Sources checked 31 October 2026. The conversations, replies and checklists below are worked examples written for this article, not quotes from the studies; the research findings are described with their limits. This is not therapy or relationship advice.
Your friend sits down, lets out a long breath and says, "Work is falling apart. My manager keeps moving the deadline and I can't sleep."
You care about this person, and you can see the answer. So you give it. "Just quit. Seriously, update your CV tonight. Or talk to HR first. And honestly, try a budget app so you have a cushion if you leave."
Four good ideas in twenty seconds. Each one is reasonable. You might even be right about all of them.
And your friend looks worse than when they sat down. Their shoulders drop. They say "yeah, maybe," in the voice people use when they want the conversation to end. Later you wonder what went wrong. You were trying to help.
Nothing was wrong with the advice. Something was wrong with when it arrived.
Why can good advice land badly?
Advice is one of the most common things people offer when someone shares a problem, and researchers who study supportive communication have spent decades asking why it so often misses.
A key early study came from communication scholars Daena Goldsmith and Kristine Fitch, who in 1997 examined how people talk about giving and receiving advice in everyday life [2]. They found that advice sits inside a set of tensions. The same words can feel like caring and like butting in. They can feel honest and like criticism. Above all, being told what to do can carry a hidden message about the listener: you haven't thought of this, or you can't handle this on your own.
Researchers call this a face threat. "Face" is the image of yourself you want others to accept: capable, sensible, in charge of your life. When someone rushes in with a fix, especially one you didn't ask for, it can quietly suggest you are none of those things. The problem is no longer just the deadline. Now it is also the feeling of being judged.
Goldsmith later pulled this work together in the book Communicating Social Support [3]. Goldsmith's argument is that support isn't a thing you hand over. It is something two people build in conversation, and the same message can help or hurt depending on how and when it shows up.
Illustrative lines written for this article, based on the face-threat idea in Goldsmith & Fitch (1997) [2]. Not everyone hears advice this way.
What did the 2009 study test?
If advice can threaten face, does that mean we should stop giving it? Communication researcher Bo Feng asked a more useful question: does the order matter?
Feng proposed what is called the integrated model of advice giving. In this model, a helpful support conversation tends to move through three moves [1]:
Emotional support. Acknowledge the feeling: "That sounds exhausting."
Problem inquiry and analysis. Ask about the situation before judging it: "What has your manager said about it?"
Advice. Only then, offer a course of action.
In a 2009 paper in Human Communication Research, Feng tested this sequence [1]. Participants evaluated supportive conversations in which the advice either came after emotional support and questions about the problem, or did not. Advice that followed the full sequence was rated more favourably than advice given without it.
The key word is rated. In studies like this one, people judge advice in conversations the researchers set up, rather than receiving it about their own problems from their own friends. That is a real limitation. A message that scores well on a rating scale may not always land the same way at 11 p.m. with a friend in tears. Later work tested the model in more than one culture, including the United States and China. A 2014 study with 572 participants in the US and 540 in mainland China found the same pattern in both: advice that came after emotional support and problem inquiry was rated higher in quality, and people said they were more likely to act on it [6].
So the finding is modest, and it is useful. It does not say advice is bad. It says advice tends to go over better when it is not the first thing the other person hears.
Worked example. The comfort → questions → advice order follows Feng's integrated model [1]; the 'feels heard' bar is illustrative, not measured.
Isn't asking questions just a delay?
It can look that way from the inside. If you can already see the answer, questions feel like wasting time.
But the questions do at least three jobs.
First, they tell your friend you took their problem seriously, which reduces the face threat. You are treating them as the expert on their own life.
Second, they give you information. "Just quit" sounds less clever once you learn they are on a visa tied to the job, or that the deadline only moved because their manager is covering for a sick colleague. Advice built on a two-sentence summary often solves a problem the other person doesn't have.
Third, they give your friend room to think out loud. Sometimes people already know what they want to do. They need to hear themselves say it.
Research on advice suggests that several features shape how it is received, including whether the person asked for it, whether the adviser seems to understand the problem, and whether the advice is actually feasible for them [4]. Asking first helps with all three.
