Cross-team friction is nobody's fault

One item on an engagement survey I worked on read:

There is effective collaboration across teams in this organization.

That pairing is what got it attention. What I concluded from the work that followed is the part people argue with: cross-team friction is almost never anybody's fault, and that is exactly why goodwill does not fix it.

The two reasonable responses

When a leader runs into this, two responses follow, usually in this order.

The first is to call a meeting with your own team. It is the group you control, so it is the group you convene. The trouble with it is obvious the moment you say it out loud: the problem is across multiple teams, and you have just gathered the one team that cannot solve it alone.

The second response is better. You go and sit down with your counterpart on the other team. Two senior people, a candid conversation, real commitment on both sides, and an agreement to do better by each other.

And then nothing changes.

What fails there is the assumption underneath the conversation, which is that somebody is choosing this. Mostly nobody is.

The problem is typically not intentional. It's systematic and requires more than just force-of-will behavior change. It requires an accurate diagnosis of the root cause of the problem and structural fixes.

What structural means here

I want to be precise about that word, because it usually gets used to mean "cultural," which is a way of saying nothing.

A structure is the set of arrangements that make one behavior cheaper than another. How the two teams' goals are written. Who is measured on what. Who has to be consulted before a decision goes ahead. Where information sits and who can see it without asking. None of that responds to effort, and all of it responds to being changed. Two leaders can leave a good conversation genuinely committed and change nothing, because commitment was never the missing input.

What the comments said

The score told us we had a problem. It did not tell us what the problem was.

That came out of the free-text comments, and the themes were narrower than I expected. People were being pulled into work at the last minute and surprised by it. Teams were carrying goals that quietly competed with each other. My own summary of the cost at the time:

It made things so much more complicated and inefficient than it needed to be. It hindered innovation and slowed things down, and was a constant source of irritation.

Sorting the comments produced four failure modes, and nearly everything landed inside one of them.

Misaligned goals. Teams that do not share a destination cannot share a route.

Unclear support models. Nobody has written down what the two teams owe each other.

Ineffective communication. The work is coordinated and the information is not.

Sharing knowledge. What one team knows never reaches the other reliably.

There was a fifth thing, and it sat in the middle rather than alongside the others. Before taking a cross-team problem to anybody, check whether your own team practices the behavior it is about to ask for.

If your first instinct on reading that list was to work out which one the other team has, that instinct is the subject of this paragraph.

What we built, and what happened next

We built a diagnostic and an action playbook around the four modes, with a specific set of practices attached to each. I also ran quarterly sessions where managers could work through their own results: how to read them, what the pattern in front of them actually meant, and where to go next. More than 2,000 managers came to those voluntarily, every cycle.

Over the following cycles the item moved. So did the overall engagement index, by one point, from 76 to 77.

One point sounds like nothing. Against the external benchmark it was the difference between the 60th percentile and the 75th. (One point. Which says more about how engagement norms are distributed than it says about us, and it deserves a piece of its own rather than a parenthesis in this one.)

What I cannot tell you

I cannot tell you the playbook caused that.

We had no usage tracking. I do not know how many people opened it, how many teams ran it, or how much of the movement belongs to the sessions instead, or to the dozen other things a company that size runs over the same cycles. Anyone who claims they can isolate a single intervention inside an enterprise engagement program is selling you something.

What I can tell you is what the diagnosis was, what we built against it, how many people chose to show up, and what the numbers did afterward. Where that stops, I stop.

There is one more thing I have said out loud about this work before:

The biased part of me says that it was pretty comprehensive and that if you deployed it correctly there shouldn't be a problem, but I know that's not realistic.

It is not realistic. There are at least five ways to run this correctly and still fail. You diagnose the wrong cause, deciding it is a gap in awareness when it is a gap in time. You diagnose it correctly and cannot get leadership to change the goals. The real cause sits outside your control, in how the organization is designed. The solution is right and the rollout is wrong. Or the fix requires moving somebody off a team and nobody is willing to make that call.

Read that list again. Only the first one is a diagnostic failure. The other four are failures of authority, execution, or nerve, which is worth knowing before you start, because it tells you what to go and secure first.

If this is live for you now

The smallest useful version of all of this takes about a month.

Name the one failure mode costing your team the most. Then check that read with two people who work across the boundary every day, because leaders are wrong about which mode it is more often than they expect to be. Pick a single practice from it. Tell your counterpart on the other team what you are doing and why. Look again in thirty days.

One thing, run properly for a month, will do more than a workshop for everybody.