Conflict at work

What conflict actually costs, and the direction it is heading

The reference study is the CPP Global Human Capital Report from July 2008, a nine-country survey by the publisher of the Thomas-Kilmann instrument. It defined workplace conflict as any disagreement that disrupts the flow of work and found it nearly universal: 85% of employees at all levels experienced conflict to some degree. US employees spent 2.8 hours per week dealing with it, which the report valued at approximately $359 billion in paid hours for that year. The follow-up matters more than the original. In 2022 The Myers-Briggs Company re-ran the research and found the average time spent had roughly doubled, to about 4.3 hours per week, with people who manage others spending over 4 hours a week on conflict. More than a third of respondents said they deal with conflict often, very often, or all the time, up from 29% in 2008. Poor communication was the most-cited cause.

Read those figures as directional survey data, not accounting. The 2008 dollar figure was built from average hourly earnings of the time and is well-aged now. The trend, though, is the point: conflict is not a marginal cost that a growing company outruns. It scales with headcount and complexity, and the time it consumes has been moving in one direction.

The five-style map and what it does not tell you

Most conflict content online traces to one model: the dual-concern framework, which maps conflict handling on two dimensions, how assertively you pursue your own concerns and how cooperatively you attend to the other person’s. Thomas and Kilmann built the best-known version in 1974, naming five styles at the corners and center of that grid. The map is genuinely useful vocabulary. Its limit is what a preference instrument can measure: it tells you which style you reach for by default, not how skillfully you execute any of them, and not whether you pick the right one for the situation in front of you.

How to read it. All five styles are situationally correct, which is the model’s own position. That is why a style label is a starting point for self-awareness and a poor measure of capability, and why skill-focused assessment works from scenarios rather than preference ratings.

Task conflict is not automatically the good kind

A popular management claim holds that relationship conflict is destructive but task conflict, disagreement about the work itself, is healthy and should be encouraged. The strongest evidence check on that claim is the 2003 meta-analysis by De Dreu and Weingart in the Journal of Applied Psychology, covering the accumulated team-conflict research. It confirmed that relationship conflict correlates strongly and negatively with team performance and member satisfaction. It also found, contrary to what textbooks had been teaching, that task conflict correlated negatively with both as well, and that the negative relationship was stronger in complex work such as decision-making and project teams.

The practical reading is not that disagreement is bad. It is that disagreement carries no automatic benefit. Task conflict helps only under conditions, chiefly when it stays on the problem, does not spill into relationship conflict, and gets worked by people who can hold it steady. In other words, the value of conflict is conditional on the handling, which puts the leverage exactly where an individual can build it: in behavior, not in temperament.

The four handling behaviors that hold up

Across the survey research, the negotiation literature, and the de-escalation evidence, effective handling keeps reducing to four behaviors. First, raise it early and directly, with the person involved. In the 2022 Myers-Briggs research, addressing conflict quickly, directly, and earlier was among the top improvements people wanted from those around them, and issues raised while small are cheap in exactly the way issues raised late are not. Second, stay steady in the moment: slow the pace, and say the other person’s point back in your own words before answering it. Paraphrase de-escalates because it proves listening without conceding anything, and a named break with a set resume time beats looping louder. Third, work interests rather than positions. A position is a single demand; the interest behind it usually admits several solutions, and asking what a position gets the other person is the move that breaks stalemates. That is the core of the principled-negotiation method Fisher and Ury published in Getting to Yes. Fourth, close it and repair it: end a resolved conflict with what was agreed, who does what, and when you will check, and repair the working relationship where the heat did damage. Conflicts that end on good feelings and nothing specific reopen.

How to read it. The interest-based version is not softer. It is faster, because it moves the conversation to the only place a solution can come from.

Why self-ratings mislead here

Conflict handling carries a documented self-perception gap. In the 2008 CPP study, 31% of managers considered themselves skilled at handling conflict while only 22% of the employees under them said their managers deal with conflict well, and the 2022 follow-up still found roughly one person in four rating their manager’s conflict handling as poor or very poor. The gap is not surprising: people judge themselves on intentions and remember the conflicts they resolved, while the people around them experience the ones that were avoided, ran hot, or reopened. It is also why a self-rating alone is a weak measure of this skill. A scenario read, what you would actually do in specific situations, compared against your self-view, shows you not just where you stand but where your self-assessment is out of calibration, which is often the more useful finding.

Where these figures come from

Tools built on this research

Common questions