The org chart ends up in the product
How we were organized shaped what we shipped, for better and worse.
There is a line from Melvin Conway that has aged better than almost anything else written about software: organizations ship their communication structure. Whatever shape your teams are in, the product comes out the same shape. I used to treat that as a curiosity. After building from zero, I treat it as one of the most important product decisions a leader makes, and one of the least examined.
You can read the org chart off the product
Open almost any product and you can guess how the company is organized. The places where two features do not quite talk to each other are usually the places where two teams do not quite talk to each other. The screen that feels like three different products stitched together is three teams that own one screen and never agreed on it. The handoff that drops things is an ownership gap nobody put their name on.
This is not a moral failing of the teams. It is physics. People build what they can coordinate, and they coordinate within their own boundaries far more easily than across them. So the boundaries you draw between teams become the seams a customer feels.
Seams are inevitable, so place them on purpose
The mistake is trying to eliminate seams. You cannot. Any non-trivial product has them. The job is to decide where they go.
A good seam runs along a line the customer already understands, and where the two sides genuinely can move independently. A bad seam cuts straight through a single user journey, so that completing one obvious task requires three teams to agree on something every quarter.
When I think about team design now, I start from the journeys I most want to feel seamless, and I refuse to draw a team boundary across the middle of one of them. Everything else can flex. The flows that have to feel like one thing get owned by one team that can actually make them one thing.
Founding teams encode their values into the product
There is a second, quieter version of this. Early teams do not just shape the structure of the product. They shape its character. The things the first handful of people care about become the things the product is good at, long after those people have moved on.
As part of a founding team, that responsibility lands hard. If the early group cares about correctness, the product is trustworthy in a way that is hard to retrofit later. If the early group cares about speed over everything, the product is fast and a little careless, and that, too, is hard to retrofit. You are not only deciding what to build. You are deciding what the company will find easy and what it will find hard for years.
Reorgs are product launches
Because structure ships, a reorg is not an HR event. It is a product decision with a long lead time and no undo button you can press cheaply.
I have learned to treat changing team boundaries with the same seriousness as changing the product, because that is what it is. Before moving a line, I ask what seam this creates, which journey it cuts, and whether the thing I am optimising for, usually speed or focus, is worth the new coordination cost I am about to introduce. Sometimes it clearly is. The point is to make the trade on purpose rather than discover it in the product six months later.
What I do with this now
A few habits that come out of taking Conway seriously.
Design the team around the experience you want, not the technology you have. The architecture will follow the teams anyway, so aim the teams at the customer.
Give every important journey a single owner who can be named. If you cannot name who owns the end to end flow, the customer already knows, because they can feel the gap.
Treat the first ten hires as a product spec. Their instincts become the product’s instincts.
And when something in the product feels disjointed, look at the org chart before you look at the design. The fix is often a line in a structure, not a screen.
The short version
Your product will look like your org chart whether you intend it to or not. The only choice you have is whether the resemblance is deliberate. The teams I am proudest of were not the ones with the cleanest structure on paper. They were the ones where the seams landed exactly where we meant them to, and the customer never had to think about how we were arranged.
Draft. Builds on Conway’s law and a few years of getting the seams wrong before getting them right. Yours to edit.