Flux
Flux was a modular multi-messaging client designed to let users manage different chat channels (Facebook, Email, etc.) in one place while owning their data. Despite raising an angel round, the startup failed due to extreme technical over-engineering, poor co-founder fit, and a "walled garden" shift in the industry as major platforms closed their APIs.
The Autopsy
| Section | Details |
|---|---|
| Startup Profile | Founders: Jan Johannes Funding: ~€70,000 Angel Investment |
| Cause of Death | Financing Failure: Yes Market Fit: Yes |
| The Critical Mistake | Technical Over-engineering: The founder spent excessive time building complex Domain Specific Languages (DSLs) in Erlang to automate REST API connections rather than just building connectors manually for an MVP. The API Shutdown: Mid-development, Facebook and Google shifted from open XMPP protocols to "walled garden" strategies, breaking Flux's core functionality. The "Contract Rabbit Hole": When the team finally tried to pivot to B2B, they got trapped in months of legal negotiations with a large German corporation. The corporation's restructuring wiped out the deal just as Flux ran out of cash. |
| Key Lessons |
|
Deep Dive
In his interview with Failory, Jan Johannes reflected on the hubris of a technical founder trying to "impress investors with tech." The Complexity Debt: Jan allowed his developers to use whatever languages they found "interesting." This led to a "Language Zoo" of Erlang, Go, Ruby, and JavaScript. While fun for the engineers, it became an absolute nightmare to maintain or scale with a small team once developers started leaving. The Investor Trap: Jan admitted that a lot of their over-engineering (like the Erlang DSL) was built specifically to show "secret sauce" to VCs. He realized too late that VCs in Berlin at the time weren't as interested in deep-tech infrastructure as he thought. He should have focused on sales and traction first. The Legacy: Flux is a classic case of "Technical Idealism vs. Commercial Reality." It serves as a reminder that your tech stack should be a means to an end, not the product itself. Jan now works as a freelance consultant, having open-sourced the Flux code and learned that "consulting is a better way to build a product than chasing VC money."
Key Lessons
Technical Idealism vs. Commercial Reality: Your tech stack should be a means to an end, not the product itself.
The "Language Zoo" Mistake: Using whatever languages developers find "interesting" becomes a maintenance nightmare.
The Investor Trap: Building "secret sauce" to show VCs instead of focusing on sales and traction.