28 October 2026 · 14 min read

Conflict in the workplace: split the argument from the friction

Most conflict at work is two capable people who disagree about something real, a few months of small digs, and a lead waiting for it to blow over.

The research is more precise about it than most advice is. One kind of conflict reliably hurts a team. The other depends.

The Monday item that never closes

Naveen leads client services at a company of about seventy people. For three months the Monday planning meeting has contained the same item.

Elspeth, who looks after the accounts, wants every new client to get an onboarding call in their first week. Tobias, who runs delivery, wants a written guide and a checklist, because calls do not scale and the same questions get answered over and over. Both have a point. Both have made it, calmly, many times.

Lately the item has changed shape. Elspeth says "some of us actually talk to clients". Tobias, eyes on his laptop, says that is one way to get nothing finished. Messages he could send to Elspeth now go out with Naveen copied in.

On Thursday both of them stop by Naveen's desk, separately. Elspeth asks whether he has noticed anything about Tobias. Tobias asks whether he really needs to be at the Monday meeting. Naveen says he will look into it, and tells himself that two good professionals will sort it out and that stepping in would only make it bigger.

Nobody has sorted it out. It has been three months.

How common is conflict in the workplace?

Naveen is invented. The numbers are not.

In an Acas survey of 4,558 working-age adults in Great Britain in August and September 2025, 44% had experienced conflict at work in the last 12 months (disagreements, disputes or difficult relationships with a colleague, manager, direct report, customer or other contact) (Acas, research by the National Centre for Social Research, How prevalent is individual conflict at work in Great Britain in 2025?, November 2025). The count includes customers and people outside the company, so not all of it is two colleagues in a meeting room. Being small is no shelter: the rate was 46% in small and medium enterprises and 42% in large organisations.

What people do next is in the same report. Among the 1,943 people who had experienced conflict at work, 19% did nothing to resolve it, 45% discussed it with their own manager and 30% talked informally with the other person involved (Acas, 2025). People could name more than one action, so those shares overlap. Nearly one in five did nothing. Elspeth and Tobias are both in the 45%, because both took it to their manager.

Doing nothing is commoner in smaller companies: 23% in small and medium enterprises, against 15% in large organisations. The report does not say why. My reading is that a smaller company has fewer people to take it to, and everyone still has to sit in the same Monday meeting afterwards.

Taking it to somebody does not always end it. Half of the people in Acas's 2025 survey who had experienced conflict at work said it was largely or fully resolved (23% and 27%), while 31% said it was mainly or not at all resolved (14% and 17%) (Acas, 2025, Table 8). Those are answers given at the time of the survey, so some of it may still have been running.

Task conflict or relationship conflict: which one hurts a team?

Researchers who study teams separate conflict into kinds, and two matter most here. A disagreement about the work itself, which option, what the answer should say, is task conflict. Friction between the people, tension and irritation and a growing sense that the other person is the problem, is relationship conflict.

What the studies found

Pooling 30 studies of work teams, De Dreu and Weingart found that both relationship conflict and task conflict were negatively correlated with team performance (mean corrected correlations -.22 and -.23) and with team member satisfaction (-.54 and -.32) (De Dreu and Weingart, Journal of Applied Psychology, 2003). These are averages across studies, and the sign matters more than the size. Neither kind went with better performance, and the link with satisfaction was stronger for relationship conflict. The authors say causation may run the other way, since poor performance can breed conflict.

A later meta-analysis of 116 studies of conflict in groups (8,880 groups) found stable negative relationships between relationship conflict and group outcomes, and between process conflict and group outcomes, but did not find a strong and negative association between task conflict and group performance; task conflict and performance were more positively related in top management teams, and where performance was measured as financial performance or decision quality (de Wit, Greer and Jehn, Journal of Applied Psychology, 2012). Process conflict is the third kind, the argument over how the work is divided and who does which part.

Read together, the two studies support a narrow conclusion. Relationship conflict reliably hurts. Task conflict is mixed, and what it does depends on the team and on what you measure. Neither says a good argument is good for a team, and neither tests a fix.

