Two of my people are in open conflict and I've been pulled in.

What to do when two of your people are in open conflict

February two thousand and fourteen. Satya Nadella has been chief executive of Microsoft for less than a month.

What he has inherited is, by most external measures, a healthy company. A hundred and twenty-three thousand employees. Profitable across nearly every quarter of the previous decade. The stock has been almost exactly flat across that decade.

What he has also inherited is a performance system that does precise damage. Stack ranking. Every team, every cycle, has to identify its lowest-performing members for removal. The system does what it was designed to do. Engineers hide information from each other. Teams withhold help from adjacent teams. The internal collaboration the company's product portfolio increasingly depends on has been engineered, by the appraisal process, into a contest.

In his first year, Nadella abolishes the system.

He reframes the culture in language that sounds like a slogan until you read it as a behavioural instruction. Learn-it-all, not know-it-all. People who could win the old system by being the most certain person in the room get rewarded for something else now. People who could survive the old system by hoarding information get rewarded for something else now.

Microsoft's market cap in February two thousand and fourteen is roughly three hundred billion dollars. By the end of two thousand and twenty-four, it sits above three trillion.

The Nadella decade is mostly the same people. The same engineers, the same product managers, the same teams. Roughly the same skills, the same intelligence, the same ambition.

What changed is what their environment was rewarding them for.

When two people on a team are in open conflict, the manager's instinct is to look at the people. The Microsoft answer is that the people are usually the wrong place to look.

So let's go to the office and work through it.

"Two of my people are in open conflict and I've been pulled in."

The feeling is squeezed.

You have been brought in by one of them, or both. Two people you respect, who have come to you one at a time, each carrying a version of the story in which the other one is the problem. The day already had work in it.

First decide whether the fight is about them or about boredom

Two choices. They look identical from outside the meeting. They need different moves.

When two personalities have built their own momentum

Choice one: the fight is about them. Two personalities that have rubbed each other wrong, and the rubbing has become its own momentum. The original cause is, by now, almost beside the point.

If that's the trap, redesign their immediate workflow so that neither can close out their own work without the other's sign-off. Not as a punishment. As a forcing function. The current environment lets them avoid each other; a changed environment makes avoiding each other more expensive than getting on with it. Don Norman wrote about environmental affordances — change the affordance and the behaviour follows. Most interpersonal fights shrink once the cost of continuing them is higher than the cost of getting over them.

When the friction is really about boredom

Choice two: they have got no one else to fight. The energy is not really about each other. It is energy generated by boredom, internal focus, or the absence of a meaningful external adversary, and the two of them happen to be in each other's line of sight.

If that's the trap, give them a shared enemy. Joint responsibility for a high-stakes external task — defending a position against another department, hitting a deadline against a competitor, holding a line with a difficult client. Muzafer Sherif's Robbers Cave experiment, in nineteen fifty-four, showed that a bond formed under external pressure almost always overwrites friction generated under internal pressure. It is not a subtle move. It works because humans are wired to take the threat that's closest.

About-them, or no-one-else-to-fight. Same two people at the same desks, two completely different first moves.

How to make it about structure, not personalities

Three tools. Move the frame from personality to structure.

The first is

Read the nine levers shaping what they actually do

MINDSPACE.

Paul Dolan, Michael Hallsworth, David Halpern, Dominic King and Ivo Vlaev published MINDSPACE in two thousand and ten, working out of the UK Cabinet Office and the Institute for Government — MINDSPACE: Influencing behaviour through public policy. Behavioural-economics distilled into an operational framework public servants could carry into a meeting.

The reason the tool exists for this scenario is that the manager's first instinct in a two-person fight is to talk to each person about their feelings. The feelings are downstream. The behaviours that surround the people are the lever you can actually move.

The unique insight is that nine specific levers shape what people do, almost regardless of what they would say if you asked them. The acronym names them. Messenger. Incentives. Norms. Defaults. Salience. Priming. Affect. Commitments. Ego. In any given fight, one or two are live. Read the room for which.

What you get is a diagnosis you can run silently. You don't ask either person what's wrong. You read the environment for which lever's pulled.

So. The nine, with the question that surfaces each one.

Messenger. Who is delivering each piece of news to the other? Sometimes the fight is the channel, not the content.

Incentives. What is each one being rewarded for, and do those rewards point them in opposite directions?

Norms. What is everyone else on the team doing? Has the unspoken rule of the team become to fight?

Defaults. What is each one's default mode under pressure, and is one being read as hostile when it isn't?

Salience. What is artificially in their line of sight — the layout, the calendar, the org chart — making the other one impossible to ignore?

Priming. What just happened, in the meeting before this meeting, that they walked in already cued for?

Affect. What feelings are attached to one side of the argument that have nothing to do with the argument?

Commitments. What has each one already publicly said, that they now can't quietly walk back?

Ego. Where is each one trying to protect a sense of themselves, and is the fight serving that protection?

Find the live one. Move that lever. The fight follows.

The second is

Write down the unspoken rule they're really fighting over

Working Agreements.

