5 Database Design Mistakes I Keep Seeing (And How to Catch Them Early)

I've reviewed a lot of database schemas over the years — my own, my teammates', inherited ones from projects I joined mid-flight. Some patterns keep showing up. Not the dramatic "we lost all our data" kind of mistakes, but the quiet ones that make every future feature harder to build, every query slower than it should be, and every migration feel like defusing a bomb. Here are five I see most often, why they hurt, and how to catch them before they reach production. 1. Missing Indexes on Foreign Keys This is the most common performance mistake in database design, and it's invisible until your tables have real data. When you add a foreign key from orders.user_id to users.id, most databases create an index on the referenced column (users.id — usually the primary key, already indexed). But they don't automatically create an index on the referencing column (orders.user_id). MySQL is the exception — it creates one implicitly. PostgreSQL, SQL Server, SQLite, and MariaDB do not. This means eve