Conflict & Feedback Guide

Before a feedback talk or a conflict conversation, check that you are ready, write the opening in situation, behavior and impact, decide what kind of conflict it is, and plan what you do if the first talk does not settle it.

Feedback effects vary widely and can be negative (meta-analysis: Kluger & DeNisi, 1996). SBI is practice, not research (Weitzel, 2000), as is Radical Candor (Scott, 2017). The task, process and relationship conflict types are strong research (Jehn, 1995; Jehn, 1997; meta-analysis: de Wit, Greer & Jehn, 2012). Psychological safety is strong research (Edmondson, 1999). Timings on this page are rules of thumb.

Use it when:

• You need to give someone feedback they may not want to hear, at any level: a colleague, a report, a client or your manager.
• Tension between you and someone else, or between two people on your team, is starting to affect the work.
• You have received feedback that stung and want to answer it well.

Skip it when:

• The issue is a pattern of performance against agreed goals: use 4.8 Expectations & Performance Canvas.
• The feedback is for your own manager about how they work: use 9.3 Upward Feedback Planner. Harassment, discrimination or safety concerns go to HR or the formal channel, not to this page.

Feedback research is more sobering than the advice genre suggests: across many studies the effects vary widely in size and direction, and feedback can make performance worse as well as better (Kluger & DeNisi, 1996). That is why structure and delivery matter. SBI is a practitioner tool from the Center for Creative Leadership (Weitzel, 2000), untested as a model; its value is that it keeps the message on observable behavior and its impact, not on character. Radical Candor is a framework from a popular book (Scott, 2017), not a researched method. The conflict side stands on firmer ground: research separates conflict about the task, about process and about relationships (Jehn, 1997), and relationship conflict is reliably harmful, while the effect of task conflict is contested. Everything here presumes psychological safety, the well evidenced belief that speaking up is not punished (Edmondson, 1999).

So use the structure as discipline, not as science. The readiness check stops you delivering feedback angry. The type of conflict decides the approach. The limit: a good opening cannot fix a relationship that needs a formal process.

Tick only what is true. If a box stays empty, prepare more or pick a better moment.

Strong answer: all five ticked, a day after the incident, with a private half hour booked in your regular one-to-one rather than a corridor.

• ☐ I am calm enough to listen, not only to talk
• ☐ I have a specific, recent example I saw myself
• ☐ I am addressing a behavior, not the person's character
• ☐ I have thought about how it may look from their side
• ☐ The time and place are private, with enough time

Illustrative example: the organization and roles are made up.

A project manager at an architecture firm in Gothenburg notices that a senior engineer keeps cutting off a junior designer in planning meetings.

Section 1: all five boxes ticked the next day; the talk goes into their regular one-to-one.

Section 2: "In Tuesday's planning, when the designer was presenting the layout, you raised the cost issue twice before she had finished. She stopped and moved on, and we lost her reasons for the layout. How did you see it?"

Section 3: on the surface a task conflict about cost; underneath, process: the team has no rule for when concerns are raised.

Section 4: the engineer had not noticed and was worried the cost point would be forgotten. Agreed: he writes concerns down and raises them after each presentation; the project manager adds a "concerns" slot to the agenda. He also tells the designer himself.

Section 5: not needed. At the check-in, the designer has presented twice without interruption, and she later raises a cost risk of her own in the concerns slot.