Esther Derby and Diana Larsen formalised Working Agreements as a team ritual in Agile Retrospectives: Making Good Teams Great in two thousand and six. The lineage is older — group-norms work traces back through Edgar Schein's Process Consultation in nineteen sixty-nine and Kurt Lewin's group dynamics in the nineteen-forties — but Derby and Larsen put the ritual on a page that an agile team could run on a Tuesday morning.

The reason the tool exists is that most interpersonal conflict between capable adults is an argument about rules they never agreed on in the first place. Whose review blocks whose merge. Whether you can interrupt someone in flow. What counts as committed. The unwritten rule is the one each person has assumed everyone else shares.

The unique insight is that the unspoken norm is the source of the friction. Once it is documented, the friction has nothing to push against. The rule becomes a third thing in the room — neither person owns it, both can point at it.

What you get is a team that can argue with the agreement instead of arguing with each other. The conversation moves up one level, and that level is much easier to manage.

So. How to run it.

Setup. A working session, ninety minutes, the whole team. Not just the two. The agreement is for everyone; the fight you are solving is one symptom of a wider gap.

Surface. Each person, on a card, names the moments in the last quarter where they didn't know what the rule was. I didn't know if I could ship without a review. I didn't know if I should respond at the weekend. No accusations. Just gaps.

Agree. Cluster the cards. For each cluster, the team writes the rule that would have made the gap not a gap. Plain language. One sentence per rule.

Post. The agreement goes somewhere visible. Not in a folder. The team's working surface — the channel header, the team page, the wall. Visibility is half the discipline.

Revisit. Every quarter. Rules that are not being followed get rewritten or removed. A Working Agreement that nobody breaks isn't a rule, it is wallpaper.

The third is

Re-anchor the argument to the customer's experience

Customer Journey Mapping.

The lineage runs through Jan Carlzon's Moments of Truth in nineteen eighty-seven, when the chief executive of Scandinavian Airlines reframed the company around the fifty million ten-second interactions a year between the airline and its customers. Service designers in the two-thousands — Adaptive Path, Live Work, the early service-design studios — turned Carlzon's idea into the visual artefact: the journey rendered as a strip across a page. The Nielsen Norman Group's contemporary writing is the canonical practitioner reference.

The reason the tool exists, when two of your people are fighting, is that two people arguing about each other's positions will keep arguing for as long as the argument is about themselves. Force them to look at the same external reality together — the customer's experience — and the argument loses its anchor.

The unique insight is that the map is a shared object. They can disagree about the customer; what they cannot do is keep disagreeing about each other once the customer is on the table. The frame has shifted, and the shift is most of the work.

What you get is a fight re-anchored to the work. Whatever they were really arguing about, the map gives them a reason to stop.

So. How to walk it.

Awareness. They don't know they need you yet. Where does the customer first encounter you — and what do they not yet know? Two people from different parts of the company will see different things at this stage. Both go on the map.

Consider. Weighing options — doubt creeps in. What are the customer's hesitations? What are they comparing you against?

Decide. The moment of commitment — high stakes. What enables the decision? What blocks it?

Use. Do they get the value they paid for? Where does the experience deliver — and where does it break down? The friction marks go here. Most of them, you'll find, sit nowhere near the boundary the two of them were arguing across.

Advocacy. Delight — they bring the next. What makes a customer recommend? Most journeys never reach this stage.

One continuous experience — emotion shifts along the path, not in the touchpoints. Pick one friction and assign it together. Both names on the fix, jointly accountable. The fight isn't resolved. The fight is now about something both of them want to be on the right side of.

That's the toolkit. One more story before we close.

The Nadella story we opened with is the move at scale — change the institution that is making them fight. The story we close with is the same move at the scale of two desks and a small conference room.

A precedent: changing what they fight over, not mediating it

Mountain View. Two thousand and ten. Kim Scott is managing a team at Google.

Two of her direct reports — pseudonymised in Radical Candor as Sarah and Jim — are in open conflict. Each of them comes to her on their own. Each of them carries complaints about the other. Each of them has evidence. Each of them expects her, as their manager, to listen carefully and take their side.

The complaints are not trivial. Each believes the other is undermining their work. Each is using her one-on-ones to rehearse the version of events in which they are the reasonable party.

What Scott does, described in Radical Candor, is bring both of them into a small conference room and tell them she will not leave the room until they have agreed on a shared goal.

She refuses to mediate their versions of events. She refuses to let the meeting be about who was right about which specific grievance. She does not litigate the past, because litigation of the past is the format the triangulation requires to keep going. The past is what they have already used to recruit her into the fight. She is not going back there.

The shared goal — a future piece of work both of them want to own and deliver together — is the format the conflict cannot survive. Sarah and Jim leave the room with one. The underlying conflict does not evaporate. It becomes a manageable tension rather than a triangulation through her.

So.

Nadella changed the institution that was making them fight. Scott changed the conversation that was letting them keep fighting. Same move, at two scales. The work becomes the only place collaboration can land.

The instinct, walking into a two-person conflict, is to mediate. To listen carefully. To take a balanced view. The move is to change what they are fighting over.

So. Your Next Move from this playbook.

Which of the two people are you quietly hoping will lose?

And what does that tell you about why you have been avoiding intervening?