A breakthrough workflow for capturing musical ideas in context and discovering compatible assets to keep your flow going.
Drop in any sound; from hummed ideas, to song sections, to MIDI, and retrieve filterable lists of sonically compatible assets. In key, on tempo, and playable beat-matched over your work.
Background
Splice is a multi-faceted subscription service for musicians, best known for its high quality catalog of royalty-free musical assets.
DAWs are digital spaces for recording, arranging, and mixing music. Pro-Tools and Garage Band are examples of DAWs.
In a DAW, musical context can be leveraged to optimize and automate how Splice content gets found, delivered, and used. Many new opportunities branch from there.
Opportunity
Early 2024, the makers of Studio One (a DAW), approached Splice about partnering on their upcoming release.
We took a request for content and grew it into something much bigger: an opportunity to release a ground-breaking AI assisted workflow for music makers.
We created a first of its kind product that can intake musical ideas and sort Splice's catalog by compatibility, while offering filtering for additional control. It's the best tool that's been released for getting unstuck and finishing music.
My Role
I led design on all aspects of the product experience. Initial concepts, iterative revisions to find the product opportunity, prototyping, web flows for sign up and account management, UX, UI, etc.
I collaborated with colleagues on invaluable pre-release research, initial webpage concepts and copy, through many iterations together until we had it right.
What Problem does it solve?
TIME TO FAILURE
One pitfall in the Splice experience had been a workflow gap between finding musical assets that are interesting, then confirming that they will work in context. The prior workflow included surface switching, context switching, and repeated editing and review to determine whether an asset could be modified to work. We wanted to shorten the time and effort to failure to something close to 0.
DILEMMA OF CHOICE
Splice has the best catalog of its kind in the market. The quality of the concepts, the execution, the sounds themselves, has been widely credited with improving the music we hear on the radio. However, as they approach the 2 million mark for assets, adding to this massive catalog will naturally have diminishing returns. The more content a cloud service offers, the more they add to the difficult problem of how a user navigates those vast choices. We sought to avoid this issue altogether with a new discovery workflow that will improve as the catalog grows.
Who is it for?
We initially had newer musicians in mind, and our first concepts were about helping people get started. We called that the "blank slate" problem. However, we failed to validate this as priority problem with our audience. Rather the opposite. Many music makers, beginner to advanced, have abundant ideas. The problem musicians care about most is finishing.
So, we made a major pivot to a product experience that assumes a user can bring their own ideas. We focused on helping them expand, extend, and complete those ideas.
We also gave careful consideration to our context as a subscription add-on to a newly purchased product, as well as the overall busy nature of DAWs. To address this, we focused on clean inviting UI and built the first ever Splice free trial that includes 100% unlocked use of the product. People can try the integration with no risk. We then added always-free assets for our most subscription wary customers. In their case, the initial product goal is for them to discover the superior quality of Splice sounds.
Preview in context
To eliminate the gap between finding a sample we love, and learning whether it works in our song, all samples preview in key, at the right BPM, and most importantly, start playback on beat over your music. Users know instantly if a sample is working, and can fly through previews as their music continues to play.
Search with Sound
This is the flagship feature. Splice has developed musical compatibility AI, a service that can pair sound by how well they fit together. The tech was initially releasing in Create, a loop builder that uses a stack of samples as reference to find the next best layer. We asked if it might be possible to use that tech, CoSo, to sort long, filterable, lists of samples rather than finding one at a time. The machine learning team supported us with an API update before we launched.
The end result is an AI assisted music making workflow that's highly user driven. Give it context and a job, and it delivers materials to inspire and fuel the next steps. It's a machine for getting unstuck; it drastically improves the speed of sample discovery, accuracy, and implementation.
The humble onboarding
Our simple onboarding was the space where we iterated the most on design. One of our foundational design goals was to be minimalist and efficient, so we mostly avoided talking to the user elsewhere in the product. The onboarding is the crucial space to explain what this tool is for, what Splice is, and entice people to try it. More than anything, through many rounds of testing and feedback, we revised our language. For example, it took 6 months to land on Search with Sound as the phrase that made sense to testers! We learned that the more we tried to explain our features, the more confused people were. When we kept the copy simple most testers intuited how the tool worked correctly.
SCREENS
Logged out homepage
Search Panel
Regular search results
Search with Sound hover
Search with Sound results
About ME
My design process is built around understanding why. I've always enjoyed trying to understand things. I'm grateful to do it for a living.
I've been in product design since 2014. I have a background in AI, e-commerce, marketplaces, creative tools, cloud-based content, and 0-1 projects. I start each project by listening.
Most recently I served as a Senior Product Designer at Splice, helping people make better music.
I love donuts, baking bread, fashion, and damn good coffee.
Contact