"That's a straw man" is frequently correct and almost never persuasive. This site exists to help people name bad arguments, so it should be honest about what naming one usually accomplishes: it ends the exchange, hardens the position, and moves the conversation from the topic to the rules of the topic.
That's not a reason to abandon the vocabulary. It's a reason to know what the vocabulary costs.
Naming a move is a claim about standing
Say "that's an ad hominem" and you've done two things at once. The first is descriptive: their argument attacked the person instead of the claim. The second is positional: you have the terminology, they made the error, and you are now the one holding the rulebook.
The second part is what the other person hears. In a live discussion, especially a public one, most people register a fallacy citation as a status move before they register it as information. The reply is usually defensive, sometimes ornamental Latin of their own, and occasionally an escalation into whether the fallacy applies at all, which is a debate about labels rather than the thing you were originally discussing.
The vocabulary is precise, which makes it feel neutral to the person using it. It rarely feels neutral to the person receiving it. Precision and coldness are hard to tell apart in text.
A bad argument is not a false conclusion
This is the part that gets skipped most often, and it's the one that undermines the entire move when someone catches you at it.
Demonstrating that an argument is fallacious does not demonstrate that its conclusion is false. Someone can support a true claim with terrible reasoning. They can attack you personally and still be right about the policy. The failure of one argument leaves the underlying question exactly where it was. Treating "your argument is bad" as equivalent to "you are wrong" is itself a recognized error, usually called the fallacy fallacy, and it's the most common mistake made by people who have just learned the terminology.
Which means a fallacy citation is a much smaller claim than it sounds like. It says: this particular path to your conclusion doesn't hold. It does not say: your conclusion is wrong, I have won, the matter is settled. If you use it as though it says the larger thing, you've made a logical error while accusing someone else of one, and any observant reader will see it.
Tone determines almost everything
The same citation lands three different ways depending on delivery.
"lol classic straw man"
"That's a straw man."
"I don't think that's my position, and I'd rather defend the one I actually hold. What I said was closer to X."
All three point at the same straw man. The first is a taunt with a technical term attached. The second is a verdict. The third does the entire job of the citation, corrects the record, and gives the other person something to respond to that isn't their own error.
The third version also happens to be the only one that leaves them room to adjust without conceding a loss. That matters more than accuracy, because a person defending their dignity has stopped evaluating your point entirely. Nobody has ever changed their mind while feeling scored on.
Where the citation earns its keep
Naming the pattern works in a narrow set of conditions, and they're worth knowing precisely.
When bystanders are the audience. You will rarely move the person you're replying to. You might help everyone else reading follow what happened. A clean, unsmug label does that efficiently, and the clarity survives after the thread dies.
When the pattern is unmistakable. Whataboutism that swerves from a specific claim to an entirely different subject is legible to anyone. If you have to write two paragraphs explaining why the label fits, the label wasn't doing the work and the two paragraphs would have been better on their own.
When it's the third time. A single tu quoque is a reflex. A pattern of them across an exchange is a structure, and naming the structure is fair and often necessary. "Every time I raise this, the response is that I've done something similar. Can we take my point on its own for one reply?"
When you're describing your own move. "I was straw-manning you there, sorry, let me take the stronger version" costs nothing and buys an enormous amount of credit. It's the highest-value use of the vocabulary and the rarest.
Where it reliably fails
When the call is close. Many real arguments sit near a line rather than across it. Appeals to authority can be legitimate. Slippery slopes can be sound when the mechanism is shown. A confident label on an ambiguous case reads as either sloppy or hostile, and both invite an argument about your competence instead of their claim.
When you're using it to end things. A citation deployed as a closer signals you'd like the exchange over on favorable terms. People can feel that, and the reliable response to being dismissed is to become harder to dismiss.
When you did it too. Naming ad hominem two lines after calling someone a hack sets up a tu quoque that will be entirely deserved, and the rest of the thread will be about that.
In private conversation. With a friend, a partner, a coworker, reaching for a Latin term mid-disagreement changes the register from "two people working something out" to "one person litigating." That shift is hard to reverse.
The move that works better
Restate their argument in its strongest form, then ask the question the fallacy was routing around.
That's it. It does everything the citation does, without the status charge. Restating charitably proves you were listening, which buys attention. Restating in the strongest form removes any straw man from your side of the exchange. And the follow-up question puts the original issue back on the table, where a fallacy citation would have replaced it with a procedural dispute.
Them: "Anyone pushing this policy has clearly never run a business."
You: "I think the concern is that the costs land hardest on small operators, which is fair. Do we have numbers on how the last increase affected them?"
No label appeared. The personal attack was quietly discarded, the legitimate worry underneath it was rescued and stated better than they stated it, and the conversation is now about evidence. If they want to return to attacking, they have to do it against a version of themselves that sounded reasonable.
The short link still has a place, and there's a separate guide to using it well. Pair it with a sentence that restates the actual question and it works. Fire it alone as a verdict and you'll be right, briefly, in an argument that stopped being about anything.
The measure of a citation is not whether it was accurate. It's whether the conversation got better in the reply after it.