Reading Naveen's Monday

My reading of Naveen's Monday: the onboarding question is task conflict, and could go either way. "Some of us actually talk to clients" is not about onboarding. It belongs to the other kind. The two arrive tangled, and a lead looking at a tangle may not know which thread to pull, which makes doing nothing the easier choice.

Signs a disagreement has turned personal

The slide from one kind to the other is easy to miss, because nobody announces it and each step looks small on the day. These are the ones a lead can see this week.

• The same item keeps coming back. New details, same two positions.
• An audience appears. A message that could go to one person now copies in a manager.
• The argument changes subject. "Some of us actually talk to clients" is a remark about Tobias.
• Someone asks to skip the meeting. The agenda is the same. The room has become the thing to avoid.
• Sides form. A colleague with no view on onboarding gets asked whose side they are on.

Two or three of those together and the disagreement is no longer the whole problem. Winning it would not end what has started.

The colleague caught between them

Hana joined Naveen's team in the spring. She looks after clients once they are live, so she needs Elspeth's notes and Tobias's checklist, and on Tuesday she asks Elspeth where one client's setup stands. "You'd better ask Tobias, he has the guide." She asks Tobias. "Ask the people who actually talk to clients." She has started copying Naveen on everything, because everybody else does.

Ask Hana whether she has experienced conflict at work and she may well say no. She is the one walking between two desks, and she is new enough to think this is how the company works.

What workplace conflict costs

The first bill goes to the people in it. Of the people in Acas's 2025 survey who had experienced conflict at work, 57% reported stress, anxiety or depression as a result and 49% a drop in motivation or commitment; 17% said it had no impact on them (Acas, 2025, Table 10). That is people's own account of what the conflict did to them, not a clinical finding.

Then the employer's. Acas's 2021 research estimated that workplace conflict costs UK employers £28.5 billion a year, just over £1,000 for every employee, using data from 2018 to 2019 (Saundry and Urwin for Acas, Estimating the costs of workplace conflict, May 2021). That is a modelled estimate of the cost to employers, and the authors call their approach conservative. The report puts the largest share of it on the end of the employment relationship, through resignation or dismissal.

None of that is on Naveen's budget. The first cost he could point to is the Monday meeting itself, where everyone else at the table has spent three months watching two colleagues have the same argument.

Should you wait for workplace conflict to blow over?

There are three easy wrong answers.

The first is to wait and see. For something small that can be a sound call: in the same Acas survey, doing nothing was more common among people who said the conflict had no impact on them than among those who reported some impact. By the third month, Naveen's is no longer small. Acas's conclusion, drawn from its wider research and not from this survey, is that doing nothing is not always effective, because unaddressed issues can stay latent and resurface.

The second is the away day with a facilitator. Neither study tested one, so this is my reading: it gives two tired people a day of goodwill and a flipchart, then sends them back to the Monday meeting, where the question and the digs both still live. One-off fixes have a poor record on a harder problem: just 6% of organisations said behaviour improved permanently following intervention, according to research by WorkNest (WorkNest, via HRreview, 21 August 2026, and our note on it). That is about toxic behaviour rather than two colleagues and an onboarding call.

The third is for Naveen to decide: calls, final, moving on. Sometimes a lead has to. It closes the onboarding question and leaves the dig and the copied-in message where they were, with one of the two now the loser, so the next dispute starts out personal.

How to resolve conflict between two colleagues

None of this needs a product.

Write the disagreement down as one question about the work

Something like "how do new clients get started with us?" Keep the friction on a separate list. Ask each of them, alone, what has been getting in the way between them, and keep that out of the question. Roger Fisher and William Ury set out principled negotiation at the Harvard Negotiation Project, and separating the person from the problem is one of its four parts. My reading of the studies above is that they are a reason to keep the two conversations apart.

Ask what each of them is protecting before you ask what they want

Ask "what goes wrong for you if you lose this one?" instead of "call or guide?" Elspeth may be protecting a client's first impression, Tobias his team's week, and those two needs need not collide. This is the interests step in Fisher and Ury's method.

