So you’re the one responding

The invitation reads differently than the ones you’re used to. Not “would you present,” not “would you sit on the panel” — “would you offer a response.” Someone else is giving the talk. Your job starts the moment they stop, with eight or ten minutes to say something useful about work you may have seen for the first time twenty minutes ago. It is a smaller, stranger job than presenting or moderating, and almost nobody explains what it’s actually for before you’re standing up to do it.

What the room is actually asking for

A response is not a second talk on the same subject, and it isn’t a rebuttal either, even when the organizer’s program calls the slot a “reaction” or a “discussant comment.” What the room wants is the thing only someone who just listened with full attention can supply: where this connects to the talk two sessions ago that the speaker didn’t reference, the assumption in slide six that would change the conclusion if it’s wrong, the question the Q&A is about to spend ten minutes circling if nobody names it first. That’s a genuinely different skill from having your own position on the topic, which is what sitting on the panel actually asks of you. A response only works in service of the talk that just happened, not as a platform for the talk you’d have given instead.

Two different jobs, depending on what you got in advance

If the organizer sent you a paper or a full draft ahead of time, your job is preparation: read it twice, mark the two or three places you’d push back or want more, and write real notes before you arrive — a response built from a week of thought reads differently than one built from twenty minutes, and the room can tell. If instead you’re responding cold, live, off the talk you just heard, the discipline is different and arguably harder: take structured notes during the talk itself, not just interested ones — a running list of claims, questions, and connections, organized as you go, so you’re not reconstructing your reaction from memory the moment you stand up. Ask the organizer which situation you’re in as soon as you accept; conflating the two is the single most common way a response comes across as underprepared.

Credit first, questions second

Open by naming, specifically, what the work actually accomplished — not a throat-clearing compliment, but the real thing: the dataset nobody else has assembled, the argument that reframed a question the field usually asks badly. This isn’t etiquette for its own sake. A response that opens with what’s wrong reads as a rebuttal before it’s finished its first sentence, and the room stops hearing the substance underneath it. A response that opens with genuine, specific credit earns the latitude to spend its remaining minutes on exactly the same critical questions — the room just hears them differently, as a colleague pushing the work forward rather than a competitor scoring points.

When you actually disagree with the work

Say so plainly, and say why, using the speaker’s own terms rather than a strawman version of their argument — the fastest way to lose the room’s trust is disagreeing with a claim nobody made. Direct the disagreement at the work, not the person delivering it, and leave the speaker real room to respond; a good response ends with a genuine question, not a closing argument. If the disagreement is about something structural — a methodology choice, a scope the talk didn’t address — frame it as a question for the room’s benefit (“how would this change if X”) rather than a verdict delivered from the response chair. You’re still, for these ten minutes, a guest in someone else’s session.