Bootstrapping vs venture capital: how Relume turned an agency into a product

Bootstrapping vs venture capital is usually framed as a choice you make on day one. Dan Anisse's story suggests a third path: start a service business, use it to learn the market and fund the product, then switch it off when the product can pay the bills.

Dan is CEO and co-founder of Relume, the AI-powered website design tool, and a founder who has gone from t-shirts to tech, with an acquisition along the way. In this episode of Now or Never, Dan sat down with Fra to talk about turning agency insight into a product, why Relume stayed bootstrapped, and how they integrate AI without losing design taste. Here's what founders can take from it.

The agency was the research project

Dan and his co-founder Adam came back from overseas with a hunch about the freelance web design and no-code space. Rather than guess at a product, they set up shop inside the problem. As Dan put it, they figured:

"the best way of learning about this space or this industry is to start an agency"

So in August 2020 they started Relume as a Webflow agency. Building site after site, they kept hitting the same wall: the repetitive work of building wireframes from scratch. Templates didn't solve that job. So they built their own internal component library and reused it on every project.

That internal tool became the nucleus of the product. Within about a year, they had the insight they'd been looking for, and it came from doing the work, not from a whiteboard session.

Productising agency work: from internal library to product

A lot of people talk about a productized agency as a services business with fixed packages. Relume went further and turned a repeated agency process into software. The way they made that leap is worth copying.

Split the founders, keep the revenue

Once the idea was clear, Dan and Adam divided and conquered. Dan kept the agency running, maintaining revenue and building up some initial capital. Adam focused on the product. In effect, the agency was their angel investor. They launched the Relume Library in November 2021.

Ship with the tools you've got

The first version had no custom engineering at all. It was Webflow as the web app, Memberstack for memberships and payments, and Jetboost for filtering. That's it. Dan's view is that you work within the constraints you have, so you're not waiting six months or blocked on hiring an engineer:

"...you work within your constraints"

He also noted there are now far better ways to build an MVP than there were three or four years ago. If a no-code stack could validate Relume, founders today have even less excuse to wait.

Build distribution before you launch

Relume didn't launch to silence. Long before the library existed, Dan and Adam were active in the Webflow community, making YouTube videos, joining podcasts and getting into conversations. Other agencies and freelancers already knew who they were, so the launch landed with an audience ready to look.

Set a clear rule for switching off the agency

Their rule was simple: cut the agency once the founders could fully pay themselves from the product's recurring revenue. They hit that within about five to six months of launch, and the agency phased out roughly a year later. Dan loved the agency, but a product that can impact thousands of people instead of five was too strong a pull.

Bootstrapping vs venture capital: how it changes the way you build

Relume is bootstrapped, and Dan was clear that it shapes almost every decision, starting with hiring. When you raise, he said, there's often an expected burn rate to hit, so you go and get as many people as you can. Bootstrapped, the question is different:

"So for us, it’s based off like what do we need? What do we need to build?"

That thinking led to their first hires being a designer and an engineer, and the team stayed at five for a long time. They've kept the habit as they've grown: work out what needs building, then who's needed to build it.

Dan isn't anti-funding. He acknowledged that scaling up fast can mean bigger enterprise value if you execute. His worry is about focus. Some founders who raise get absorbed in raising and building a company, and:

"they’ve kind of lose sight maybe in like achieving product market fit, which equals revenue, which equals validation"

Bootstrapping removes that option entirely:

"If we don’t build a product that people buy, we don’t have revenue, we don’t pay people. It’s as simple as that."

His point is that even funded companies eventually need to be profitable, and bootstrapping gives you the reps in being capital efficient early. If you're weighing up how to fund design and product work at an early stage, our guide to how founders should buy design covers the trade-offs from the other side.

A design-led team with light management

Relume now has a team of around 25. Dan said the jump from a team of eight or nine to double that was less about headcount and more about running three or four projects in parallel.

Their structure is unusual. They have four designers to nine engineers, a much higher ratio of designers than most product companies. Each designer works on at most two projects, which gives them the focus to own the outcome and, in practice, run the project. The single product manager's job is to unblock teams rather than manage them. Dan would rather scale small pods, roughly one designer to two engineers, than add layers of management.

To keep everyone across the work, every team presents at an end-of-week showcase called "wins of the week", so growth, sales and support see what product is shipping and vice versa.

Keeping design taste while integrating AI

Fra asked the question every design tool is wrestling with: how do you add AI without losing taste? Dan's answer is that Relume builds AI on top of a system. The components already bake in decisions about spacing and heading-to-body type ratios, and the style guide adds design variables. AI works inside those rules, which improves the odds of good output and keeps it consistent.

He sees that component system as Relume's point of difference in an AI era, because:

"AI can take you super wide, AI can lose consistency, AI is not good at sometimes understanding context"

We agree strongly. When AI makes generating screens nearly free, the system underneath and the taste encoded in it become the thing that separates good output from generic output. It's why we spend so much time on design systems, and why we've written about why so much AI-generated UX looks the same. If you're new to the idea, start with what a design system is.

Dan applies the same philosophy inside the business. Relume uses AI across design, research and repetitive support tasks, but the goal is to help people do more, not to replace them. The same goes for the product: helping web designers take on more projects or be more profitable on each one.

What to take from Relume's path

• Go and work inside the problem. An agency can be the best market research you'll ever do, and it pays you while you learn.
• Use services revenue as seed capital. It buys time to find the insight without giving away equity.
• Set a clear exit rule for the service business, like founders being fully paid from product revenue.
• Hire for what needs building, not for a headcount target.
• Put AI on rails. A strong design system is what keeps AI output tasteful and consistent.

There's more in the conversation, from Dan's first acquisition to Relume's plans for its users. Listen to the full episode with Dan Anisse on Now or Never.