MVP dilemma: Improve the architecture or go live?
I have built a MVP for sodim.dev But i am not very happy with the architecture design, and i want desperately to change the architecture to more of stateless microservices. But given all the nice posts and blogs, including those of YC specifically; i also don't want to spend too much time improving the app which doesn't have any user yet. What should be the wise decision here?
GO LIVE!
Nobody cares about your architecture except you :)
Go live! I had the exact same question last week - go live with Next.js api pages that timed out after 10 seconds for more advanced requests, or wait until I replaced with serverless functions. Instead, went live with alpha users, and built in the reliable serverless functions in the meantime, throughout the week (and personally explained to them the limitations at the time)
I think it’s essential to establish a cadence of releases and showcases, even with suboptimal architecture. This week on indie hackers, users discovered a niche cookie/browser problem which makes the service unusable, that could have arguably been preemptively found with E2E testing - the embarrassment of that blunder is offset by the momentum gain of continuing to release and get feedback…with E2E being implemented this weekend :)
In short: Go live!
Go live! Completely agree with @xavier, no one cares about how your app is implemented. If it works, roll with it and refactor later.