Uppbeat
Overview
Reduced churn by 15% by redesigning the cancellation flow to clarify options, address user concerns, and introduce retention offers that kept more subscribers engaged.
Approach
Looked into the data to understand the main causes for cancellation.
Identified key issues with the flow in place.
Provided help along the journey to solve the user's pain point before they cancel.
The problem
Uppbeat is a platform with royalty-free music, sound effects, and motion graphics, serving over 3.5M creators - mostly hobbyists and early-career YouTubers. While user growth was strong, churn was a critical challenge:
15–18% of churn happens within the first month, with most churn concentrated within the first 7 days and renewal dates.
After month one, ongoing churn stabilizes around 5-6% per month.
~9% of users are cancelling because they received copyright claims.
26% of users are cancelling because they struggle to find the assets they need.
The insight
Cancellation wasn’t just an exit point - it was an overlooked moment where we could educate users, address pain points, and potentially retain them.
My process
Research & analysis
Mapped out the top cancellation reasons and grouped them into “supportable” (e.g. asset discovery, copyright issues) vs. “non-supportable” (e.g. no need, budget).
Ran user surveys to better understand friction points in asset search and copyright protection.
Benchmarked cancellation journeys from streaming and SaaS products to find patterns that balance business needs with user trust.
Through research, I mapped the optimized cancellation journey and uncovered key flaws in our existing flow.
Design iterations
Intervention points: Instead of a linear cancel button, users first selected their reason. We then surfaced tailored solutions:
Playlist Generator (AI-powered track recommendations).
Direct access to the catalog team via Discord.
Priority support for technical issues or claims.
Ethical alternatives: We tested versions that offered easy exit but also clear options to “fix it now.”
Education moments: Clarified what cancellation meant - loss of unlimited downloads, copyright protection, and premium catalog access.
Post-cancel alerts: Gave users an easy path back to re-subscribe.
Final designs
Impact
👉 Drop-off at intervention screens increased from 40% to 58%, showing users paused to reconsider.
👉 ~10% of users who engaged with suggested solutions (e.g., AI playlist) ultimately retained their subscription.
👉 Churn attribution to “supportable reasons” (finding tracks, copyright claims) dropped by 15%.
👉 Post-cancel banners led to 3% of users returning to re-subscribe within 14 days.
Learnings
Legal & compliance alignment needed earlier: Some ideas had to be reworked late in the process.
Infrastructure readiness: Existing features like the AI Playlist Generator and safelisting were critical to supporting interventions.
Transparency built trust: users appreciated being informed about what they’d lose instead of being blocked.
For the future
💡 Plan flexibility: Beyond incentives, we could also explore offering options to downgrade to a more affordable plan or pause the subscription.
💡 Experiment with a full-page experience: The current side modal has limited space, so in the future it may be worth testing a full-page flow to provide a smoother, more engaging experience - especially on tablet and mobile.
