Framer's growth story is usually told as a pivot story. The company started as a prototyping tool, discovered that prototyping was not a large enough market, and became a website builder. That is true, and it leaves out the part worth copying: how the founders worked out what to build next.
The answer was not a new technology. It was a year of talking to users.
After years of building increasingly advanced prototyping software, Framer's founders hit a problem many startups reach. The product was good, customers existed, investors had backed the company, and growth had stopped. In a 2026 retrospective published by Framer, CEO and co-founder Jorn van Dijk said the prototyping business plateaued at roughly $4–5 million in annual recurring revenue, and that the team spent about a year on education and evangelism before accepting that something more fundamental had to change.
Instead of adding features, the founders went back to users and interviewed hundreds of them. A different problem emerged. Designers were not struggling to make better prototypes. They were frustrated that work they had already done had to be rebuilt somewhere else.
The Framer That Existed Before the Pivot
Jorn van Dijk and Koen Bok did not set out to build a website platform. The pair had worked together at Sofa, an Amsterdam software studio known for polished Mac applications. Facebook acquired the studio in 2011, and the founders spent more than two years inside the company before leaving to build something of their own. Van Dijk's retrospective describes several early experiments before the team settled on interactive prototyping.
The problem Framer addressed was familiar to software designers. Static screens could show what an application looked like. They could not show how it behaved. Framer closed that gap, letting designers build interactive prototypes with transitions, animations and real interactions.
The product worked and revenue grew, but the market did not develop the way the founders expected. Framer reached roughly $4–5 million ARR and stalled. The company had built a technically impressive product inside a limited market, which raised an uncomfortable question: was the problem the product, or the problem it was trying to solve?
When Growth Stopped, Framer Started Listening
The founders first tried to make the existing thesis work. They spent about a year educating the market, arguing that high-fidelity prototyping would matter more over time. It did not happen quickly enough. Van Dijk has described the position as a gray zone between having meaningful revenue and having genuine product-market fit, which left three options: sell the company, quit, or pivot. They chose to pivot.
Stalling at that size is not unusual in our Deep Dives. Submagic’s ARR sat at roughly $5 million for seven to eight months before it cut prices and launched a separate product for agencies. Framer’s plateau lasted longer and needed a bigger answer: not a new price, but a new market.
They did not begin by brainstorming new products. They began with users, and they changed the question. Not "what should we add to Framer", but "what problem are people actually trying to solve".
The research deliberately left the product's boundaries
That distinction did the work. Instead of researching how people used Framer, the founders looked at the workflow around it. What happened after a designer finished a prototype? What happened when a design had to become a real product? Who took over, what tools did they use, and what work had to be repeated?
The recurring answer was rebuilding. A designer creates something. Someone else recreates it in another environment to make it real. The tools change, the people change, but the work is duplicated. Framer's own account of those interviews puts it plainly: "rebuilding is inefficient." Van Dijk's shorthand in a later Y Combinator talk was blunter, that people hate rebuilding.
That reframed everything. The company had spent years asking how to make prototypes more realistic. Now it could ask why the prototype needed to be rebuilt at all.
The Product Changed Because the Problem Changed
The obvious change was the move from prototyping to website building. The deeper change was how the company understood its customer. The old Framer saw a designer as someone who needed a better way to communicate an idea. The new Framer saw a designer as someone who wanted to ship it. A prototype is an intermediate step. A website is an outcome.
Once rebuilding was identified as the frustration, the direction followed. Rather than adding another layer between design and development, Framer could remove the layer. Design the website, publish the website. The professional website builder launched in May 2022, and Framer's 2023 Series C announcement described the target explicitly: designers were creating work in tools such as Figma and then rebuilding it in website development platforms.
A Year to Admit It, Ten Months to Do It
The pivot timeline as van Dijk has described it. Dates and figures come from Framer's own retrospective.
Source: Framer's 2026 retrospective with Jorn van Dijk. These are the founder's recollections and company-reported revenue, not audited figures, and a story told after it worked tends to look tidier than it felt.
