Second date ideas: what to do, what to talk about and when to ask

Updated October 5, 2026 · 15 min read

A second date is for finding out whether the first one was luck. Choose a format by what you want to learn: a walk tests conversation, an activity tests ease, dinner tests both and costs most. One experiment covers when to text afterward. We found no study on how many days to wait, so that part is opinion, labeled.

We looked at the pages that rank for second date ideas and at the research on who ends up meeting again. The forum threads cited here are all from Mumsnet, a British parenting site whose users are mostly women; we did not check its audience figures. Anything that is our opinion is labeled. For the first meeting itself, see our first date ideas, or the wider first-date guides.

Why the format matters more than the venue

We read the lists that rank for second date ideas, among them Today (updated August 30, 2024) and Utah State University Extension (May 17, 2024), and Bumble’s page of expert tips. None of them cites any dating research (Today, 2024; USU Extension, 2024; Bumble, undated). That is partly fair, since little of the research is about second dates as such. It does mean each idea arrives without a reason attached.

Our view: a second date has one job, which is to tell both of you whether you want a third. The first date showed that the profile was a person. The second shows whether that person is still good company once the nerves wear off. No checklist settles that in advance. In one speed-dating study, a model given more than 100 self-report measures could not predict who would want whom in particular (Joel et al., 2017). So the table below sorts ideas by what each one lets you find out, and treats the venue as a detail.

Second date ideas, sorted by what each one tests

There is no best second date idea in general, only good ones for a particular question. The table sorts 24 of them into seven formats. Our pick for the fun ones: the activity and live-show rows, though fun is not a scored column. The cheapest are walks, markets, museums on free days and parks.

The scoring is ours; nobody has measured any of it. Pressure is how much the format leaves you facing each other with nothing else to look at. Cost is per person and rough, since prices vary by place. Exit after an hour means either of you can end it politely at that point without it meaning anything.

Cheap is where the arguments start. In one Mumsnet thread, a poster whose date had suggested sitting in the park read it as low effort, and the 458 replies argued cost against effort, with some suggesting a counter-proposal (Mumsnet, 2024a). In another, a woman who had planned the first date was offered a dog walk for the second and read it as a preview of who would be doing the planning from then on; the replies split between moving on and defending walking dates (Mumsnet, 2024b).

Our view: a free date is fine, and saying why you picked it does most of the work. “The park, because there’s a coffee cart by the pond and I owe you better coffee than last time” is a plan. “The park” on its own can read as no plan at all. The hard-exit rows suit a second date you already expect to go well, since they commit you to a couple of hours either way.

What to do on a second date

Our rule of thumb: an activity first, a sit-down second, and the whole thing under three hours. Doing something at the start gives you a shared subject and somewhere to look besides each other. Moving to a table afterward shows whether the conversation holds without props. Our guess: ending inside three hours makes the third-date question easier to answer. We found no study on date length.

Some pairings on that logic: minigolf, then a drink nearby. A market, then lunch at whichever stall won the argument. A gallery, then coffee to disagree about what you saw.

Our rule of thumb on who pays: whoever asked offers, and either answer is fine. We found no research on who pays on second dates.

On kissing, we found no research about second dates in particular, and we are not going to invent any. Our view: consent comes before timing. Ask, or slow down enough that the other person has room to answer without words. A turned head, a step back or a “not yet” is a no for that evening, and it needs no debate. Bumble’s undated page of expert second date tips says to ignore timing rules (Bumble, undated). The first-date version, with the research on kissing, is on our kiss on the first date page.

What to talk about on a second date

The research here is about conversation in general, and it points at questions. Across three studies of live conversations, people who asked more questions, follow-up questions especially, were better liked. In face-to-face speed dating, people who asked more follow-ups got more agreement to a second date. Participants did not see any of this coming (Huang et al., 2017).

A second date has something the first one lacked: last time. A question about something they told you a week ago is a follow-up with a delay, and follow-ups went with being liked (Huang et al., 2017).

The ten second date questions below are ours, not tested at scale. They run from easy to closer, on the logic of Aron et al. (1997), where self-disclosure escalated a step at a time. That study was a 45-minute lab task in which previously unacquainted pairs took turns with questions that grew more personal, and it produced more closeness than small talk did. Closeness, not love. Whether the pairs were matched, whether they expected to like each other, and whether they were given a goal of getting close produced no significant closeness effects (Aron et al., 1997). The opening round is in our first date questions.

Callbacks (1–3). The details are invented; swap in theirs. A callback is a follow-up by definition, and follow-ups are what went with being liked (Huang et al., 2017).

1. So did Thursday’s interview go the way you wanted?
2. Did you finish the book you were complaining about, or did it win?
3. Is your sister’s band still playing the wedding?

Specifics (4–7). A specific question tends to get a story, and a story has details to follow up on. That part is our reasoning, not a finding.

4. What does a normal Tuesday evening look like for you?
5. What’s the most you’ve spent on a hobby you then quit?
6. Who in your family are you most like, and do they know it?
7. Which place do you keep going back to, and why that one?

Closer (8–10). In Aron’s procedure, the personal questions came after the easy ones (Aron et al., 1997). Same order here, at a fraction of the depth.

8. What do people usually get wrong about you at first?
9. What are you trying to get better at right now?
10. What made you say yes to today?

