Early-stage founders often ask which technology will scale to a million users. It is a fair question, but it is usually the wrong one to optimise for. At the start, the goal is to learn what customers actually want as quickly and cheaply as possible.
Choose boring, proven tools
Well-established tools have better documentation, larger communities, and fewer surprises. That means faster development and easier hiring later. New, exciting technology is great for experiments, not for the product your business depends on.
Our default starting point
For most web products we start with Next.js and TypeScript on the front end, a PostgreSQL database, and an ORM like Prisma. It is fast to build with, easy to deploy, and comfortably handles far more traffic than most startups see in their first years.
Do not build what you can rent
Authentication, payments, email delivery, and file storage are solved problems. Using trusted services for them saves weeks of work and avoids security mistakes, which lets your team focus on the features that make your product different.
Keep it one codebase as long as you can
Microservices solve problems that large teams have. For a small team, they mostly add deployment complexity and slow everyone down. A single, well-organised codebase is simpler to change, test, and understand.
Plan for change, not for scale
Your product will change a lot in its first year. Clean code, sensible database design, and automated deploys matter far more than an architecture built for traffic you do not have yet. You can scale a working product; you cannot scale one nobody uses.
