Standing Out in a Crowded Plugin Market w/ Gavin Burke (Future Audio Workshop) | Ep 40

Three Things You Can Take Away From This Episode

• Hand AI the tasks you already know how to do, and keep the architecture and the sound in your own hands.
• In a crowded plugin market, your best protection is an idea that solves a real problem for a specific group of musicians.
• Launch small, let customers tell you what your product is, and only then spend money telling everyone else.

Gavin Burke has been building audio software since 2005. He was one of Arturia’s first employees, working on the analog modeling for the Prophet V and Jupiter-8. He founded Future Audio Workshop and released Circle and Circle 2, served as CTO at Output through the development of Arcade, then came home to build SubLab, the 808 bass synth that is now the best-selling instrument on Plugin Boutique.

His newest plugin, Notes, is a much bigger product: a chord engine, samplers and machine learning that pulls chords out of a loop. It’s also the one teaching him the most right now.

In this livestream, Josh Hodge and Gavin talk about how a plugin stands out in a crowded market in 2026, from paid ads and influencer fees to giving away something better than anyone else does. They also cover why plugin interfaces still matter, and what happens to the developer’s role when Claude writes better code than a 20-year veteran.

1. Gavin Went Out on His Own to See His Ideas Through

Gavin arrived at Arturia as an intern straight out of a master’s in signal processing at Queen Mary. The company had 8 to 10 people and was building a new audio library from scratch. The person doing the analog modeling had just left, so Gavin stepped into the role, working through DAFx papers and Julius Smith’s research to model the oscillators and filters and their nonlinearities across the spectrum.

Years of tuning resonant filters also cost him most of the hearing in his left ear. Keep that in mind next time you crank the resonance on headphones.

What pulled him toward running a company was control over the whole product. In a team, “best idea wins,” and he wanted to see his own vision through, including the marketing and the customers. His biggest lesson from the switch is resilience. “In a lot of ways it’s last man standing,” he says. If you keep going, you get there.

2. Start With Gut Instinct, Then Hand the Refining to Your Customers

Josh asked whether Gavin builds from data or from instinct. His answer has 2 stages. The first idea comes from a need he feels himself, the kind that shows up in the shower in the morning. After that, his own biases become a problem, and feedback has to take over.

“I think there’s a bit of magic in the idea,” he says, and he doesn’t expect data to produce it. Where data earns its place is refinement. Gavin reads everything: customer support tickets, YouTube comments on tutorials, KVR threads, anywhere the plugin gets mentioned. If he loves a feature and nobody uses it, he drops his attachment to it.

3. If You Can’t Explain Your Product in a Single Line, Your Customers Will Do It for You

SubLab explains itself. You press a key, you hear a hard-hitting 808, and it’s louder than every other ad on a phone speaker. Notes took about 8 man-years to build and does many things at once, and the team compressed it into the line “a new way to play chords.”

Gavin now calls that a mistake. The line didn’t say what problem Notes solves. So they held back on a big launch and ran a test release instead, then listened. Buyers on Plugin Boutique, Reddit and YouTube started calling it an “ideas machine,” something they open when they need a starting point. The team is now rewriting the website and the messaging around what users said.

Paid ads stay off until the message is right. Gavin doesn’t want to spend money advertising a pitch that doesn’t land. If you have a feature-rich plugin, expect the message to take longer than the code.

4. Hand Claude the Tasks You Know How to Do, and Keep the Architecture

Gavin resisted Claude Code for a long time. He compared it to Google Maps: rely on it to get from A to B and you lose your sense of direction. His team was allowed to use Claude for individual functions, but he’d seen agents make assumptions and change files across a codebase, which turned pull requests into a mess.

He tried Claude Code after his summer break and was surprised by how far the models had moved in 3 months. The clearest example was a graphics bug on macOS that affected both SubLab and Notes. Experienced developers had spent months on it without cracking it. Claude found it in about 5 minutes.

He still draws a line. Codebase architecture and DSP stay with a person who understands them, because employees and teams change and someone has to know what’s going on inside the code. He wouldn’t let a model loose on a new product the size of Notes, or trust it to make DSP sound good at a professional level. Tuning algorithms by ear is where he thinks developers keep their edge. Packaging scripts, installers and hunting a known bug for a month are what he’s happy to hand off.

5. Code Is Worth Less Now, and the Craft Moves Up a Level

Future Audio Workshop’s coding standard is simple: if it looks like good JUCE code, it’s probably fine. They don’t write unit tests. Testing means using the plugin themselves and sending it to beta testers.

Gavin is blunt about the rest. After 20 years, he thinks the latest Anthropic models write better code than he does, and he’s had to accept that a lot of his skill set is now redundant. The habit of concentrating on one narrow problem until it’s fixed still counts. Gavin thinks it moves toward structuring products instead of structuring code.

His image for this: are you studying the shovel, or digging the hole? Claude has taken over most of the shovel study. For new graduates, his advice is to aim that same detail-obsessed mindset at instrument design. How could a synthesis method be presented in a new way? What does a producer actually hit a wall on? Those questions still need someone who cares about the details.

