Merging the Org Doesn't End the Argument. It Lets the Author Approve Their Own Pull Request.
Watch the video version:
For a stretch of my career, every decision on my team ran through me. Fifteen direct reports plus consultants: software engineers, QA, DevOps, technical writing, business analysts, and a scrum master, with no management layer between me and any of it. I approved and merged pull requests. I rewrote acceptance criteria on stories. Every call that crossed from one discipline into another was mine.
From the outside it looked fast. Nobody had to win an argument before work moved, because nobody was arguing. QA and DevOps used to tell development "not yet." Once every disagreement came to me as a single call, they stopped settling anything with each other, and after a "not yet" lost a couple of times, they stopped saying it. Delivery won every time, and nobody ever decided that it should.
Every argument waited in my calendar, behind the approvals and the merges. The long-horizon work only I could do, the architecture and the people, lost to whatever was urgent that day. I knew exactly what the team needed, and the structure made delivering it not impossible, but damn close. That gap is where the frustration and the madness came from.
What changed it was three managers: QA, DevOps, and development. Within a month, the development manager owned approvals, merges, and release management. The arguing came back. Each side had someone whose job was to argue it again, and the disagreements moved out of my calendar and back between the people who could settle them.
This month Canva told its people it was merging Product, Growth, and Platform into one organization with fewer management layers. I make no prediction about how it goes for Canva; nobody has shown yet whether this shape works or fails. The case for it is one every technical leader has heard: the fighting was overhead.
Put product, growth, and platform under one leader, and you get one roadmap, one set of priorities, and no more negotiation across org lines. Decisions get faster because nobody has to win an argument first. Some of that fighting was noise, and losing the noise is a real gain.
Engineers know this move from the other side. When an author and a reviewer keep disagreeing, nobody fixes it by letting the author approve their own pull request. The disagreement was the check. Platform and growth are the same pair. Platform protects what has to keep working two years from now; growth needs what ships this quarter. Both are legitimate, and neither can be removed without breaking something the business needs. That is a Structural Bearing™, and the argument between its two sides was how the organization knew what each side was paying.
Merge them under one owner, and the argument moves inside one person's head, where nobody can hear it. If that owner is you, the people who used to argue each side stop settling it with each other and send it up to you. Your calendar fills with whatever is loud and urgent, so the quiet side's concerns never reach the point where you decide them. The system learns that disagreements get escalated, not worked out, and that the same side wins. Nobody chose that. It is what happens every time.
If you lead a technical organization, you have inherited both sides of a bearing already, or you will. The leaders who hold a merged org well keep the argument running. Merge the reporting line, never the argument. Keep a reviewer: give each side someone whose job is to argue it, and a fixed point where that person is heard before you decide. That is a seat at the decision itself, before the call ever reaches you.
Take the last three trade-offs you settled between the two sides you now own, and write down which side won each one. If the same side won all three, the other side has nobody arguing for it, and the quiet in the room is one side losing every time. Name the person who will argue that side, and put them in the room before your next call on it.
The Structural CTO™ uses Structural Bearing™ to name the tension at the center of a growing company, then reads each seat for what it demands against what one person can carry, the question every merged org puts to the leader who now owns it: https://TheStructuralCTOBook.com
Why a company paid like a thirty and built like a ten is already paying for the lag, and how to see a cost the P&L never shows. (Chapter 1)
