When I Realized the Problem Wasn't the ERP
For years I thought the software was the problem. Then I set foot in a factory and realized I was looking at the wrong side.
For years I thought the software was the problem.
Slow ERPs. Ugly interfaces. Processes that looked written by someone who hated human beings.
I was convinced that building something more modern would fix everything.
Then I actually stepped into an industrial environment.
And I got something simple: the ERP wasn't the problem.
The problem was the distance between what happens on the factory floor and what the system can see. Which is a lot. Too much.
What the System Cannot See
In the warehouse and in production I saw things a pure developer would never notice.
Materials that exist but aren't where they should be. Orders blocked by components worth two cents. Production stopped for days. Purchasing finding out when it's already too late. Operators who already know something is going wrong long before the system does.
And that's when I started asking the real question.
Is the problem that the part is missing?
Or is the problem that the information arrives too late?
Because in most SMEs, the real flow doesn't go through software.
It goes through people.
Through:
- phone calls
- Excel sheets
- the memory of the guy who's been there for thirty years
- shouting between warehouse and production
Reducing Operational Latency
And this changed how I see industrial software. All of it.
I realized many companies don't need yet another giant ERP.
They need lower operational latency.
Meaning: the time between:
- the moment a problem appears
- and the moment someone decides what to do
That's it. Sounds obvious. It isn't.
A Simple Dashboard That Works
Years ago I built a very simple internal dashboard.
No hype. No AI. No absurd architectures just to look cool on LinkedIn.
Flask. Jinja2. Basic JavaScript.
Operators entered job orders, item codes and quantities. Production and purchasing finally saw the same data.
And the weird part: people actually used it.
They opened it during meetings. Operators worked on it. They started trusting the flow.
It wasn't pretty. Wasn't modern. It worked.
Real Validation
And that's when I got it.
Validation isn't when a founder says "great idea."
Validation is when a production manager opens your software on their own while trying to prevent a production stop. Without you asking. Without you even knowing they're opening it.
From that moment I started seeing software differently.
No more web pages. No more CRUD. No more modern frontends just to say I can build them.
Just systems that have to reduce friction, delays and operational chaos.
And that's probably when I stopped thinking like a web developer.
And started thinking like someone trying to understand how a company actually works.
I don't know if that's a "systems engineer." I know it's different.