Skip to content

Organisational psychology

Team conflict: separate the task from the relationship

Handle team disagreement by separating task, process and relationship issues, using shared criteria, evidence and a clear decision owner.

Alessandro ScuottoPublished 8 min read

Two colleagues disagree about a launch. One wants more testing; the other wants to meet the date. Soon the conversation moves from the product to the people: “You always slow things down” and “You only care about the deadline.”

The original decision has become harder to examine. Evidence is mixed with accusations, and each person now has a relationship to defend as well as a proposal.

Handling team conflict starts by separating what is being disputed. Is it the work itself, the way responsibility is distributed, or behaviour between people? You may need to address more than one layer, but discussing them as one vague interpersonal problem usually obscures the decision.

Understand three different questions

Task disagreement concerns the content of the work: which design, interpretation or course of action is better. Process disagreement concerns how the work is organised: who does it, who decides and how resources are allocated. Relationship conflict concerns tension between people, including irritation and personal hostility.

These categories help organise a conversation. They are not fixed identities assigned to employees. A task disagreement can acquire a personal dimension; an unclear process can repeatedly bring the same people into tension.

Karen Jehn’s research on intragroup conflict distinguished the effects of different kinds of conflict. Do not simplify that research into “task conflict is always good.”

De Dreu and Weingart’s meta-analysis found negative associations that challenge such a simple rule. Context matters. Your aim is a decision process where relevant differences can be examined, not a higher level of conflict.

Identify the decision before asking people to cooperate

“Please align” is an unclear instruction when two people are optimising for different outcomes. Write the specific decision that the team must make.

For the launch example, it might be: whether to release the current scope on the planned date, narrow the scope or move the date. Those alternatives are more useful than “quality versus speed,” which makes each position sound morally complete.

Name the criteria. Which failures would make a release unacceptable? Which delay has a real consequence? Which missing information could materially change the choice? Include constraints that someone else owns.

If the decision was already made, do not present the meeting as an open debate. Explain what remains open: implementation, a new risk or a justified request to revisit the choice. An honest boundary is better than a conversation where contributions cannot affect anything.

Separate observations from accusations

Ask each person to describe the relevant episode without stating the other person’s motive. “The last two review requests arrived after the scope deadline” can be checked. “You do not respect my work” needs a different kind of conversation.

You should not require people to strip all emotion from their account. Frustration can signal a recurring problem. But emotion does not establish an intention, and an alleged intention should not become the factual basis for the project decision.

A facilitator can restate the practical issue and ask whether the description is fair. Record points of agreement and unresolved facts. If the conversation repeatedly returns to personal attacks, pause rather than treating endurance as a sign of commitment.

The psychological safety guide explains how to respond to difficult information while preserving standards. Speaking up and respectful conduct belong together.

Use independent positions before a live debate

Ask participants to write a short position in advance: their recommendation, supporting evidence, main uncertainty and what would change their mind. Limit the document to the actual decision.

Independent preparation helps you distinguish different reasoning from different presentation styles. A confident speaker should not gain automatic authority over a carefully documented risk.

During the discussion, ask someone to summarise the other position accurately before responding. They need not agree with it. The purpose is to verify that the argument being answered is the one being made.

Do not turn this into a compulsory empathy performance. If someone reports inappropriate conduct or a serious power imbalance, the immediate need may be an appropriate reporting route, not a joint exercise.

Make decision rights explicit

A team can understand each other and still disagree. Consensus is one way to decide, but it is not the only legitimate one.

Name who is accountable for the choice and whose input is required. State whether a particular function has an approval obligation. A project owner cannot quietly override a specialist control simply because the deadline is close.

Agree the point when the decision will be made and what happens if information remains incomplete. This prevents the conversation from restarting with each new message.

Once the owner decides, record the rationale, the unresolved risk and any review trigger. Committing to execution does not require pretending that everybody preferred the choice. A colleague can disagree and still honour an understandable decision process.

A hypothetical disagreement over an AI feature

Imagine a product lead and a customer-support lead reviewing an AI answer assistant. This is a hypothetical example. Product wants to launch to learn from use; support worries about incorrect answers reaching customers.

