Let's create luxury problems!

"Suppose we get a million users!" Or why choosing Amazon Web Services can be a foolish choice.

If you simply want a sandwich, you don't start off by building a kitchen first, just because you might want a thousand sandwiches later on! It seems a no-brainer, but many projects are approached in this way.

When people start a new project they often plan for perfection. They prepare for the all kinds of eventualities and therefore start with strong fundamental foundations. Sounds professional, but it often distracts from the real issues.

Generally, the most important issues of any new IT project are not about scalability or security. But rather, do the essential parts work? And, did we deliver something the end user and the customer actually want and envisioned?

Creating success is more important than building solutions in case of success.

Once that is achieved, which is already quite a feat in itself, other issues like scalability, design and security should be addressed.

So, when starting a project, you don't need a full-fledged (AWS) configuration with load balancers, database replicators, a cluster of application servers, etcetera. All because of an unlikely scenario: "Suppose we get a million users!" You can, in principle, start with shared hosting. Or VPS. I have done so and even postponed registering a new domain name, I just set up a subdomain. It means that you can start the same day with releasing the first version. Customers generally like this practical approach.

I know this all sounds very unprofessional, but professionalism is often an excuse for very elaborate preparation (read: missed deadlines).