I used to think I was good at managing conflict. What I was actually good at was avoiding it.

Early in my management career, I’d watch two engineers disagree about an architectural approach and my instinct was to smooth things over. Find the middle ground. Suggest we “take it offline.” Anything to get past the uncomfortable moment and back to something that felt like progress. I told myself I was being diplomatic. In reality, I was being a coward, and I was making things worse.

It took me longer than I’d like to admit to understand that the absence of disagreement isn’t harmony. It’s silence. And silence in a team is far more dangerous than a heated debate about whether to use a message queue or direct API calls.

The Comfort Trap

Technical people, and I include myself firmly in this category, tend to gravitate towards conflict avoidance. We’re problem solvers by nature. We like clean solutions. Interpersonal conflict doesn’t have a clean solution, so we’d rather pretend it isn’t there.

Patrick Lencioni’s work on team dysfunction nails this. His pyramid of dysfunctions starts at the bottom with an absence of trust, and the very next layer up is fear of conflict. The two are inseparable. If people don’t trust that their colleagues have good intentions, every disagreement feels personal. Every challenge to an idea feels like a challenge to the person. So people stop challenging. They nod along in meetings and then complain in Slack DMs afterwards.

I’ve seen this pattern play out in teams of every size, from a twelve-person start-up where everyone was too polite to tell the CTO his pet feature was a dead end, to a hundred-person engineering org where entire architectural decisions went unchallenged because nobody wanted to be the one to rock the boat. The symptoms look different at different scales, but the root cause is the same: people have learned, implicitly or explicitly, that disagreement isn’t safe here.

The irony is that these conflict-avoidant environments don’t actually avoid conflict. They just push it underground. Instead of a direct conversation about technical trade-offs, you get passive-aggressive code reviews. Instead of honest feedback about someone’s proposal, you get corridor conversations and factions forming. The conflict is still there, it’s just become toxic rather than productive.

Kind vs Nice

One of the most useful distinctions I’ve come across in management literature is Camille Fournier’s framing of being kind versus being nice. They sound similar, but they’re fundamentally different approaches to leadership.

Being nice is telling someone their work is fine when it isn’t. Being kind is telling someone they’re not ready for promotion yet and then sitting down with them to map out exactly what they need to work on to get there. Being nice avoids the uncomfortable conversation. Being kind has it, because having it is the right thing to do for that person.

When I started looking into this distinction more carefully, I realised how much of my early management style was optimised for niceness rather than kindness. I’d soften feedback until it lost its meaning. I’d frame problems as “opportunities” until nobody understood there was actually a problem. I thought I was being supportive. What I was actually doing was depriving people of the honest input they needed to grow.

This plays directly into conflict management. A nice manager lets two team members’ simmering disagreement go unaddressed because raising it would be awkward. A kind manager names it, creates space to work through it, and helps both parties understand each other’s perspective, even when that process is uncomfortable for everyone involved, the manager included.

Fournier describes the “Conflict Avoider” as a management anti-pattern, and contrasts it with the “Conflict Tamer.” The Conflict Tamer doesn’t seek out drama or enjoy difficult conversations. They simply recognise that addressing conflict early and directly causes far less damage than letting it fester. They don’t turn a blind eye, and they don’t add unnecessary drama either. They address things clearly, calmly, and without making it a bigger deal than it needs to be.

Not All Conflict Is Equal

One thing that shifted my thinking on this was Will Larson’s distinction between transient conflict and recurring unresolved conflict. They’re completely different beasts.

Transient conflict is healthy. Two engineers arguing passionately about database schema design, then reaching a decision and moving on, that’s exactly what you want. It means people care about the work, they feel safe enough to disagree, and the team has a mechanism for reaching resolution. If your team never has these moments, something is probably wrong.

Recurring unresolved conflict is the problem. The same two people having the same fundamental disagreement every sprint, with no resolution and increasing resentment, that’s corrosive. It drains energy, it slows decisions, and it eventually forces other team members to pick sides.

The manager’s job isn’t to eliminate conflict. It’s to ensure conflict is productive and that it resolves. When I started framing it that way, my whole approach changed. I stopped trying to prevent disagreements and started trying to ensure we had the structures and trust to work through them.

