How Structure Shapes Communication

The last page was honest about politics — the human behaviour that fills the gaps structure leaves. This closing page of the part does something that ties the whole thing together: it shows that the structure itself, before any politics, is already deciding what gets communicated, to whom, and where the message dies. The org chart isn’t just a picture of who reports to whom. It’s a map of how information is allowed to move.

Here’s the promise. By the end of this page you’ll have a diagnostic instinct that will serve you for the rest of your career: when communication breaks, look at the structure first. Most “communication problems” are not really about people being bad at talking. They’re about a message being asked to cross a boundary the org chart never built a bridge for. Once you can see that, a huge amount of frustrating, blame-filled confusion turns into something you can actually locate — and route around.

The principle: what you build mirrors how you’re organized

In 1967 a computer scientist named Melvin Conway noticed something odd and durable. He wrote, roughly, that any organization that designs a system will produce a design whose structure copies the organization’s own communication structure. It became known as Conway’s Law, and although he was describing software, the insight is far broader and applies to any organization that makes anything — a product, a report, a patient’s care, a customer’s experience.

Strip it to plain terms: the thing you produce ends up shaped like the group that produced it. If four separate teams build a product, the product will tend to have four separate-feeling parts, with awkward seams where the teams had to hand off to each other. If a hospital’s care is split across departments that don’t talk, the patient’s experience will feel split across departments that don’t talk — appointments that don’t connect, a story retold five times. The output inherits the shape of the org chart because the communication that made it inherited the shape of the org chart.

That’s the deep version. There’s a simpler, everyday version that matters just as much for you: the same is true of messages. A message travels well exactly where two people are organizationally close — same team, same manager, a real reason to talk — and travels badly, or not at all, where the chart puts a wall between them. The org chart is a communication map. It just doesn’t advertise itself as one.

The evidence — Conway’s Law is more than a slogan

Conway’s original 1967 observation sat as a memorable maxim for decades, but researchers have since found real support for it. Studies of large software projects have generally found that the boundaries in a product line up with the boundaries between the teams that built it, and that components built by groups who communicate poorly tend to have more defects at their seams. The exact figures vary by study and aren’t worth memorizing. The robust finding is the direction: how a group is wired to communicate shows up, measurably, in what it makes. You cannot fully separate the shape of the output from the shape of the org.

Team boundaries are where information gets lost

If output mirrors structure, then the most important lines on the org chart are not the boxes — they’re the lines between the boxes. Every boundary between two teams is a place where information has to be deliberately carried across, and anything that has to be deliberately carried is something that can be dropped, delayed, or distorted.

Think about how communication actually works inside a team versus across one:

INSIDE a team                    ACROSS a boundary
- shared context, shared goals   - different context, different goals
- overhear each other            - must be explicitly told
- ask a quick question           - schedule a meeting / file a request
- notice when something's off    - only learn when it's already a problem
- one manager, one priority      - two managers, two priorities

Inside a team, communication is nearly free — it happens in the air, exactly like the five-person cafe from the start of this part. Across a boundary, every one of those free mechanisms disappears. Nobody overhears; nobody notices; the quick question becomes a meeting; the shared goal becomes two competing goals. So information doesn’t just travel slower across a boundary — it travels worse. It arrives late, incomplete, or subtly warped to fit the receiving team’s assumptions.

This is why the phrase “it fell through the cracks” is so common and so revealing. The cracks are real, and they’re always in the same place: the cracks are the boundaries between teams. Nothing falls through the middle of a team. Things fall through the gaps the structure created.

In practice — the handoff that quietly rotted

A warehouse has a receiving team and a stocking team. Receiving logs what arrives; stocking puts it on shelves. Between them: a boundary. One week a supplier changes its box sizes. Receiving notices — it’s obvious to them, boxes are their whole world — but there’s no channel that carries “box sizes changed” across to stocking, because that fact never mattered inside receiving’s own goals. Stocking discovers it the hard way when a pallet won’t fit the racks, three days and one jammed aisle later. Nobody was lazy or careless. The information existed, fully known, ten metres away. It simply had no bridge across the boundary — and the org chart never built one because the two teams answer to different managers with different priorities.

