One of the things that surprised me most during my time at Google was how long simple things can take.

Not the big ambitious projects. Those were expected to be hard. I mean the small things. Updating a feature. Changing a default. Getting two teams to agree on a shared approach. Things that should have taken days somehow took months. And the frustrating part was that nobody was being difficult. Everyone was trying their best.

When organisations grow, the number of people who need to align on any decision grows with them. And that cost doesn’t scale in a straight line. It compounds. Two people agreeing on something is easy. Ten people across four teams, with different managers, different priorities, and different definitions of success, is an entirely different problem. Nobody designed it that way. It just emerges.

What makes it so hard to address is that the cost stays invisible. Each individual interaction seems minor. One more email. One more meeting. One more person who needs to be consulted before anything can move. None of it feels significant on its own. Together, it quietly makes everything slower and more expensive than it should be.

The thing I saw work most consistently at Google was keeping teams as small and focused as possible for as long as possible. Not because small teams are inherently better, but because every person you add to a problem increases the coordination cost in ways that are easy to underestimate and hard to undo.

The instinct when things slow down is to add more people, more process, more oversight. This was known as “coordination headwind” internally at Google. In my experience, that usually made things worse. The better instinct is to ask what you can strip away.