 Exploration
⚙️ Key Design Considerations
🗣️ Research & Insights
If we suggested location-specific recommendations, users would keep coming back to check what their city was listening to, increasing engagement and retention.
During a Hackathon Week, we ran a design sprint to quickly iterate on concepts based on the workshop insights. This experience reinforced my ability to:
-
Facilitate structured ideation with cross-functional teams.
-
Work quickly and pivot between concepts based on technical feasibility.
-
Translate workshop findings into tangible design solutions.
1. How would we collect location data?
-
Location services (high accuracy, low opt-in rate).
-
Manual city selection (medium accuracy, more user control).
-
P-based or App Store location (low accuracy, but easy implementation)
2. Where would users engage with location-based recommendations?
-
Home Page (Medium friction, medium conversion).
-
Onboarding (High friction, but could increase activation).
-
Library/Stations (High intent, strong potential for engagement).
3. What would the user experience look like?
-
A SoundCloud Radar page surfacing trending artists and genres in the user’s city.
-
Recommendations based on live event feeds from third-party platforms like Resident Advisor.
-
Social-based recommendations suggesting artists that followers were engaging with locally.
After evaluating technical constraints with engineers, we moved forward with an MVP experiment, focusing on location-based recommendations for music discovery.
After pairing with engineers, we identified technical roadblocks and adjusted the scope to focus on location-based discovery as an initial experiment.
Our goal was to run a test to evaluate user engagement, using SoundCloud’s existing recommendation engine enhanced by location-based inputs.
Although I left the company before the test concluded, this project reinforced the importance of starting with technical feasibility and closely collaborating with engineers early in the design process.
Outcome
💡 Solution
🏁 Impact
🧠 Learnings & Takeaways
Our MVP solution focused on two key areas:
1. Location-Based Music Discovery
-
Users would opt into location services or manually select a city to receive personalized music recommendations.
-
SoundCloud would surface popular and emerging artists in that area, helping users discover local talent.
-
Created a module on the home page allowing users to browse music trends in different cities (e.g., “Current,” London, Berlin).
-
Developed SoundCloud Radar pages for each location, featuring local artists, trending tracks, and genre-specific trends.
2. Event-Based Music Recommendations
-
Integrated with third-party event feeds (e.g., Resident Advisor) to suggest upcoming concerts and club nights based on a user’s listening habits.
-
Used social connections to show events attended by artists a user follows.
-
Suggested related artists performing locally, creating a seamless connection between streaming and live music experiences.
This initial concept was developed as an experiment, with plans to iterate based on user engagement data.
Unfortunately, I left SoundCloud before seeing the final results of the experiment, and I believe the idea was never fully implemented. However, I later noticed Spotify launched a near-identical feature, surfacing concert recommendations based on listening habits and location data - a similar concept to what we explored!
Had I stayed at SoundCloud, I would have pushed harder for a full rollout, as I saw firsthand how this feature could enhance engagement and bridge the gap between digital streaming and real-world music experiences.
What This Project Taught Me
-
Start with the technical. I learned how to collaborate with engineers early to ensure concepts were viable before investing in design.
-
Agility and rapid iteration. Working in a hackathon-style sprint taught me how to generate, test, and pivot ideas quickly.
-
Workshop facilitation. This project strengthened my ability to lead structured ideation sessions, a skill I’ve since refined through my AJ&Smart Design Sprint certification.
If I Had More Time…
I would have advocated more aggressively for this feature to be fully developed, especially after seeing Spotify successfully execute a similar idea.
Pick my brain 🧠