Soundtrap/Spotify
Contributions
Product proposals
UX/UI
Prototyping
Team
UX, Design, Prototyping (Myself)
UX, Design, Prototyping (Andreas W)
Product thinking (George M)
Product thinking (Joakim P)
Engineering (Johannes B)
Engineering (Pedro N)
Engineering (Song L)
Engineering (Johan)
Engineering manager (George B)
Sr product manager (Raquel S)
Project Lead (Per E)
Spotify Leadership (Gustav, Charlie, Sten)
We explored AI-assisted music tools for creating song covers by building upon Soundtrap’s existing platform. With features like guided steps, AI instruments and personalised workflows, we envisioned a product within the Spotify ecosystem that drastically lowers the barrier of entry into music creation.
Unlocking strategic relationships between artists and fans
We imagined new value chains between artists and creators, through fan engagement and content remixing, A creation tool personalised to your music taste.
We bridged the gap between inspiration and creation by embedding the experience directly into your spotify library.
Early explorations.
Early explorations were driven by ideas around plain language creation, hiding complexity and step by step guidance.
Ver 1.0
01 Hides complexity with track groups.
02 Guides user with steps.
03 Allows big changes with few clicks through Ai powered instruments
01 Dive deeper with expanding groups.
02 Access to full studio.
First Principles guiding the product.
Limited access to research and testing constrained our team in gathering feedback and direction. We distilled the problem to its core and defined a set of First Principles to guide the development.
Creation over modification.
To give a sense of ownership we gave greater control over instrument sounds by including more steps.
Unique yet familar.
To let users leave their mark we gave them innumerable instrument permutations to play with.
Sound shaping over arrangement.
To make things easy we automated a lot of the arrangement work and let them focus on creativity.
Instrument UI.
To foster quick exploration we adopted an creation canvas with interactive circles that represented an instrumentalist. These could be tweaked and arranged easily.
Project structure.
Breaking down a song into its distinctive parts helped to guide the user, hide complexity and focus on the most crucial elements.
such as drums, vocals, bass.
i.e. intro, verse, chorus.
Final Iteration.
In our final iteration we introduced time-stamped lyrics as a way to navigate the song. Given the context of making song covers, it was a more appropriate way of interacting with the timeline than conventional UI.
A dedicated creation canvas allowed users to play around with various instrument sounds. Novel interaction patterns, first principles thinking and rapid prototyping, helped us create this final proof of concept.
Demo of instrument interactions.
This prototype shows how Ai powered instruments could be used to create a signature sound. It went on to get patented
Defining success as Pilot readiness
We defined success around the following -
User Impact - Faster time to completion of a project
Business impact - Cohort sign in rate.
System impact - Patent output, reusability of soundtrap architecture, scalability
This correlated to -
-
Task Success Rate: Percentage of users completing a song cover workflow (target: >70% in remote usability sessions with 10-20 participants simulating the demo interactions).
-
Time on Task: Average time to create a basic track (target: <15 minutes with the milestone set at one audio track recording)
-
Pilot readiness >70% interest shown from internal users
Final outcomes
Task Success Rate: 80% percentage of users completed the song cover workflow
Time on Task: Average time of 10 minutes to create a basic project .
Patent output: 1 patent achieved