Us vs. Them Is a Survival Strategy: Why Agile Transformations Split Organisations in Two

The team that became a tribe

Eight months into the transformation, the team is delivering better than it ever has. They are also becoming increasingly convinced that the rest of the organisation is the enemy.

The team is a product squad at a fintech — technically excellent, measurably improving, and well-regarded by the coach who has been working with them since the start. In the last two months something has shifted. Every conversation that touches on stakeholders has the same flavour: "they don't understand value," "management just wants to control the process," "the business is stuck in the old way of thinking." The coach has been nodding. The team's frustrations are not baseless. Stakeholders are sometimes poorly aligned. Management questions have occasionally been clumsy. But each new instance of friction is being read as confirmation of an existing verdict rather than as new information. The category of "management" has stopped containing individuals. It has become a stance.

The coach notices, in their own body, a mild warmth when they are with the team and a mild guardedness when they are with stakeholders. They have not examined this. They have attributed it to appropriate professional alignment.

The coach had become part of the team's picture of the world. They had not noticed when it happened.

This is not a communication problem. It is not solved by a stakeholder alignment workshop or a cadence of transparency. The team has organised itself around a split — a division of the world into those who understand and those who threaten. The mechanism is involuntary, rational in the short term, and self-reinforcing in ways that make it progressively harder to correct. It has a name, a structure, and a resolution. The resolution is not a workshop.

Why groups split

A group undergoing transformation has limited psychological space to hold complexity. When the demands of change exceed what the group can process — when the uncertainty is too great, the losses too unacknowledged, the new requirements too disorienting — the group simplifies. It divides the world into safe and dangerous, trustworthy and threatening, us and them. This mechanism is called splitting. Van Beekum (2015), drawing on Klein and Bion: "dividing the world into good and bad, the bad comfortably projected onto the outside world, creates clarity, and with clarity comes comfort."

Splitting is not pathology. It is a developmental simplification — the same mechanism that allows a team in its early months to cohere around a shared identity before it has the complexity-tolerance to hold differences. The problem is not that it happens. The problem is when it solidifies from a phase into a permanent defensive structure, and the clarity it provides becomes the team's primary product.

Alongside splitting, a second mechanism operates: the team puts its own unwanted qualities — its rigidity, its fear of failure, its compliance with structures it claims to oppose — into the organisation and then relates to the organisation as if those qualities genuinely live there. This is projective identification. The mechanism is previewed here because it is inseparable from splitting; the full argument belongs to the next article in this series. For now: when a team describes management as rigid, inflexible, and resistant to new thinking, it is worth asking what the team has put there that it is not yet examining in itself.

When splitting takes hold, the team's energy shifts from its primary task toward protecting its group identity. Bion called this the basic assumption group in paranoid-schizoid mode — the state in which the group's organizing principle is survival rather than work. The ceremonies continue. The metrics are reported. But the group's real energy is going into boundary maintenance: confirming who is in, who is out, and who threatens the perimeter.

Three transformation conditions that produce splitting

Agile transformations create the specific conditions in which splitting is most likely. Three conditions recur.

The first is identity disruption. The transformation asks people to move from specialist to generalist, from expert to learner, from a directed team member to an autonomous owner. Each of these moves requires giving something up. The loss is real and rarely acknowledged. Splitting provides a narrative for the loss that places its source outside the self: it is not that the transition is genuinely difficult, it is that the organisation is obstructing it.

The second is loss without ceremony. Old roles, old working relationships, sometimes old reporting lines disappear without grief rituals. People are expected to be energised about a change they did not choose and may not have wanted. The unexpressed grief does not vanish. It becomes resentment with a narrative attached — a narrative that explains why the loss was someone else's doing.

The third is the promise and reality gap. The transformation was sold with autonomy as a benefit. The team discovers that genuine autonomy is available — and so is genuine accountability. The weight of real ownership was not in the sales pitch. The disillusionment is real, and it looks for an external cause. "Management gave us autonomy in name and kept control in practice" is frequently true. It is also frequently a way of not examining what the team is doing with the autonomy it does have.

Chosen trauma and chosen glory

Volkan, via Van Beekum, describes the patterns by which groups carry their histories as identity markers: the chosen trauma — a founding wound the group cannot put down — and the chosen glory — an exceptional past achievement the group's self-image requires to be protected from contamination.

In Agile contexts: the chosen trauma is the bad project, the arbitrary deadline that broke the team, the executive decision that killed something they built. The wound is real. The conclusion drawn from it — that constraints are attacks, that management cannot be trusted, that history will repeat — has become absolute. Every subsequent difficulty is filtered through it.

The chosen glory is the quarter when everything clicked, the product that shipped against the odds, the proof that this team operates at a level the organisation does not fully appreciate. This story must be protected. Stakeholder friction is experienced as an attack on the story, not as a project management problem. The team's identity depends on the story remaining intact.

