Every problem looks like a job posting
Revenue's flat, so someone writes a job posting. Deployments are slow, so someone writes another one. The posting is the default answer to every operational question, and the org chart is the tool that holds the answers.
Most companies grow this way. Not by design. By reflex. A problem appears and the first move is to put a person in front of it. It works, so nobody asks whether it was the right mechanism.
A year ago the hiring stopped. Not as a policy - the need just disappeared. The tools changed. A small team started doing work that would have required a department.
The distance between seeing a problem and fixing it collapsed. No handoff documents. No standing meetings where someone's job was to make sure everyone else was looking at the same screen. If something broke, the person who noticed was the person who fixed it, and the fix shipped before the meeting would have started.
The conviction was strong. Small team, sharp tools, no overhead.
Then a meeting with a client's team. The question: how many roles existed because two systems didn't talk to each other. The room went quiet. Someone started counting on their fingers and didn't run out of fingers.
The roles were real. The people in them were good at what they did. But the work itself - carrying a number from one screen to another, reformatting a report so a different department could read it - existed only because of a gap. Nobody had gone back to close it.
That part still holds. Most headcount is reflex.
But there was someone in that room who had been the bridge between two systems for years. They didn't just carry the numbers across. They caught the ones that were wrong. They knew which fields lied and which reports drifted. The gap wasn't just a gap. It was a space where their judgment lived, and automating the bridge would have automated them out of it.
The line isn't clear. Most of the time, the job posting is a reflex and the reflex is expensive. Some of the time, the person in the gap is doing something the system can't see. Telling which one it is from the outside - that's the part that doesn't get easier.
Related work
Engraving Automation
Orders with hundreds of custom variants. Manual processing taking 45 minutes each.
See this project →