Why We Started Asking Whether People Actually Met Again

When we began thinking seriously about adult friendship, it was easy to look for familiar product signals: did someone sign up, show up and enjoy an activity? Those signals matter. They can also create a misleading sense of progress. A great first evening is not the same as a developing friendship. The deeper question for g3t became whether people found a natural way to see one another again—and whether that happened because both people wanted it.

The tempting simplicity of event attendance

Attendance is observable. A booking exists, people arrive and a record can show that an event happened. Product teams can build dashboards around those facts. But attendance alone says little about the relationship that follows. A person might love the restaurant, dislike the social situation or enjoy one particular conversation but never meet the person again. Another might quietly build a friendship without rating the activity as spectacular. Treating a social product like an events business therefore risks optimizing the easiest part to count.

What our earlier work made difficult to ignore

Our former food-centered product, thinnan, revealed practical problems with planning and repeat encounters. The g3t direction emerged from the question of why an introduction sometimes fails to become a second meeting. Internal interviews suggested that even interested people can hesitate to initiate or coordinate. Those observations influenced our product thinking, but they are not population-level research claims and should not be used as published percentages without approved methodology and consent checks.

The shift in framing matters more than the size of any internal number. Instead of assuming that a successful first encounter will naturally lead to another, we began asking what stops a second plan: unclear interest, scheduling friction, the fear of being rejected or simply not knowing whether another person wants the same thing. Those are different problems. A useful product needs to distinguish them instead of treating all drop-off as lack of chemistry.

Why even a second meeting is not the finish line

Two people can meet twice and remain acquaintances. A third encounter can reveal new parts of a relationship: whether conversations continue naturally, whether plans are reciprocal and whether meeting starts to fit into ordinary life. That does not mean three meetings 'make' a friendship. It means third encounters could be an informative product outcome if the measurement is accurate. We should not publish an internal goal as though it were observed friendship formation.

Designing measurements that respect people

There is a difficult privacy boundary in any service concerned with who likes whom. People need room to express interest without exposing another person's private choices or movements. A product should avoid turning social preference into a public ranking. It should also distinguish a notification sent from a meetup actually attended. We can count a reconnection opportunity, a reservation, a confirmed attendance and a mutually desired repeat encounter as different stages—not collapse them into one feel-good success metric.

The right question for a founder dashboard

Instead of asking only 'How many people came?', we want to ask 'Did the product help people keep an interaction going when they otherwise might have lost it?' Answering that rigorously requires definitions, consent, reliable attendance records and enough time to observe repeat behavior. It also requires acknowledging that some people simply will not click. Neither the product nor the founders can manufacture mutual friendship, and the promise should never be that they can.

The principle we want to earn

A useful social product should make repeated encounters easier, more respectful and less logistically demanding. It should not declare friendship achieved because two profiles matched or because a member bought a ticket. This is an editorial account of a founder hypothesis and research direction, not a claim that a particular feature has launched or that its effectiveness has been proven. The real evidence will come from what members choose to do again.

A more demanding definition of success

For a social product, counting signups or event tickets is relatively straightforward. Counting friendships is not. We can observe that two people attended one event, but we cannot infer the quality of their relationship from that record alone. Did they enjoy each other's company? Did they choose to meet again? Did a second meeting actually happen? Were they comfortable enough to make future plans? Those questions require more care than a dashboard showing increasing registrations.

This is why our editorial approach distinguishes product hypotheses from outcomes we have verified. We can describe the question we are exploring—how to make follow-up meetings easier—without claiming it has been solved. A user is not evidence of friendship merely because they paid for an experience. Any future public claims about repeat attendance, satisfaction or connection should be accompanied by their measurement period and method. Our learning process should remain visible, including when a promising mechanism has not yet produced the intended behavior.

A responsible public account of building g3t would explain what we can measure at each stage and what remains uncertain. Showing up twice is an observable behavior, but it is not an automatic claim about emotional closeness. A mutually chosen reconnection is a stronger signal of interest than an algorithmic suggestion, but it still may not become a lasting friendship. Even an interview report that sounds compelling has selection limits and can reflect the kinds of people willing to speak with a startup. Founder storytelling becomes useful when it shares these constraints rather than turning every lesson into a triumph. This story remains subject to founder approval before publication.

Sources and verification

How many hours does it take to make a friend? — Hall

Are You Your Friends’ Friend? — Almaatouq et al. (2016)

References are provided to support the specific research or public context discussed above. Correlation does not establish causation, and no meeting or app can guarantee friendship.

For another starting point in Paris,explore g3t’s current Paris page. Check the live details there before making plans; this article does not imply that an upcoming feature has launched.