Van Beekum (2015): "The projection of feelings of badness outside the organization then produces a state of illusory goodness and self-idealization." The team does not experience itself as splitting. It experiences itself as accurate.

What the coach gets pulled into

Splitting teams do not only create internal divisions. They recruit external figures into each pole. The good-object position — the one who understands, validates, and represents the team's perspective — is available and the coach is structurally vulnerable to filling it. The coach is present, invested, and by professional disposition inclined toward the team's experience. The recruitment happens quietly, through accumulation rather than through a single request.

Four signs indicate the coach has been recruited. The first is a feeling of warmth and loyalty toward the team paired with mild suspicion or guardedness toward stakeholders — experienced as natural professional alignment rather than as a position the team has assigned. The second is framing conversations with management as "representing the team's perspective" rather than as working with the organisation as a whole system. The third is finding the team's narrative self-evidently correct and not subjecting it to the same scrutiny applied to stakeholder positions. The fourth is feeling instinctively protective when the team's autonomy is questioned — not as a reasoned judgment but as a reflexive response.

The coach who has been recruited into the good-object position can no longer see the team whole. The split-off qualities the team has projected outward — its own rigidity, its own fear of failure, its compliance with the structures it claims to oppose — are invisible, because the coach has confirmed rather than questioned the projection. The coach's reading of the stakeholders is now partly the team's reading. The coach has become a mirror that only shows the team what the team already believes.

When the coach who holds the good end of the split ends the engagement, the team often splits again — sometimes onto the incoming coach, sometimes onto a newly available external target. The dependency was structural, not personal. The coach was not beloved for who they were. They were filling a role the split required someone to fill.

The coach's feeling of loyalty is itself an enactment

The warmth, the mild guardedness, the instinct to represent the team's perspective in leadership conversations — these are not independent professional judgments. They are enactments: the team's split recruiting the coach into a role the split requires. The question from that frame applies directly here: What role did the team need someone to play, and did I notice I was playing it?

Pre-recruitment is common. Coaches who have spent years in corporate environments before moving into development work often carry a personal version of the split the team is enacting: good work (coaching, learning, human-centred practice) on one side; bad work (budget decisions, management politics, organisational hierarchy) on the other. The team's narrative resonates not because it is accurate but because it matches something the coach brought into the room.

The politics of revelation

Van Beekum, drawing on Lawrence (2015), distinguishes two political orientations available to the practitioner working with a split system. The politics of salvation tries to remove the group's pain by solving the presenting problem: run the communication workshop, align the stakeholders, fix the friction. The politics of revelation makes the unconscious functioning visible — names what is happening without resolving it prematurely, and creates the conditions for the group to see itself.

Salvation is faster and feels more immediately useful. A stakeholder alignment session produces visible output. Revelation is slower and produces something more durable: a team that can see its own defensive structure and therefore has the option of relating to it differently.

For Agile coaches: running a communication workshop is salvation. Saying, I notice that when we talk about the product steering group, the language shifts — they become "them" rather than people with their own constraints. I'm curious what that's about — is revelation. The first move improves the surface. The second touches the mechanism.

Not confirming the narrative

The practitioner who consistently validates the team's good-object/bad-object division reinforces the split rather than working with it. Not confirming the narrative does not mean contradicting it. The team's frustrations are often legitimate. The coach can acknowledge that a stakeholder was clumsy, that a management decision was poorly communicated, that a constraint is genuinely frustrating — without accepting the global conclusion that all stakeholders are opponents and all management is hostile. The difference is between validating a specific experience and endorsing an absolute position.

Naming the split directly

When the split is sufficiently visible and the team has sufficient safety, naming it directly is available: I notice that when we talk about the product steering group, something changes in the room. They become a category rather than a set of people with their own constraints. I'm curious what we'd lose if we held them as people.

This is not an accusation. It is an observation offered for the team's examination. The team may reject it — may insist that the category is accurate, that the coach is not seeing what the team sees. The rejection is data. A split that can be named and rejected is closer to becoming visible than a split that has never been named at all.

Helping the team reclaim its projections

The diagnostic question for split teams: What have we put outside ourselves that we are not examining? This is not therapeutic exploration. It is organizational sense-making. A team that has projected its rigidity onto management can be asked: Where are we being rigid in ways we don't discuss? A team that has projected its fear of failure onto the organisation can be asked: What are we most afraid of getting wrong? The questions are not comfortable. They are useful.

Working with chosen traumas

The founding wound does not need to be processed. It needs to be named. When the team's split is anchored to a specific event — the product that was killed, the restructuring that cost them the colleague they trusted most — making that anchor conscious gives the team the option to carry it differently. It does not dissolve the history. It disaggregates the conclusion from the fact.