That sequence carries the lesson. Identifying the problem took about a year. Building the answer took nine to ten months. The hard part was not the website platform. It was working out which problem was worth building for.
The Old Framer Had Already Built the Foundation
Framer did not throw away what it had built. The team already had years of experience with browser-based visual design, interaction, responsive behavior, components, collaboration and sophisticated rendering. Those capabilities had been used to simulate production. Now they could produce it.
This is the part that makes the story repeatable and also limits it. The interviews did not tell the company to build an entirely new technology stack. They revealed where its existing capabilities would be worth more. A pivot that lands this well usually depends on that overlap, which is why "talk to users" is necessary advice but not sufficient advice.
The Numbers Followed the Reframing
A prototype helps teams communicate and test ideas. A website does something else. It represents a company, generates leads, supports marketing, recruits employees and serves as the primary public interface. That widened the customer base from designers to founders, marketers, agencies, startups and eventually larger organizations. The customer was no longer paying for better prototyping. They were paying to create and operate something the business needed anyway.
Where the Money Arrived After the Pivot
Funding and the scale disclosed alongside it.
Sources: Framer's Series C and Series D announcements and TechCrunch's reporting on the August 28, 2025 round. ARR, user counts and the break-even claim are company statements, not audited results.
Those figures do not prove that user interviews caused the growth. Product development, pricing, marketing, distribution, the creator ecosystem, enterprise adoption and the broader expansion of no-code software all contributed. What the interviews changed was which market Framer was competing in at all.
Users Became More Than a Source of Feedback
After the pivot, users became part of the distribution system. Designers built templates, produced websites, taught other designers, made tutorials, referred customers and built sites for clients. Feedback improved the product, better work attracted more users, and those users produced more feedback.
What the Creator Economy Pays Out
Figures Framer published in 2026 covering the previous year.
Source: Framer's 2026 announcement of Framer 3.0 and the new Community. Marketplace earnings are typically concentrated among top sellers, and nothing here contradicts that.
Declining that cut is a choice we have seen before: beehiiv takes no share of the subscription revenue its writers earn and charges for the platform around it instead.
Listening Also Changed Who Framer Was Building For
The original Framer was a specialist tool for designers. The website builder had a far broader audience, and the company expanded into CMS, analytics, collaboration, experimentation and enterprise features. By 2025 Framer described its ambition as becoming the platform where a company runs its entire .com, and TechCrunch reported that businesses had become the majority of its new customers.
That was a second-order effect of the original discovery. Once Framer understood that users wanted to eliminate rebuilding, it could ask what else could be removed from the workflow. The question pushed the company past website creation into website management.
Framer 3.0 put the AI inside the canvas
The same logic runs through the June 2026 release of Framer 3.0, which introduced AI agents that work directly inside the design and publishing environment. The agents can generate pages, modify layouts, work with CMS content, write copy and SEO metadata, build code components and audit a site. Framer also added branching, so agents and teammates can experiment without touching the live site, and connected external agents through a CLI.
The notable part is not that Framer added AI. It is where the AI operates. Rather than generating something that then has to be moved into another workflow, it works in the environment where teams already design, review and publish. The user gives an instruction, the system produces something, the user checks and edits it, and the result ships from the same place. The technology changed. The principle did not.
Framer Did Not Find Its Next Product in a Brainstorm
The interesting part of this story is not the website builder. It is how the company found it. The founders did not discover the opportunity by imagining a more ambitious version of what they had. They found it by listening to people explain where the workflow broke down.
The problem was never that designers could not prototype. They could. The problem was that their work stopped being useful the moment someone had to rebuild it somewhere else. Framer's opportunity was hiding inside that frustration, and once the founders saw it the strategy became simple: do not make people rebuild what they have already built.
It is worth being precise about what this does and does not prove. Framer's pivot worked because the interviews found a bigger problem that its existing engineering could already address, and because the company survived long enough at $4–5 million ARR to act on what it learned. The lesson is not that customer interviews produce growth. It is that they are how you find out which problem you are actually in business to solve. For another company that grew by removing a repeated chore rather than adding a feature, see our investigation into how LLM Gateway turned one API into $1 million in payments.