6. The Idea Is Now the Moat

Gavin remembers when launching a plugin meant NAMM, Musikmesse, boxed software and distributors, all paid for before you’d sold a copy. Now you can build something, put it on your own website or Gumroad, and see if anyone cares. He’s watching ordinary producers do exactly that on Instagram, and some of their plugins stick.

Plugin companies used to fire features at the wall, and the feature list was the moat because few people could build something that complex. Gavin says that protection has mostly gone. What makes a new plugin stick now is solving a real pain point, obvious or not.

Circle 2 taught him this the hard way. The original Circle was a hit wavetable synth, released around the time of Massive and reviewed in Sound On Sound. By the time Circle 2 arrived, Serum had “out-featured us.” His conclusion: nobody needs another wavetable powerhouse next to Serum and Pigments, and Notes was planned with that lesson in mind.

7. Plugin Interfaces Carry Culture, and a Generic UI Can’t

Josh raised the prediction that user interfaces will matter less now that anyone can ask Claude for a dashboard or a CRM. Gavin agrees for business software, which carries no culture. Plugins are different because they’re instruments people use to express themselves.

A ’90s chiptune synth with a Winamp-style skin is a wink to its audience that says “we’re on the same page.” Gavin described a demo where a new model built a web piano sampler that played back MIDI by tapping its own keys. It worked, but it was something we’ve had for 30 years, and nobody will form a relationship with it.

Josh’s view is that people already treat DAWs as part of their identity (“I’m an Ableton person”), and plugins are heading the same way. When 2 plugins have near-identical feature lists, the interface is often what decides it. Gavin points out that even Ableton’s attempt to stay out of the way produced a recognizable minimal style. As Josh put it, “no identity is an identity.”

8. You Get 2 Seconds of Attention, and Trust Decides What Happens Next

Gavin treats marketing as letting people know something exists and why they should care. Most buyers are flicking through Instagram and will give you about 2 seconds. Hoping a good product finds its audience on its own is, in his words, “a little risky in 2026.”

So he’s comfortable with paid ads, as long as there’s something real behind them. His favorite recent example is a plugin whose founder ran Instagram ads telling his own story, why he built it, with simple audio demos. Gavin bought it. Without the ads, most people would never have heard of it. Without the product, the ads wouldn’t have mattered.

On measurement, Meta is where he can actually see return: pay for a free giveaway, collect 1,000 emails, count how many convert. He’s never gotten YouTube ads to attribute sales properly, though turning them off produces small dips. For his hip-hop audience, KVR banners don’t make sense, although KVR’s charts used to move sales the minute a release hit the feed back in his Arturia and Circle days. Reddit worked well for SubLab, especially when the team replied to comments under their own ads.

Support is part of marketing. Some customers arrive angry because other companies have burned them. Gavin’s wife, who runs customer support and co-runs the company with him, changed how he thinks about handling those moments. Showing up consistently with updates and answers is how you respect the time and money people give you.

9. With Influencers, Pay for Time, Add an Affiliate Link and Build a Series

When SubLab V1 launched, YouTube creators were hungry for new plugins and there was little money involved. The most impactful video ever made about SubLab came from Andrew Huang, who made a video on 808s entirely on his own. It has around half a million views and “kicked us into the big league,” and Gavin used the episode to thank him.

That era is mostly over. A layer of managers and fees has arrived in the last 2 to 3 years, and some quotes give Gavin sticker shock. His structure now is a fee that covers the creator’s time on a single video, plus an affiliate link. If it converts, he builds a longer relationship with a series of tutorials that help both the creator’s audience and his customers.

Josh added a practical tip: skip the people at the very top and approach the hungry ones on the way up, the way a new label signee is more reachable than the label owner. If you’re part of the scene you’re building for, you probably already know who they are.

10. Surprise People With What You Give Away

Future Audio Workshop hasn’t run much paid advertising lately. Even the Notes launch went mostly through the mailing list, which the company has been building since 2008. Gavin’s go-to-market “hack” is to give away something better than what everyone else gives away. Free samples used to be enough, and now people expect a free plugin. Then follow it with a paid product that clears the same high bar.

He suggests new developers do the math on their first release. You could sell 10 copies at $20, or give it away and put 1,000 surprised people on your mailing list for the paid version.

Josh brought up Steve Duda, who gave Serum 1 owners Serum 2 for free because he’d promised lifetime upgrades, with an optional “pay what you like” note. Gavin’s caveat is realistic. A company with 100 people on payroll may need to charge for an update. If your business can afford to overdeliver, do it, because it pays you back.

Closing

Gavin has spent 20 years watching the valuable part of plugin development move: from modeling nonlinearities by hand to today, when Claude can fix a months-old bug in 5 minutes. What hasn’t moved is the part that can’t be handed off: knowing which problem your users actually have, how your instrument should feel, and how to earn the trust of the people who give you their time and money. Notes is his live test of that, and he’s rewriting its story based on what his customers told him.