When do people actually want advice?
Some people, some of the time, want a fix straight away. A friend whose car won't start may want the number of a good mechanic, not a conversation about their feelings. Treating every problem as a request for emotional support can be its own mistake.
That is why receptiveness matters. In a study of what predicts how open people are to advice, Feng and Erina MacGeorge found that features of the problem, the adviser and the recipient all played a part [5]. Advice that was asked for, and that came from someone seen as caring and knowledgeable, generally met a better reception.
You can't read minds. But you don't have to. You can ask.
A rule of thumb adapted from the integrated model of advice giving [1] and research on asked-for advice [5]; not a tested script.
Why do we fix so fast?
If listening first works better, why is the fixer reflex so common?
Part of it is kindness. Seeing someone in pain is uncomfortable, and offering a solution feels like doing something. Part of it is competence: being useful feels good, and a clear answer is a gift you can wrap quickly. And part of it is simple habit. Many of us have been praised for solving problems at work and at school, and we bring that reflex home.
None of this makes you a bad listener. It makes you a normal one. The fix is not to bottle up your ideas forever. It is to hold them for a minute.
A small, practical way to do that is to notice the first sentence that starts forming in your head. If it begins with "just," "you should," or "why don't you," that is the reflex. Swap it for one sentence that acknowledges the feeling and one question about the situation. Your advice will still be there afterwards, and it will probably be better.
What can you say instead?
The simplest move is the one in the video: before you solve, ask.
"Do you want advice, or do you want me to listen?"
It sounds almost too simple. But it does what the research describes. It recognises that your friend gets to decide what kind of support they need, and it turns advice from something done to them into something offered to them.
Other versions work too:
"Do you want ideas, or do you mostly need to vent right now?"
"I have a couple of thoughts. Want them now, or later?"
"That sounds really hard. What's the worst part of it for you?"
If they say "just listen," then listen. Reflect back what you heard. Ask what happened next. If they say "I'd love ideas," you now have permission, and you have the context from the questions to make your advice fit.
Three questions on the findings described in this article [1][2][5].
The one thing to try this week
The next time someone you care about brings you a problem, notice the fix that rises in your throat. Let it wait. Say one sentence about how hard it sounds, ask one question about what is happening, and then ask whether they want advice or a listener.
Advice isn't the mistake. The timing is.
Sources
Feng, B. (2009). Testing an integrated model of advice giving in supportive interactions. Human Communication Research, 35(1), 115–129. https://doi.org/10.1111/j.1468-2958.2008.01340.x. Caveat: participants rated advice in conversations set up by the researcher, not advice about their own problems; ratings may not match real-life reactions.
Goldsmith, D. J., & Fitch, K. (1997). The normative context of advice as social support. Human Communication Research, 23(4), 454–476. Caveat: qualitative work with a limited, mostly American sample; it describes tensions people report rather than measuring effects.
Goldsmith, D. J. (2004). Communicating Social Support. Cambridge University Press. Caveat: a scholarly synthesis and argument, not a single experiment.
MacGeorge, E. L., Feng, B., & Thompson, E. R. (2008). "Good" and "bad" advice: How to advise more effectively. In M. T. Motley (Ed.), Studies in Applied Interpersonal Communication (pp. 145–164). Sage. Caveat: a review chapter summarising earlier studies.
Feng, B., & MacGeorge, E. L. (2006). Predicting receptiveness to advice: Characteristics of the problem, the advice-giver, and the recipient. Southern Communication Journal, 71(1), 67–85. Caveat: self-reports from university students recalling advice; correlational, so it can't show cause.
Feng, B. (2014). When should advice be given? Assessing the role of sequential placement of advice in supportive interactions in two cultures. Communication Research, 41(7), 913–934. https://doi.org/10.1177/0093650212456203. Caveat: participants judged described conversations and reported intentions to act, not actual behaviour; two countries are not all cultures.
Watch the episode
One email when we publish. Research, product decisions, and what teams report back.