Why cross-functional work fights the grain

Now you can see why cross-functional work — work that requires several different teams to move together — is so reliably painful. It isn’t that the people are difficult. It’s that cross-functional work asks information to flow along paths the structure actively discourages. It fights the grain of the org chart.

Picture the chart as wood grain. Communication runs easily along the grain — up and down your own reporting line, sideways within your own team — because that’s the direction the structure is built to carry it. Cross-functional work asks messages to run across the grain, boundary after boundary, each one a place where context is lost and priorities collide. The work is hard not because it’s complex in itself, but because it’s swimming against the current the structure sets up.

Everything you learned earlier in this part shows up here as a specific kind of friction on that path:

• Layers slow messages down. In a tall hierarchy, for two frontline people on different teams to align, the message may have to climb up one reporting line to a common boss and back down another. Every layer adds delay and a chance for the message to be re-interpreted. A message that climbs five layers and descends five is barely recognisable when it lands — the corporate version of a game of telephone.
• Silos block messages entirely. A divisional or heavily functional structure creates walls with no doorway. Two teams with different goals and no shared manager have no natural reason — and often no channel — to tell each other things. The information doesn’t slow down; it simply never leaves the room.
• Centralization funnels messages through a bottleneck. When decisions live at the top, communication has to route through the few people authorised to decide. Those people become a chokepoint: everything waits on their attention, and whatever they don’t have time to read, in effect, didn’t get communicated.

So “we have a communication problem” is almost never the whole story. The useful question is which structural feature is doing the damage: too many layers (slow), too many silos (blocked), or too much centralization (funnelled). Each has a different fix, and you can’t choose the fix until you’ve named the shape.

The common trap — treating a structural problem as a people problem

When a cross-team effort keeps stalling, the instinct is to blame people: “they don’t communicate,” “that team is a black box,” “we just need better alignment.” So the response is more meetings, sterner emails, a new group chat, a workshop on “communication skills.” These rarely fix it, because the problem was never the people’s willingness to talk — it was that the structure gave them no natural channel and two conflicting sets of goals. Piling communication effort on top of a structural gap is like bailing a boat instead of patching the hole. The tell: if the same handoff keeps failing no matter who is staffed on it, you’re looking at a structural problem wearing a people costume.

The diagnostic instinct: read the structure first

Here’s the single most useful habit to take from this page. When you hit a communication breakdown — a decision nobody can get, a fact that never reached the right people, two teams building slightly incompatible things — resist the urge to ask “who dropped the ball?” first. Ask instead: “what boundary did this information have to cross, and was there ever a bridge across it?”

That reframes a blame question into a design question, and design questions are solvable. Walk it through:

The breakdown                    Look here first
"nobody told us"          →      a team boundary with no channel across it (a silo)
"it took forever"         →      too many layers between the people who needed to align
"we're still waiting on   →      a centralized bottleneck: one over-subscribed
 a decision"                     decision-maker everything routes through
"we built the wrong       →      two teams, two contexts, a handoff that
 thing"                          distorted the message in transit

Notice that each symptom points at a specific structural feature you already understand from this part. That’s the payoff of reading the whole thing as one argument: communication isn’t a separate topic bolted on at the end. It’s the thing that flows through everything else — so every structural choice you studied is also, quietly, a decision about what gets said and what gets stuck.

Case study — the 1986 Challenger launch

On January 28, 1986, the Space Shuttle Challenger broke apart shortly after launch, killing all seven aboard. The engineers who built the affected part had genuine, specific concern about launching in the unusual cold — the danger was, in a real sense, known inside the organization. But that concern had to cross several boundaries and climb several layers to reach the people making the launch decision, and along that path it was softened, reframed, and pressured until what arrived at the top was not the alarm that had left the bottom. The later investigation pointed hard at communication and organizational structure, not merely at the physical part. The lesson isn’t about rockets. It’s the sharpest possible version of this page: when a message has to cross too many boundaries and climb too many layers, the most important information can be fully known by the organization and still never truly arrive where the decision is made. Hedge the details as you like; the structural lesson is not in dispute.

Routing around the friction — ethically

None of this is a counsel of despair. You usually can’t redraw the org chart — that’s not your job, and reorganizing has costs of its own. But once you can see the structural friction, you can route around it deliberately and honestly. This is the constructive flip side of the diagnostic instinct, and it’s where this part hands off to the rest of the book.

