Splice

Project Brief

I was a Splice user before I joined the team.

A friend and I used Ableton Live to write music together, syncing our sessions with the Splice desktop app. We both went to an event at their office in New York, met the team, and saw how scrappy the whole thing was. Soon after, I joined as the full-time product designer, taking over from a contract designer who had been carrying the marketing site, the product interface and the direction of the product all at once.

He left a month or two after I started, so I took on what he had been carrying. The company was about ten people then. It grew to over a hundred and launched two businesses that did not exist when I arrived.

Nearly everyone at Splice was a musician. That’s the unglamorous reason the early decisions were good: we all knew what it was like to work in a half-digital, half-analog mess.

The product they hired me onto

Splice originally offered free, unlimited backup and version control for your projects. By syncing through the desktop app, your sessions were saved, versioned, and shareable. At the time, other cloud storage moved a session without understanding it: no version history, no samples resolved, no guarantee it would open on the other end.

To make this work, the system needed to open and understand the session file. That is the part that turned out to matter. Instead of treating a synced project as a black box, Splice could show you the tracks inside, what they were named, and which plugins were on them.

During that time, I designed the syncing interface, the project dashboard, and the marketing pages.

The most challenging thing we built was the Community, which became a product in its own right. Synced projects could be published in an editable form, for other people to remix, sample, or finish. We built the DNA Player so those internals could be read on the open web, and so an artist could annotate one specific moment and say what a decision was actually for. Sometimes we released a session from an established artist, and you could open it up and see how a finished track was put together.

We discovered that asking someone to publish a music file, with no contract and no expectation attached to it, is an enormous ask. We had no way of knowing that when we built it. And at the same time, the Community was one of the main ways Splice grew while it was still finding traction. Both of those are true, and they never resolved.

But the sessions people did choose to publish told us two things.

People clearly enjoyed making music in a browser, which is where Beat Maker came from. From the sessions they had chosen to publish, we could see which plugins producers actually used, so we put out an early plugin directory. That became the schema for the ecosystem: the one underneath Splice Sounds and, later, the Rent-to-Own plugin marketplace.

Splice retired the sync product in 2023. The company said it hadn’t been a focus since 2017, and that keeping it running slowed other work. That’s true. It’s also incomplete. The product that got retired is how we knew what to build next.

Beat Maker

Beat Maker lets anyone with a browser create a beat using samples from Splice Sounds. There’s no need for an account, an install, or a digital audio workstation. A step sequencer is one of the clearest ways to show how a beat is built.

I worked on Beat Maker with a UX designer who later moved into product management, because his work on flows kept running into the business logic underneath. He designed the step sequencer wireframe. I did the visual design and the implementation, and we worked out together how the interface should behave.

Beat Maker also sold Splice Sounds. That is how a startup could justify giving away a free instrument: you learn to make a beat using samples you can then license. Knowing that does not diminish what it did for people who had never opened a DAW. It explains how it got built at all.

It found an audience we did not design for. Beat Maker ended up used in classrooms by Building Beats, a nonprofit that teaches music production to young people. In November 2018 Splice partnered with them on a sample pack, The Fantastic Sounds of Jay Dee, marking J Dilla’s influence on production, and donated the proceeds to their programs.

Splice Desktop

The desktop app had a problem that sounds trivial written down. It stayed in the toolbar and hid itself the moment you focused another window — which, for a producer, is every moment. Experienced users told us they wanted it reachable without leaving the digital audio workstation they were working in.

That is not a small complaint. With a creative tool, a minor inconvenience is enough to make the tool useless: the interruption costs more than the thing it interrupts. The behavior we were protecting was dragging a sample straight from Splice into a timeline, which only works if the app is there when your hands are already moving.

We redesigned the app in 2017, working as a product designer with a product manager and a UX designer, and it shipped in May 2018. The app became resizable between two deliberate footprints: a compact window that sits beside a DAW, and an expanded one for browsing properly. Search moved to the middle of the experience rather than the edge of it. It is a layout that adapts to what the user is doing, not to the size of their screen.

The desktop app was built with Electron, so my CSS shipped as a native macOS and Windows surface. That is the whole reason one person could carry web, marketing and desktop at the same time: it was one skill set rather than three. It is also where the work I now call design engineering actually started.

Around the same time Splice went through a full rebrand. R/GA London designed the logomark and initial strategy, and OMSE made the custom typeface (a cut of Modern Era). I rolled that system out across the desktop app, the web experience and the marketing site, alongside a restructuring of much of the front-end code that carried it.

Rent-to-Own

The plugin business started with Xfer’s Serum in August 2016. Three days free with the whole instrument unlocked, then $9.99 a month, each payment counting against the $189 it cost to buy outright. No interest, no premium, and you could stop paying and pick it up later without losing the ground you had made.

The design problem was not the price. It was that the entitlement system underneath it is invisible, and it has to be trustworthy to somebody who is trying to finish a track and is not thinking about licensing at all. The desktop app was the thing doing the authorising: it checked periodically, gave you three days offline before it needed to look again, and stepped out of the way entirely once you had paid the instrument off.

So the states are the product. Trial started, and when it ends. Payment taken, and how much closer that puts you. Paused, and what you keep. Declined, and how to recover without losing access mid-session. Paid off, and the app is no longer required at all. Each of those is a moment where someone either trusts the arrangement or starts to suspect it, and none of them are the moment they came to make music.

Splice Sounds was designed in part by another product designer, who left not long after it launched. The product manager and I carried it afterwards, through the iteration and the engineering, which is the less glamorous half and the one that takes years.

What we got wrong

We said yes to too many things.

We could have focused on the Sounds business, and only that, much earlier — if we had decided sooner to stop supporting the sync product. It is genuinely hard to know which features to bet on while you are inside it, and we never wanted to turn our backs on people who used something every day. That instinct is a good one and it still had a cost.

That’s why the call finally had to be made by a CEO who arrived after all of us. It was a tough one, and it was correct.