Depressive integration

The resolution of splitting is not the elimination of disagreement with stakeholders. It is the recovery of complexity — the team's capacity to hold management is sometimes wrong and management has legitimate constraints as simultaneously true; to be frustrated without defining identity through the frustration; to see individual stakeholders as people with their own histories rather than as representatives of a category. Van Beekum calls this depressive integration, borrowing from Klein. The plain translation: the team can see the whole picture again.

Moving from survival to work

Bion distinguished two modes in which groups operate. In work-group mode, the group's energy goes toward its primary task — the actual work it exists to do. In basic assumption mode, the group organises around an unconscious shared assumption that provides psychological safety at the cost of the task. For a team in paranoid-schizoid splitting mode, the basic assumption is: we are good, they are threatening, and our job is to protect the boundary. The primary task — building software, delivering value, doing the work — is not abandoned, but it becomes secondary to the boundary-maintenance function.

The shift from survival back to work is not dramatic. It does not require a resolution of the split or a reconciliation with stakeholders. It requires enough integration that the team's energy becomes available again for something other than the boundary.

Three markers indicate the shift is underway. The first is that stakeholders begin to be described with specificity rather than as a category. Martin is being rigid about the scope constraint rather than management doesn't understand us. The individual replaces the category. Complexity returns. The second is that the team can acknowledge its own contribution to friction without the acknowledgment collapsing into self-condemnation. We were unclear in how we presented the trade-off becomes sayable. It does not destabilise the team's sense of competence. The third is that curiosity replaces suspicion in cross-functional conversations. Stakeholder questions are examined for what they might be pointing at rather than for what threat they represent.

In the early months of a transformation, some degree of in-group solidarity is both natural and necessary. The team needs to cohere. The problem is not that it bonds against a shared challenge. The problem is when the bond calcifies — when the challenge becomes a permanent identity rather than a situation the team is moving through.

Timing matters. The coach who names the split too early, before the team has sufficient safety, risks being relocated from the good-object to the bad-object position — another voice that doesn't understand. The coach who names it too late risks having the naming absorbed into the split itself. There is a window. It is usually visible in the moments when individual team members step briefly outside the shared narrative — a moment of genuine curiosity about a stakeholder's position, an acknowledgment of the team's own part in a friction. These moments do not last. They indicate the split is not yet total, and the window is open.

For the coach: staying whole

The coach's own splitting risk is real and specific. Coaches who spent years in corporate or hierarchical environments before moving into development work often carry a personal version of the split the team is enacting: good work (coaching, learning, human practice) on one side; bad work (budget decisions, management politics, organisational constraint) on the other. The team's narrative confirms what the coach already knows. The confirmation feels like alignment. It is recruitment.

The supervision question for coaches working with splitting teams is not am I being appropriately supportive? It is: Toward whom am I feeling loyalty right now, and what am I not seeing about the other side? This is not a call for neutrality. The coach is not a neutral party and does not need to perform neutrality. It is a call for full perception. Loyalty that prevents seeing is not professional alignment. It is a defended position.

When the split has become visible enough to work with directly, the coach can name their own recruitment to the team: I've noticed I've been feeling quite protective of this team in conversations with the steering group. I'm curious what that's about — whether it belongs to me, or whether it's something the team needs someone to carry. This move requires the coach to have worked through their own history in supervision first. Named to the team without that work, it becomes a confession or a manipulation. Named with that work behind it, it is an invitation for the team to examine what it has needed the coach to hold.

The coach who stays whole — who can see the team's genuine strengths and its genuine defensiveness, who can see the stakeholders' legitimate constraints and their genuine limitations — is more useful to the system than the coach who has chosen a side. Not because neutrality is a virtue, but because the whole-system view is the one the team cannot yet hold for itself. That is precisely why the coach is there.

What the tribe is protecting against

The team did not become a tribe because it failed. It became a tribe because the transformation was genuinely overwhelming, the losses were real, and a tribe was the most available form of safety. The coach's job is not to dissolve the tribe. It is to help the team outgrow the need for it.

The split does not need to be eliminated to be worked with. It needs to be made visible. When the team can see the split as a response to a situation — rather than as the truth about the situation — it has already begun to integrate. The chosen trauma can be named. The projection can be questioned. The category of "management" can be disaggregated back into people with their own constraints, histories, and legitimate concerns.

Splitting operates at the group level. What operates at the individual level — what happens to specific people, including the coach, when the group deposits its split-off emotional material into them — is projective identification. That mechanism, and how to work with it, is the subject of the next article in this series.

The team's tribal narrative is not wrong. It is incomplete. Making it complete — helping the team hold the complexity it has had to simplify — is the relational coach's most demanding and most consequential contribution.