The silly creatures are at it again. They have merged two companies and expected the codebases to fall in love. They expected the databases to hold hands and the APIs to understand each other because the marketing department said “synergy.”
They are wrong.
Welcome to Mergers, Acquisitions, and the Art of the Unplanned API.
Conway’s Law is the ghost in the machine#
You cannot merge two organizations and keep the old architecture. It does not work that way. The system’s architecture reflects the communication patterns of the organization that built it. Structure is not an accident. It is a direct output of how teams interact.
When you merge two companies, you are not just adding headcount. You are adding communication bottlenecks. You are adding silos. You are adding new reporting lines and new political boundaries. And the architecture follows. The teams design components around the boundaries they already use to coordinate work and fast flow. If the communication is a mess, the modules will be a mess. The data flows will be a mess. The interfaces will be a mess.
Technical redesign without organizational redesign produces drift, fragmentation, and integration pain. You cannot just refactor the code. You have to refactor the conversation. If you want a different architecture, you must first change the team structure and communication pathways.
The Unplanned API#
In a merger, the “Unplanned API” is the interface between the two new entities. It is the messy, undocumented, high-bandwidth zone where the “us” meets the “them.”
This is where Collaboration Mode lives.
Collaboration is a short-term, high-bandwidth interaction used when teams need to jointly explore, design, or solve a complex problem. It is intentionally temporary. It is expensive. It increases cognitive load and coordination cost. But it is necessary to resolve uncertainty.
The problem? The silly creatures do not want it to be temporary.
They want to stay in Collaboration Mode forever. They want to sit in meetings for six hours to decide what colour the button should be. They want to “align” on every technical decision. They think alignment is the goal.
Alignment is the starting line.
If long-term collaboration becomes a coordination sink, you have failed. Used sparingly, it accelerates breakthroughs and reduces long-term friction. Used permanently, it kills velocity. It kills morale. It kills the thing you bought the company for.
The Steward and the Conductor#
This is where the leaders come in. They are not here to write code. They are not here to debug the API.
They are here to be the Steward. They are the translation layer. They help peers and executives understand what the organization is doing and why it matters through clear messaging. Without this, the teams work at cross-purposes and progress stalls.
They are here to be the Conductor. They make the music without playing an instrument. They provide clarity to remove organizational anxiety. They articulate a clear direction early to align the organization and reduce confusion.
If the leader cannot articulate the “why,” the team will fall back on the “how” of the old company. And the old company is gone. The old company was a different alien species.
The Megaproject Trap#
Mergers are often treated as megaprojects. And megaprojects fail. Not because they are inherently impossible, but because they are executed inside a system that structurally favours failure and attrition.
Political incentives reward starting megaprojects, not finishing them. Leaders gain prestige from announcing the merger. Contractors profit from overruns. Accountability is diffuse. Inexperienced teams are selected for political reasons.
The result? Unrealistic estimates. Locked-in commitments. Irreversible execution. Escalating failure. Sunk-cost continuation.
The “Unplanned API” becomes a black hole. Teams get stuck in high-bandwidth collaboration. Leaders get stuck in political signaling. The architecture fragments.
The intervention#
- Define the boundaries. Use Conway’s Law. If you want a modular architecture, create modular teams. Don’t expect the architecture to change if the org chart stays a pyramid of confusion.
- Use Collaboration, then drop it. Solve the uncertainty. Build the shared understanding. Then get the hell out of the way. Return to lower-cost interaction modes.
- Lead with clarity. The leader’s job is to reduce anxiety and provide direction. Not to be in every meeting. Not to be in every decision.
- Stop the megaproject illusion. Acknowledge that the merger is a system with biased estimates and political pressures. Plan for the drift. Plan for the fragmentation. Don’t pretend it’s a clean handover.
The silly creatures think the API is technical. It’s not. It’s organizational.
If you fix the org chart, the API fixes itself. If you don’t, you’ll be in Collaboration Mode for the next two years.