Two teams that have to work together and do not

Nobody in either team would describe themselves as uncooperative. Ask them separately and you get two internally consistent accounts: the other team is slow, or unrealistic, or does not understand the constraints we work under. Both accounts are honest. Both are describing the same structure from opposite ends.

This is why collaboration workshops that ask people to communicate better tend to change nothing. The teams are already communicating. What they are not doing is seeing the shape of the thing they are both inside.

The three failures that get called a silo

They need different responses, and lumping them together is why one workshop rarely fixes anything.

Information that does not move. Each side holds something the other needs and neither knows it. Nobody is withholding; the handover point simply does not exist. This is a process problem wearing a cultural costume.

Incentives that genuinely conflict. Sales is paid on closing, delivery is measured on margin, and the friction between them is not a misunderstanding — it is the compensation structure working as designed. No amount of goodwill fixes this, and a workshop that implies otherwise insults everyone in the room.

Accumulated history. Something went wrong two years ago, somebody was blamed, and the relationship has been transactional since. Both sides can name the incident. Neither has said it out loud in a mixed room.

The first is fixable in a session. The second needs the sponsor to change something and the session's job is to make that undeniable. The third needs the incident surfaced before anything else will land, which a well-run debrief does and a facilitated discussion usually does not.

What produces the shift

The mechanism that works is structural, not motivational: build a situation where no group can finish alone, and then do not tell them that.

What follows is reliable enough to plan the debrief around. Each team works its own piece in isolation for the first stretch, usually well. Then progress stops. Somebody crosses the room — and who that is, and how long it took, is the most useful observation of the day. Information gets exchanged badly at first, because each team explains its piece in its own vocabulary and the other hears noise. Eventually someone translates rather than repeats, and the thing unlocks.

That sequence is a compressed version of the working relationship, and everyone in the room has just lived it rather than been told about it. The debrief question is then simple and unanswerable in a defensive way: how long did it take us to go and ask, and what does that cost us on a Tuesday?

The four observations worth writing down

A facilitator running this session should come out with specific, timestamped notes. Generic ones produce a generic debrief. These four are the ones that reliably carry an hour of conversation.

Who crossed the room first, and at what minute. Not who was most senior, and not who was loudest — who decided that asking was better than continuing. In most sessions this person is the same one who does it at work, and naming that in front of the room is worth more than any model.

How long the group worked in parallel before anyone checked alignment. Teams routinely spend a third of the available time building things that turn out to be incompatible. The number is usually shocking when stated aloud.

The moment translation replaced repetition. The first attempt at exchanging information almost always fails, because each side explains in its own vocabulary. Someone eventually rephrases in the other side's terms. That switch is the skill the session is actually about.

What got abandoned silently. A plan that stops without anyone saying it has stopped is the single most transferable pattern in the room, because it happens constantly at work and nobody ever notices it happening.

What the session cannot do

Be honest with the sponsor about this before booking, because the disappointment otherwise lands on the session.

It will not fix conflicting incentives. It will make them visible and specific, in front of the person who can change them, which is usually the most valuable thing available — but the change itself happens afterwards or not at all.

It will not survive a missing decision-maker. Two teams reaching a shared conclusion without the person who arbitrates between them get to have the same conversation again next quarter.

It will not repair a relationship where one side genuinely acted badly and it has never been acknowledged. Surface that first, separately.

Who should be in the room

Both teams, in the same room, mixed at the tables. This sounds obvious and is routinely got wrong — sessions are booked team by team for scheduling reasons, which reliably produces two groups that are now each more articulate about the other's failings.

Include the handover points specifically: the people who actually pass work between the teams, not only their managers. They hold the detail, and they are usually the ones who have quietly built workarounds nobody has documented.

Around twelve to thirty people, if a single debrief conversation is going to matter. Above that, plan for table-level reflection and reports.

What to do the week after

The session produces a shared reference and a short list of specific frictions. Both decay quickly unless something is put in place before everyone leaves the room.

Two things are worth committing to on the day, in writing, in front of everyone:

One handover that changes shape. Not a principle — a named point where work passes between the teams, with an agreed response time and a named owner on each side. The session will have produced three or four candidates; pick the one both teams groaned at.

A date to look at it again. Ninety days, in the calendar before anyone leaves. Not another workshop; twenty minutes with the same people asking whether the handover actually changed.

Anything more ambitious than this tends not to survive contact with the following Monday, and a small change that holds is worth more to the relationship than a long list that quietly lapses.

Where HDX fits

Flow is built exactly on the mechanism above: three teams, separate tables, one locked room, sixty minutes, and no team holding enough of the puzzle to finish alone. The physical materials matter here — boxes, padlocks, hidden ink — because a team can see another team is stuck without being able to see why, which is the whole problem in miniature.

At larger scale, or where the relationship is between more than three groups, Falcon Eye runs the same dynamic as trade between villages: reputation accumulates across rounds, villages remember who honoured a deal, and bluffing turns out to be expensive.

Two things worth reading: how to debrief an experiential session, because on a collaboration session the debrief is the entire intervention, and when a physical game beats a digital one, because for interpersonal behaviour the format choice is not neutral.

---

If two teams in your organisation each believe the other is the obstacle, tell us what the room needs to leave with.