Too Few Bridges
Silos aren’t a failure to fix. They’re what specialization looks like from the outside.
I’ve run hundreds of employee surveys over the last twelve years. Different companies, different industries, different vendors. One item shows up in the bottom five almost every time: cross-functional communication.
Every few years, in nearly every one of those organizations, someone declares war on it anyway. New reporting lines get drawn. A cross-functional initiative gets announced. Someone rolls out a collaboration tool nobody asked for. The walls, everyone is assured, are finally coming down.
A few years later, I run the survey again. Cross-functional communication is still sitting in the bottom five.
Why does the same problem survive its own solution, in company after company?
The easy answer is that the last attempt wasn’t serious enough, this time the reorg will actually stick. But an easier answer worth taking seriously: maybe the walls aren’t a failure to be corrected. Maybe something is rebuilding them on purpose.
I’m not the only one who’s noticed. Gillian Tett found something similar studying the institutions that fell hardest in 2008. The same structures that let people get exceptionally good at their work were the structures that kept the organization from seeing across itself. Not a coincidence sitting next to a competence. The same mechanism, doing both jobs at once.
That’s worth sitting with, because it means the standard diagnosis has the causality backwards. Silos don’t form because communication breaks down. Communication breaks down because silos formed first, and silos form because specialization works. Engineers cluster with engineers, accountants with accountants, nurses with nurses, because people solving the same kind of problem develop a shared language faster than people solving different problems ever will. Give that cluster enough time and it stops being a reporting line. It becomes a professional community, with its own standards, its own shorthand, its own sense of what good work looks like. What an outsider calls a silo is, from the inside, usually just a discipline.
So the war on silos has been fighting the wrong enemy the whole time. It isn’t bureaucratic residue. It’s the same mechanism that produced the expertise the organization is trying to keep.
Which is where it gets uncomfortable, because the fix everyone reaches for doesn’t distinguish between the two. Flatten the boundary hard enough and you don’t just improve the flow of information, you also flatten the thing that was generating deep knowledge on either side of it. Shared standards blur. Professional identity thins out. The org chart gets easier to draw and the organization gets harder to trust with anything specialized. Nobody actually wants a legal team that reasons like a product team, or a finance team that thinks like marketing. The boundary that frustrates collaboration is the same boundary that makes the expertise worth collaborating with in the first place.
So the organization is stuck wanting two things that appear to cancel out: deep enough specialization to be good at anything, and enough visibility across specialties to not be blindsided by what it can’t see. Tett’s finding is really about the second failure, not the first. Smart institutions didn’t miss the 2008 crisis because any single desk was incompetent. They missed it because no one occupied the space between desks.
That space has a name. Organizational researchers who study it call the people who work there boundary spanners, individuals whose value isn’t fluency in every discipline, but the ability to move an idea from one discipline into a form the next one can use. Not a manager forcing two departments into the same room. Not a reorg that erases the line between them. Something closer to a translator who lives on the border and answers to both sides.
Healthy organizations, it turns out, don’t get there by eliminating silos. They get there by staffing the space between them, deliberately, the way they’d staff any other function they depend on. Rotational assignments that send someone across the border and back. Embedded roles that sit inside one team while reporting into another. Communities of practice that pull specialists together around a shared problem without dissolving their home discipline. None of it requires anyone to stop being a specialist. It just requires someone, somewhere, whose job is to make sure the specialists aren’t the only ones who know what they know.
Which changes the question a leader should actually be asking. Not “how do we break down silos,” a question that treats specialization itself as the defect. Instead: where do ideas, decisions, and relationships need to cross a boundary, and who or what is making sure they can?
One question tries to erase the structure. The other assumes the structure earned its place and asks what it’s still missing.
Perhaps organizations don’t suffer from too many silos.
Perhaps they suffer from too few bridges.
Field Questions
- Where would cross-functional communication land if I measured it here, and have I actually asked?
- Is this boundary still protecting real expertise, or has it hardened into habit nobody questions anymore?
- Who is actually staffing the space between these teams right now, or is that job going undone?
- If I flattened this boundary tomorrow, what would we gain in coordination, and what would we quietly lose in depth?