How to ask for a second date

One experiment covers when to ask. According to a summary of the study, Teichmann and colleagues ran a pre-study with 100 people, then a main experiment with 543 adults in the US and UK. Participants read a first-date scenario in which the other person texted immediately afterward, the next morning or two days later. The next-morning text produced the highest intention to keep seeing that person; both the immediate text and the two-day delay scored lower. Women were more sensitive to the timing than men. The effect ran through how reciprocal and how reliable the texter seemed (Teichmann et al., 2026; PsyPost, 2026).

Our reading: it was a vignette, and participants never saw what the message said. The study speaks to when, not to what. Our guide on what to text after a first date takes the what from here.

On what, the evidence is thinner. In the emails of 105 online daters, people who discussed what they wanted in a partner tended to report a higher likelihood of a second date (Sharabi & Dykstra-DeVette, 2019). That is self-reported and correlational.

Who asks is where Mumsnet splits. In a 2014 thread, a poster had heard nothing 24 hours after a Saturday date; replies mentioned a Tuesday rule, and several told her to message him herself (Mumsnet, 2014). In a 2025 thread, both people had said they wanted another date before he went quiet; most replies said interested people act within a day or two, and some said to propose a plan yourself (Mumsnet, 2025).

Our view: the person who wants the second date asks for it, the next morning or close to it, with a day, a place and a time in the message. A plan gives the other person something specific to say yes to, and something just as specific to decline.

Three asks, ours, not tested at scale:

• “Last night was good. Minigolf on Saturday at 2? I’ll find the course with the windmill.” A day, a time and a detail they can picture, sent the morning after, which is when the experiment’s text did best.
• “I still owe you a rematch on the otter question. Thursday at 7:30, same bar?” A callback to the first date with the plan attached, and callbacks are follow-ups (Huang et al., 2017).
• “I’d like to see you again. The market on Sunday morning, then coffee?” Interest stated plainly before the plan. The timing study found that perceived reciprocity mattered, but its messages had no wording, so this is our guess.

For the texting around the ask, see our guide on how to flirt over text. If nobody asks at all, it can look like ghosting, going quiet instead of saying it is over, though after one date it can just as easily be two people each waiting for the other.

How long to wait between the first and second date

Three questions get filed under this one, and the evidence behind them is uneven.

The middle row answers a different question. Ramirez and colleagues studied a national sample of online daters and how much they talked before meeting face to face, not the gap between dates. Curvilinear means the link was not a straight line, and the abstract does not say which way it bent. It cannot tell you when to meet again.

The third row is opinion with numbers attached. Sara Tick, a therapist quoted by Today, recommends scheduling the second date within five days of the first (Today, 2024). Tawkify, a matchmaking company, says no more than one week (Tawkify, 2025). When we searched on October 4, 2026, Google’s AI Overview said within one to seven days and cited a Reddit thread and Tawkify. So the machine summary rests on a forum and a matchmaker.

Our view: within the week, and the plan matters more than the day. A date ten days out with a place and a time is a firmer yes than a vague “soon” sent the next morning. What the surveys say about the dates after this one is on how many dates before a relationship.

What a second date can and cannot tell you

A second date can show you how someone behaves twice. It cannot tell you whether the two of you will last, and one speed-dating study found that more than 100 self-report measures could not predict that either.

The clearest evidence comes from speed dating. Joel, Eastwick and Finkel gave a machine-learning model more than 100 self-report measures from speed daters whose dates lasted four minutes. It predicted 4% to 18% of how much a person desired others in general and 7% to 27% of how much others desired that person. It could not predict who would want whom in particular (Joel et al., 2017). Each person’s general pattern, how much they wanted others and how much others wanted them, was partly predictable. How two specific people would feel about each other was not.

Expectations and decisions can come apart. After six-minute speed dates, attraction and nonverbal immediacy, the closeness people signal through eye contact, posture and smiling, predicted how rewarding people expected further contact to be. They did not predict the decision to meet again (Houser et al., 2008).

A third study followed online daters into the first date. Sharabi and Caughlin surveyed 186 online daters before a first date and 94 after it. Whether the date went well could be predicted from features of the online impression and from how the relationship was developing (Sharabi & Caughlin, 2017). We read the abstract, which gives no effect sizes, so we cannot say how strong those links were.

Our reading: none of these studies looked at second dates directly, but they lean the same way. What people say about themselves predicts little about a particular pair, and first impressions shape expectations more than choices. One more date tells you more than any list. What it can show is behavior that repeats, the kind of thing a green flag is made of: they remembered what you told them, and they turned up when they said they would.

How to use this

1. One format, one question it answers. Pick from the table by what you want to know: whether the talk holds, how they lose at minigolf, what they choose when handed the pick. The venue is a detail, and so is the price; our date cost calculator splits it three ways.
2. A text the next morning with a day in it. In the one experiment on timing, the next-morning text did best, and both the immediate text and the two-day wait scored lower (Teichmann et al., 2026). A day, a place and a time give the yes something to attach to.
3. An hour you can leave after. That is the last column of the table. A format with an easy exit makes staying a choice, for both of you.
4. The third-date question, asked out loud. Near the end, or the next morning: “I’d like to do this again. You?” A clear yes is not a relationship, but it is a step out of the talking stage. Our view: treat a maybe with no day attached as a no.