Stop 08 · Technical Fluency for PMsTopics · 3/11

Databases 101 for PMs

In short

"Can we just add a field?" sounds tiny. Knowing a little about tables and relationships tells you when it really is tiny and when it touches half the system.

A basic mental model of tables, rows and relationships — enough to understand why 'just add one field' isn't always trivial, and to ask informed questions about data migrations.

The purple blob writes in a huge ruled register book on a wooden desk.

Think of it like this

The admissions register

The college office keeps a register: one row per student, columns for name, roll number, course. A separate fee register lists payments by roll number. Linking them by roll number is exactly how relational databases work.

Watch it happen

Tables linked by keys

Database

users

orders

id (key)

name, phone

user_id (link)

total, status

1/5A database holds tables. Each table stores one kind of thing: users, orders, products.

The big ideas

What you need to know

Tables, rows, columns

Like a spreadsheet with strict rules: each column has a type, each row is one record.

Keys connect tables

Splitting data into linked tables avoids copying the same info in many places.

Indexes speed lookups

An index is like a book's index. It makes search fast but adds cost to writes.

SQL vs NoSQL

Relational databases use SQL and tables. NoSQL stores like document databases trade some structure for flexibility.

See it in code

Walk through the code

sql

1SELECT u.name, COUNT(o.id) AS orders2FROM users u3JOIN orders o ON o.user_id = u.id4WHERE o.created_at >= '2026-10-01'5GROUP BY u.name6ORDER BY orders DESC7LIMIT 5;

1/4Pick the columns you want: the user's name and a count of their orders.

Top five customers by order count since 1 October. Being able to read a query like this helps you ask data questions directly.

Save yourself the pain

Common mistakes

Oops

Assuming every new field is a five-minute change.

Do this instead

Ask: is it a new column, a new table, or a change in how things link? Effort differs a lot.

Oops

Asking for "all the data" in one big export.

Do this instead

Ask a specific question. Precise queries are faster and safer for production databases.

Your mission

Write your first query

~30 min

Quick check

In an orders table, what does a user_id column usually do?

Go deeper

1Sketch a simple database schema (e.g. users, posts, comments) for an app you use.
2Learn the difference between a relational (SQL) and non-relational (NoSQL) database at a high level.
3Understand why changing an existing field's type can be a risky migration, not a quick fix.