Can you deploy in one step?

Conversation on April 1:

— When is your release?

— April 15.

Conversation on May 1:

— When is your release?

— May 15.

— But it was supposed to be April 15 🤔

— When we tried to release, we discovered that our solution could not be deployed to the required platform. Now we’re rebuilding it.

I see two to three cases like this every year.

Three months of work — and no one checked whether the solution could be deployed at all.

A simple Hello World deployed in week one would have saved them months.

Technical people naturally tend to focus mostly on the code; that’s what is expected of them. However, if you combine engineering thinking with a managerial one, the results will be strikingly better.

Let me give an example.

Early in my career, Joel Spolsky influenced me a lot.

His radical question still matters: “Can you make deployment in one step?”

He believed deployment must be so automated that anyone can press a button and ship the product.

This forces you to build systems that make it impossible to deploy garbage 🗑️

It’s about faster iterations.

Automation shortens the path from idea → result → feedback.

Fast iterations are a managerial idea — about shortening the managerial cycle: the time from task to feedback. Here it is combined with an engineering process.

And when you combine engineering + managerial thinking,

you become more valuable.

💰 And that always converts into money: company profit and your salary.