Agree what will decide it, once

Choose a standard that does not depend on who argues better: a date, a cost, or a trial of each approach for a month with a review at the end. Then argue about the standard once, instead of arguing about the dispute every Monday.

If what you are looking at is more than friction between people who disagree, it is not yours to settle alone. Take it to whoever your company has named for concerns.

The weak point in all three is the next Monday, when the meeting is full and the old exchange is quicker.

Common questions about conflict at work

When should a manager step in on a conflict at work?

Step in when the argument stops being about the work. A disagreement over which option is right can run for a while and may even help. Once people are commenting on each other, copying in an audience or skipping meetings, waiting mostly gives the two kinds of conflict longer to tangle.

Should you talk to two colleagues together or separately?

Separately first, then together once the question is on paper. Apart, each can say what is getting in the way without performing for the other. Together, they need a written question about the work and an agreed way of deciding it. A joint meeting before that risks staging the same argument with a witness.

What if the conflict is between two teams, not two people?

Use the same tools one level up. Find what each team needs, write the dispute as a question neither owns, and agree a standard for deciding it. In Muzafer Sherif's studies of rival groups, contact alone did not end the rivalry and a shared goal that needed both sides did, so a joint social is a weak substitute for that goal.

How MeaningFuel helps when a disagreement has hardened

The daily insight: Respect, Negotiation and Emotional Intelligence

Nobody can choose conflict as a daily subject, because a problem is what a module is for. What a company chooses are the values and skills it wants to work on, and each working day everyone gets one short insight on one of them. The nearest here are Respect as a value, and Negotiation and Emotional Intelligence as skills.

In an ordinary week Elspeth might meet a line on listening to understand, and Tobias one on disagreeing without contempt. Neither will close the Monday item. They keep how people treat each other in the working week, including the weeks when nobody is fighting, which is when that habit is cheapest to build.

The module: On the Same Side

On the Same Side is the module for a team where a disagreement has hardened into two fixed positions. It is a short course of daily lessons: ten short lessons across two working weeks, one each working day, for a team's named people. Whoever runs the account, or leads the team, books it for the team, never for a person, so nobody is sent on it for being the problem. It invites everyone on the team, and those who join get one lesson each working day.

In one lesson, each person is asked to turn a live tension into a question neither side owns, and to sit on the same side of it as the colleague they disagree with: the same side of a table, a call or a shared document. The lesson borrows from Muzafer Sherif's studies of rival groups, in which contact alone did not end the rivalry and a shared goal that needed both sides did.

Another lesson starts from a dispute that keeps coming back: Kai insists a fix happens now and Lin insists it waits, and underneath, Kai lives with the daily pain of the problem and Lin has promised a customer a date. Those two needs do not conflict, and nobody had asked. Later lessons ask for at least three options before anyone commits, and for one agreed standard to settle the team's most recurring dispute.

It draws on research by Roger Fisher and William Ury, and on CREW, the US veterans' health service's civility programme.

What Naveen can read, and what he cannot

Insights shows how much of a team takes part and how the insights land, pooled, with nobody named. It would not tell Naveen about the Monday item. A team can take part every morning with two of its people copying in their manager, and no screen shows how one person answered. During the module he sees whether the team is doing it, and never anyone's answers or what they write, because nobody is honest about a dispute the lead can read.

What the module does is give the habit ten working days of practice, so that keeping the person and the problem apart does not rest on one lead noticing on the right Thursday. It is not mediation, and it will not settle a grievance. Where the competition is built into the structure, with rankings that grade people against each other or promotions only one person can win, the module says so upward, because norms cannot outvote incentives for ever. It promises no outcome.

The item on next Monday's agenda

Elspeth and Tobias may never agree about onboarding. They only need to agree on what will decide it. What Naveen has been leaving alone for three months is the part nobody needs to win: the dig and the copied-in message. Pull that apart from the question, and the question goes back to being what it was at the start, a real disagreement between two capable people, with an answer to find.

• Workplace Conflict
• Internal Conflict
• Team Leadership
• Negotiation
• Workplace Culture