The facilitator identifies three questions. The task question is which cases the tool handles acceptably. The process question is who reviews uncertain answers. The relationship issue is that previous concerns were described publicly as obstructive.

The team addresses the conduct issue separately and removes that language from the working discussion. For the product decision, it compares a full launch, a limited internal pilot and a release covering only reviewed cases.

A named owner decides within the organisation’s approval rules. The chosen plan specifies who checks answers and which failure would suspend use. The team records the alternatives it rejected and why.

The example offers no claim that the disagreement disappeared. It illustrates a way to keep the product choice, work allocation and interpersonal issue from being resolved through the same argument.

Use this disagreement brief

Before the next contested decision, ask the people involved to complete the same fields. Keep the brief available to whoever owns the final choice.

Swipe or use the arrow keys to see every column.

FieldWhat to recordWhy it matters
DecisionThe choice to make and its deadlinePrevents debate about unrelated grievances
AlternativesReal options, including narrowing or pausingAvoids a false choice between two extremes
CriteriaQuality, delivery and required controlsMakes trade-offs discussable
EvidenceFacts, examples and their limitationsSeparates information from confidence
UncertaintyWhat remains unknownPrevents assumptions becoming hidden facts
AuthorityDecision owner and required approvalsShows how the disagreement can end
Review triggerNew evidence that would reopen the choiceKeeps the decision firm without making it permanent

The brief does not make every value conflict measurable. Two functions may legitimately weight consequences differently. It makes those weights visible so an accountable person can choose, rather than allowing an argument to decide by exhaustion.

Handle written conflict with enough context

Text makes it easy to answer the latest sentence and forget the underlying decision. Long threads also accumulate fragments of evidence, sarcasm and unspoken assumptions.

When the exchange stops producing new information, move the decision into a short document. Summarise the options and disputed facts. Use a conversation when the disagreement needs clarification, then record the outcome.

Do not resolve a sensitive interpersonal issue in a public channel. Public visibility can help track a project decision, but it can also expose an individual unnecessarily.

Respect time zones and accessibility. A late message should not require an immediate defence from someone who is offline. Give people a reasonable opportunity to review evidence before treating silence as agreement.

Know when facilitation is the wrong response

Harassment, intimidation, discrimination, threats and retaliation need the relevant organisational procedures. A manager should not reframe them as a healthy debate or ask the affected person to find a compromise.

Repeated conflict may also reflect a structural problem: incompatible objectives, insufficient capacity or overlapping authority. Coaching people to communicate better will not solve a resource allocation that nobody can authorise.

Use the organisational psychology guide to examine that wider arrangement. For decisions that keep returning to the manager, delegation and accountability helps specify authority. For conflicting priorities and repeated reconstruction, examine team cognitive load.

Review the process, not how quiet the team becomes

A silent meeting is not evidence that conflict has been handled. People may have lost interest or decided that speaking is costly.

Look instead at whether contested choices have owners, whether evidence is considered, whether personal attacks stop and whether decisions remain understandable after the meeting. Ask the participants what unresolved work or conduct issue remains.

If a decision is repeatedly reopened, identify the cause. The review trigger may be unclear, new information may genuinely matter, or somebody may be bypassing the agreed route. Each possibility requires a different response.

This discipline resembles the distinction between observation and interpretation in consumer psychology. Behaviour is a starting point for a question, not immediate proof of a motive.

Choose one disagreement this week. Write the decision and the alternatives. Separate the evidence, work allocation and interpersonal issues. Give each the right owner and the right process.

If a digital programme is stuck in the same debate, tell me which decision cannot reach an owner. We can map the trade-off and create a practical route to a decision.

Share this article

LinkedInXWhatsAppEmail

Apply this to your project

Tell me the decision you are facing, the constraints and what you have already tried.

Complete all three fields to send your message.

Organisational psychology

Motivation at work: give autonomy a clear shape

Understand motivation and autonomy at work: clarify choices, support competence and design responsibilities people can exercise in practice.

8 min read