Bootstrapping Flagsmith to $3m ARR
Flagsmith founder Ben Rometsch explains how an agency side project survived years of near-zero revenue, why enterprise buyers unlocked a path to $3 million ARR, and how patient peer recognition became a stronger compass than starting with a financial target.
Ben Rometsch is the founder of Flagsmith, an open-source feature flag and remote configuration service that was approaching $3 million in annual recurring revenue at the time of this conversation.
Its path was unusually patient: spare agency capacity funded the product, early hosting revenue barely covered the distraction, and a sweat-equity partnership with Polychrome supplied the go-to-market knowledge required to serve enterprise demand already arriving inbound.
The Q&A below is experimental — extracted from the episode transcript with AI. If you spot a mistake, let us know at jack@scalingdevtools.com.
How did Flagsmith begin?
Our London software agency used downtime to build ideas, then scored them against practical criteria. Feature flags stood out because we were the target customer, a useful MVP did not require enormous engineering, and a request for Firebase web support had attracted unusual interest on GitHub. Three of us began building in spare time in 2017, years before creating a separate company around it.
What happened during those first quiet years?
Almost nothing. The project was called Bullet Train, had a name that could not rank in search and acquired perhaps one or two hosted customers a month on $29–$100 plans. A single larger website could send five times everyone else’s traffic and require agency engineers to nurse the API for revenue worth roughly an hour of work. Continuing the project was genuinely in doubt.
Why keep going when it was barely making money?
Small positive loops accumulated. Reaching $1,000 MRR suggested that more SDKs, better hosting and steady product work might eventually become a profitable company, even if you had to squint to see it. We spent almost no cash beyond infrastructure and supplied the business through engineering time. Because the agency paid the bills, the project could improve at a pace its early revenue could never have financed.
What changed between $1,000 and $10,000 MRR?
We met the three founders of Polychrome at exactly the right moment. They supplied a small amount of money and committed their own work until Flagsmith reached $40,000 MRR. Their backgrounds in companies such as Amplitude and Twilio brought enterprise finance and sales knowledge; we brought engineering, open source and product. The partnership filled the capability gap rather than merely adding capital.
Where was the enterprise opportunity hiding?
Large companies were already asking how to self-host, but we told them to use the open-source code because we did not understand what they were trying to buy. They wanted a license, a contract, insurance, Kubernetes deployment help and someone accountable for sizing and support. Polychrome understood that package and its price. The demand was inbound; the missing piece was knowing how to turn it into an enterprise product.
What did the partnership make possible financially?
The investment covered the agency revenue gap while several of us worked on Flagsmith full-time. We modeled the runway to $40,000 MRR because that level could pay cloud costs, me and a small engineering team. If growth had lagged, we could have reduced spending and waited for revenue; in practice, enterprise licenses sold well enough that the fallback was unnecessary.
How did Flagsmith reach nearly $3 million ARR?
The company was approaching $3 million ARR after growing about 250% in the previous year and reinvesting into the team. Its position also became clearer: Flagsmith focuses on engineers improving delivery rather than making experimentation and product analytics the center. Enterprise customers may simply need to decouple deployment from release, add approvals and auditability, and safely operate old, complex systems.
What does “simple” mean to an enterprise customer?
It means straightforward hosting, an understandable platform and excellent help getting teams into production. The technical change can still feel consequential when mission-critical systems and regulated processes are involved, so adoption may take a year. Open source gives buyers visibility and the ability to contribute; some paying customers have sent several pull requests after the sales process because they can directly improve the SDK or core product they rely on.
What should an earlier-stage founder take from the story?
Do a little useful work consistently and let it compound. We began because building a successful open-source project and earning recognition from peers was intrinsically exciting; an unexpected contribution could feel as meaningful as the first sale. Stars, issues, contributions and repeat users reveal value. If you focus on that community response, a commercial opportunity can emerge more naturally than beginning with an arbitrary revenue target.