Running dry is what happens when you think the conversation is yours to fill. It isn't. It's yours to extract.
"I don't know what else to say." Look closely at what that sentence assumes: that the next thing must come from you — a new topic, generated from nothing, on the spot, while under mild social pressure. That's the hardest move available in conversation. It is also, measurably, the worst one. You've been reaching for the hardest and least effective option and concluding you're bad at this.
People ask a mean of 6.7 questions in a fifteen-minute conversation. When researchers coded 368 of these conversations by question type, the differences were not subtle. Huang, Yeomans, Brooks, Minson & Gino 2017, JPSP 113(3), 430–452
| Type | What it is | % of all | Effect on responsiveness |
|---|---|---|---|
| Follow-up | Draws on what they just said | 40.5% | +1.92 p<.001 — and the only type that also raises liking |
| Full-switch | New topic, unconnected | 27.6% | −2.83 p<.001 — actively harmful |
| Mirror | Bounced straight back ("You?") | 19.0% | +0.16 n.s. — inert |
| Introductory | Opening enquiry | 5.5% | +0.98 n.s. |
Now read your own failure mode against that table. "I don't know what else to say" is the panic of someone searching for a full-switch. You're hunting for a new, unconnected topic — the category that makes up 27.6% of all questions asked and that carries a negative effect of −2.83 on how responsive you seem.
You have been straining to produce the one move that damages the conversation, and interpreting your difficulty in producing it as a deficiency. The difficulty is real. The move is just the wrong one.
Here is the entire mechanism: they have just handed you the raw input. A follow-up question doesn't come from your stock of interesting topics, because it doesn't come from you. It comes from the last thing they said. You are not filling a gap; you are picking up something already on the table.
Which means the supply problem dissolves. You cannot run out of follow-ups while the other person is still talking, and if they stop talking, that's a different problem with a different fix (see the exits work in Floor Mechanics).
The mechanism the researchers identified is responsiveness — the sense of being listened to, understood, and cared about. It fully mediated the effect on liking. A follow-up is the cheapest available proof that you processed what someone said. Nothing else you could say demonstrates it as economically.
| Detail | Take one concrete noun from their answer and ask about it. "…Lisbon — how long were you there?" |
| Cause | Ask how the thing came about. "How did you end up doing that?" |
| Experience | Ask what it was actually like. "What's the part people don't expect?" |
That's the whole toolkit. Three shapes, applied to whatever they just said. You will notice none of them require you to know anything.
You'd expect people to have noticed something this useful. They haven't. In the same studies, question-askers did not predict any of the effect: predicted liking was 5.16 for high-question-askers versus 5.22 for low. No difference at all.
People do not know that asking questions makes them more likeable. Which means the internal signal telling you you're asking plenty, or that you ought to contribute something of your own now, is not measuring anything real. Don't consult it.
Once "ask follow-ups" lands, there's a characteristic way competent people execute it badly, and you should know about it now rather than acquire it and unlearn it later.
Boomerasking: asking a question, letting them answer, then immediately pivoting to give your own answer. Three flavours — ask-bragging, ask-complaining, ask-sharing. Brooks & Yeomans 2025, JEP: General 154(3)
So: if you want to share your own answer, just share it. Direct disclosure beats the question-shaped version by fifteen points. The question wrapper isn't politeness — it's a tax you pay for the appearance of interest.
Three conversations a day where you run a chain: their answer → your follow-up → their answer → your follow-up. No topic changes until the third exchange, at minimum.
Once a day, hold two minutes without introducing a single new topic. This will feel uncomfortable and slightly artificial. The discomfort is the diagnostic — it shows you how often you were switching topics to relieve your own tension rather than theirs.
Three days at three chains.
Follow-up ratio. After each conversation, jot two numbers: questions asked, and how many were genuine follow-ups — meaning they used something from the previous answer. Divide. Target above 60%, from a likely baseline near 40%.
Be strict with yourself on what counts. "So what do you do?" following an answer about their weekend is a full-switch wearing a friendly tone. If you could have asked it before they spoke, it isn't a follow-up.
Venue: two nights at Barcelona Language Exchange — target six separate people per night. This drill is unusually well suited to a language exchange, because people there are explicitly primed to talk about themselves and there is no ambiguity about whether questions are welcome.
From memory: what's the one-sentence test for whether something is a real follow-up, and what's the target ratio this week?
Test: could I have asked this before they spoke? If yes, it's a full-switch, not a follow-up. Target: above 60% of your questions being genuine follow-ups.
Huang, K., Yeomans, M., Brooks, A. W., Minson, J., & Gino, F. (2017). "It Doesn't Hurt to Ask: Question-Asking Increases Liking." Journal of Personality and Social Psychology, 113(3), 430–452. Free PDF from HBS
Read Study 1 and the taxonomy table. If you read one paper from this entire curriculum, read this one — it's the shortest path from "I know I should ask questions" to knowing which ones actually do anything.