One of the more uncomfortable conclusions I've reached over the years is that a lot of software is built backwards.
Not because developers are bad.
Not because teams are incompetent.
Because we're often rewarded for the wrong things.
Features get attention.
Architecture doesn't.
Announcements get attention.
Documentation doesn't.
New capabilities get attention.
Maintenance doesn't.
As a result, many projects begin with the visible parts of the product and only later discover they neglected the foundation.
We Start With Features
The typical software conversation goes something like this:
What features should we build?
What should the dashboard look like?
What integrations should we support?
What can we put on the roadmap?
What can we announce next?
These aren't bad questions.
They're just usually asked too early.
Because before discussing features, we should probably understand the system those features will live inside.
The Foundation Determines The Experience
Imagine building a house.
Most people don't start by choosing paint colors.
They start with:
- The foundation
- The structure
- The plumbing
- The electrical system
The visible details matter.
But they depend entirely on the invisible systems beneath them.
Software isn't much different.
The user interface is visible.
The architecture isn't.
The application is visible.
The workflows aren't.
The features are visible.
The systems supporting them aren't.
Yet those systems often determine whether the software succeeds.
Products Are Built. Systems Are Designed.
This is one reason I've become increasingly interested in systems thinking.
Features are built.
Systems are designed.
Features solve individual problems.
Systems solve categories of problems.
Features create functionality.
Systems create consistency.
When software focuses entirely on features, complexity tends to accumulate.
When software focuses on systems, complexity tends to organize itself.
Documentation Reveals The Truth
One thing I've noticed is that documentation exposes whether a system was designed or assembled.
A well-designed system tends to produce clear documentation.
A poorly designed system often requires increasingly complicated explanations.
If a workflow requires several pages of documentation to explain, the issue might not be the documentation.
It might be the workflow.
Documentation acts like a mirror.
It reveals architectural decisions that would otherwise remain hidden.
The Real Product Is Often Invisible
The longer I build software, the more I realize that users rarely experience individual features.
They experience systems.
Users don't experience:
- Authentication
- APIs
- Database queries
- Deployment pipelines
They experience:
- Reliability
- Speed
- Simplicity
- Confidence
Those experiences emerge from the system as a whole.
Not from any single feature.
Why Builders Fall Into This Trap
The reason software often gets built backwards is understandable.
Features are easy to demonstrate.
Systems are harder to showcase.
A feature fits in a screenshot.
A system doesn't.
A feature can be announced.
A system is usually invisible.
As builders, we naturally gravitate toward the visible work because it's easier to communicate.
The challenge is that invisible work often creates the most value.
Start With The System
I've started approaching projects differently.
Instead of asking:
"What features should this have?"
I try to ask:
"What system am I building?"
Once that becomes clear, many feature decisions become obvious.
The system creates constraints.
The system creates priorities.
The system creates direction.
Features become easier because the foundation already exists.
Final Thoughts
I don't think features are unimportant.
Far from it.
Users ultimately interact with features.
But features are only one layer of a larger structure.
The most successful products I've used weren't successful because they had the most features.
They were successful because the underlying system was thoughtfully designed.
The software felt coherent.
The workflows felt natural.
The experience felt intentional.
That's why I've started believing that many software projects are built backwards.
We start with the visible pieces.
Then scramble to create the system underneath them.
Maybe we should reverse that process.
Build the system first.
Then let the features emerge from it.
Top comments (0)