Larson’s advice to approach conflict with curiosity rather than judgement resonated with me. When two people disagree, the interesting question isn’t “who’s right?”, it’s “what does each person know or value that’s leading them to a different conclusion?” More often than not, the disagreement is actually a signal that there’s important information the team hasn’t fully surfaced yet.

Building Structures That Help

Hoping people will handle conflict well on their own is a strategy, but it’s not a great one. What I’ve found works better is creating explicit structures and processes that make healthy disagreement the path of least resistance.

Larson describes LinkedIn’s approach to structured escalation, where there’s a clear, known process for what happens when people can’t agree. The beauty of a structured escalation process isn’t that it gets used constantly, it’s that its existence changes behaviour. When people know there’s a fair mechanism for resolving disagreements, they’re more willing to engage in them. When there’s no process, people either avoid the conflict entirely or it escalates in messy, political ways.

Fournier makes a related point: don’t rely on consensus. Consensus-driven decision-making sounds democratic and fair, but in practice it often means that the most conflict-averse person in the room has a veto. If anyone can block a decision by simply not agreeing, you’ve created a system that rewards avoidance and punishes bold thinking. Better to be clear about who owns the decision, ensure all perspectives are heard, and then let the decision-maker decide. People can disagree and commit, but only if the process is transparent and they trust it.

One practical approach I’ve adopted is what Stone, Patton, and Heen call “learning conversations.” The idea is straightforward: when there’s a disagreement, approach it as an exploration rather than a debate. Instead of each side arguing their position, you try to understand the other person’s perspective, what information they’re working from, what concerns are driving their view, what they value. It sounds simple, almost obvious, but it requires genuine discipline to do well. Our natural instinct in disagreement is to advocate, not to inquire. Flipping that takes practice.

I’ve also found Ian Nowland’s thinking on bikeshedding useful here. Sometimes what looks like conflict is actually just poorly structured discussion. When someone accuses a debate of being bikeshedding, the productive response isn’t to shut it down, it’s to reframe. Bring requirements, not solutions. If two people are arguing about implementation details, step back and ask whether they actually agree on the requirements. Often they do, and the “conflict” dissolves once the conversation is at the right level of abstraction.

When Conflict Can’t Be Resolved

Not all conflict is resolvable, and I think it’s important to be honest about that. Sometimes two people simply can’t work together effectively. Sometimes a fundamental disagreement about direction reflects genuinely incompatible values or priorities.

The framework I’ve found most helpful for thinking about these situations comes from Anil Kambil’s adaptation of Albert Hirschman’s work: Exit, Voice, Loyalty, and Optionality. When you’re in a difficult professional relationship, you broadly have four options. You can exit, leave the situation, move teams, change roles. You can use voice, raise the issue, advocate for change, have the difficult conversation. You can choose loyalty, accept the situation as it is and commit to making it work. Or you can create optionality, develop alternatives so you’re not trapped.

What I like about this framework is that it’s honest. It acknowledges that “just talk it out” isn’t always sufficient. Sometimes the kind thing, for everyone involved, is to acknowledge that a working relationship isn’t functioning and to find a structural solution rather than endlessly trying to fix an interpersonal one.

As a manager, I’ve had to make calls like this. Moving someone to a different team isn’t a failure of conflict resolution, sometimes it’s the most mature outcome. The failure would be leaving two people in a situation that’s making them both miserable and degrading the team’s output, just because addressing it feels uncomfortable.

The Ongoing Work

I won’t pretend I’ve got this completely figured out. Managing conflict well is genuinely hard, and I still catch myself defaulting to avoidance when I’m tired or stressed. The difference now is that I recognise the pattern and can correct for it.

The thing that’s helped me most is reframing what conflict means. It’s not a sign that something’s gone wrong. It’s a sign that people care enough to disagree. The leader’s job isn’t to prevent it, it’s to ensure the team has the trust, the structures, and the skills to work through it productively.

If your team never disagrees, don’t congratulate yourself on building a harmonious culture. Ask yourself what people aren’t saying, and why they’ve decided it’s not worth the risk.