The Real Numbers Behind Modern Dating (2024–2025)

On apps: the gap between usage and success

The 2024 Hinge annual report — which Hinge has a commercial interest in framing optimistically — still contains a striking data point: the majority of its active users in their sample had not been on a date through the app in the previous three months. They were using the app. They were not, by this metric, dating through it.

Match Group's broader data tells a similar story from a different angle. Across their platforms, they track what they call conversion: the rate at which matches become conversations, conversations become scheduled meetings, meetings become second meetings. Each step loses a significant percentage of the original group. By the time you reach a third in-person date, you are looking at a small fraction of the starting pool.

None of this is hidden. The companies publish it in forms that allow them to claim progress — more first dates than last year, better conversation rates in this demographic — while the underlying picture remains: app-based dating produces far fewer actual relationships per user-hour invested than the apps' marketing suggests.

Who is actually meeting people offline

Pew Research's 2023 dating survey found that among adults who had started a relationship in the previous five years, apps and online platforms accounted for about 22 percent of how they met. The most common method was still through friends or family — about 29 percent. Work and school combined for another 18 percent. Apps are significant but not dominant, which is not what you'd conclude from the cultural conversation around them.

The demographic breakdown matters. For adults under 30, apps are more common than any other single method — but still account for under a third of new relationships in the Pew sample. For adults over 45, they drop sharply: under 15 percent. The idea that modern dating is primarily app-mediated is roughly accurate for one age cohort and significantly overstated for most others.

What doesn't show up clearly in any of the major surveys: the hybrid path. A large number of people who describe meeting through a mutual friend or at a social event made initial contact or had significant early communication digitally. The categories that made sense in 2005 are fuzzier now. Meeting someone 'in real life' increasingly just means the first face-to-face encounter happened offline, even if everything before it was on a screen.

The loneliness numbers

The Surgeon General's 2023 advisory on loneliness gets cited a lot, usually for the headline statistic that roughly half of American adults report measurable loneliness. What gets cited less is the breakdown by relationship status and age.

Young adults — ages 15 to 24 — reported the steepest increase in social isolation over the past two decades, steeper than any other age group. This is the cohort that grew up with smartphones from early adolescence and entered dating life already accustomed to digital-first social interaction. Whether the technology caused the isolation or the isolation accelerated technology adoption is a causal question the data doesn't resolve. What it does show: the assumption that digital connection is an adequate substitute for in-person contact has not held up at the population level.

For what it's worth, the loneliness data does not show that partnered people are uniformly less lonely than single people. About a third of people in committed relationships report significant loneliness. The variable that tracks most tightly with connection quality is not relationship status but relationship satisfaction — which is a different thing. A bad relationship at close quarters is lonelier than chosen solitude.

What the 'dating is worse than ever' narrative gets right and wrong

The cultural consensus — stated explicitly across dozens of opinion pieces and implicitly in the framing of most dating-app coverage — is that something has gotten measurably worse about forming relationships over the past decade.

The evidence for 'worse' is real. Marriage rates continue to decline. Age at first marriage continues to rise — currently 28.2 for women, 30.0 for men, per US Census Bureau 2023 data. The share of adults who are not in a partnership has increased. The Pew loneliness data points in the same direction.

The evidence against 'worse than ever' is also real, and less discussed. Long-term relationship satisfaction, among people who are in relationships, has not declined in the survey data. The relationships that do form appear to be, on average, more chosen and less driven by social obligation than historical baselines. The researchers who track intimate partner violence consistently find it has declined over the same period that relationship formation has become harder. The median relationship that forms today is more equitable in its internal structure than the median relationship of 40 years ago.

The honest picture: forming relationships has gotten harder for some people, particularly younger adults; the relationships that do form are often better than previous eras suggested was possible; and the distributional question — who gets access to good relationships and who doesn't — has become more visible even if not necessarily worse.

The metric no one is tracking that matters most

All of the above data measures things that are relatively easy to count: relationship status, app usage, dates per month, loneliness self-report. None of it measures what I'd actually want to know, which is something like: among people who are actively trying to form a relationship, how many are making decisions based on an accurate understanding of the other person?

Anecdotally — based on what people describe in clinical research, in surveys, in the qualitative parts of the data — the answer seems to be: not many, in the early stages. Early dating tends to involve a process of projection that is at least partially replacing actual perception. You're often in a relationship with the version of the person your attachment needs require, not the person they actually are.

No metric captures this well. But it might explain more variance in relationship outcomes than any of the variables that do get measured. The right question might not be 'why is forming relationships harder' but 'what would it mean to form them more accurately' — and that's a question the surveys haven't figured out how to ask.