Deciding what to build with.
WHAT TO CHOOSE ON
What you and your team can build and maintain quickly Availability of people who know it Maturity of libraries for what you need Hosting cost and availability
WHAT NOT TO CHOOSE ON
Performance benchmarks Novelty What large companies use
WHY THAT LAST POINT
Their problems are not yours, and their solutions carry costs you cannot absorb.
WHAT MATTERS MOST EARLY
Speed of change.
WHY
You will rewrite much of it as you learn what the product should be.
WHAT MATURE FRAMEWORKS PROVIDE
Authentication, authorisation, background jobs, migrations and testing, already solved.
WHY THAT MATTERS
Building those yourself consumes months and produces nothing a customer values.
WHAT TO AVOID EARLY
Microservices Multiple languages Elaborate infrastructure Anything requiring specialist operations
WHY MICROSERVICES SPECIFICALLY
They solve organisational problems you do not yet have, at substantial operational cost.
WHAT TO START WITH
One application, one database.
WHEN TO SPLIT
When a specific part genuinely needs separate scaling or isolation.
WHAT TO ESTABLISH ABOUT HOSTING
Cost at your expected scale Whether it can be paid for from here Where data will reside