A few moves, all of them clean:

• Build the missing bridge yourself. If a handoff keeps failing at a boundary, create the tiny channel the structure never built — a standing five-minute check-in with your counterpart on the other team, a shared note, a single named person on each side who owns the seam. You’re not going around anyone; you’re adding the doorway the wall was missing.
• Shorten the path. If a message is dying in the layers, find the shortest legitimate route — sometimes a direct, transparent conversation with the person two teams over (with your manager in the loop, not behind their back) beats sending it climbing up and down two reporting lines.
• Make the invisible boundary visible. Often the fix is simply naming the seam out loud: “This work crosses three teams and nobody owns the middle — can we agree who does?” You’re not accusing anyone; you’re pointing at the structural gap so the group can close it on purpose.
• Carry context across, not just facts. Because boundaries strip context, the highest-value thing you can send across one is the why, not only the what. The receiving team has different goals; a bare fact will be re-interpreted to fit them, but a fact plus its reason survives the crossing.

The line between routing around friction and playing politics is the same line the previous page drew: do it in the open, in service of the shared goal, and in a way you’d be comfortable explaining to everyone involved. Building a missing bridge is service. Quietly withholding information to gain an edge is the opposite. The structure creates the gaps; your integrity decides what you do in them.

How it connects

This page is the hinge between how organizations work and how you communicate inside them. It closes this part by showing that structure — the chart, the layers, the silos, the bottlenecks — is really a communication map in disguise. And it opens directly onto the Communication part, which is the craft of moving a message well despite the structure, and onto Managing Up, much of which is the skill of getting information to cross the boundary between you and the people above you before it gets softened into nothing. Read the machine first; then you’ll know exactly where your communication has to work hardest.

Try this

This week, pick one recurring “communication problem” you’ve been treating as a people problem — a handoff that keeps failing, a decision you can never get, a team that feels like a black box. Draw the two (or more) parties as boxes and mark the boundary the information has to cross. Then answer one question: is there an actual channel across that boundary, or has everyone just been hoping the message would find its own way? If there’s no channel, sketch the smallest bridge you could build yourself — one check-in, one shared note, one named owner of the seam — and propose it in the open this week.

Reflect

1. Think of something your organization produced that felt disjointed — a clunky product, a fragmented customer or patient experience, a confusing report. Can you see the team boundaries showing through in the seams of the output?
2. Where does information reliably get lost around you? Name the specific boundary it has to cross — and whether that boundary has any real channel across it.
3. Recall a “communication problem” that got blamed on people. On reflection, how much of it was actually structural — layers, silos, or a bottleneck — wearing a people costume?
4. Of the three structural frictions (layers slowing, silos blocking, centralization funnelling), which one does the most damage where you work? What would the specific fix for that one be?
5. Where could you ethically build a missing bridge — a small, open channel across a boundary the org chart never bridged? What’s stopping you from proposing it?

Show reflections

1. This is Conway’s Law in your own life. The point is to see the mirror: disjointed output almost always traces back to teams that didn’t (or couldn’t) communicate across a boundary. If you can spot the seam in the product, you’ve found the seam in the org — they’re the same line.
2. The useful discipline is to get specific about the boundary, not the people. “Sales and delivery don’t talk” is a structural diagnosis; “that team is bad at their job” is a dead end. Naming whether a real channel exists tells you immediately whether this is fixable with a bridge or needs something bigger.
3. Most people find far more of it was structural than they’d assumed — the same handoff failing regardless of who’s staffed on it is the giveaway. The value here isn’t to excuse everyone; it’s to stop spending energy on blame and redirect it at the gap, which is the only thing you can actually change.
4. Each friction has a different fix: layers → shorten the path or push the decision down; silos → build a channel or a shared goal across the wall; centralization → unblock or work around the bottleneck. Naming which one dominates is what turns “we should communicate better” (useless) into a concrete, targeted move.
5. This tests whether you’ve turned the diagnostic instinct into action. A good answer names a real seam and a genuinely small bridge — and is honest that what usually stops us isn’t difficulty but the fear of overstepping. Doing it in the open, in service of the shared goal, is exactly what keeps a bridge from becoming politics.