Somewhere between 10% and 60%, and I'd rather explain that range than pretend it's one number.
I went looking for a single figure and came back with about six, all published as fact, all measuring different people. Here are the three you'll see most often.
Three numbers, three different populations
- 10% of partnered US adults met their current spouse or partner on a dating site or appPew Research Center, survey fielded July 2022
- 39% of heterosexual couples who got together in 2017 met onlineRosenfeld, Thomas & Hausen, PNAS, 2019
- 27% of couples who married in 2025 met through a dating appThe Knot's annual newlywed survey
None of them is wrong. They count everyone who is partnered, versus couples who formed in one particular year, versus people who just got married. Which one you want depends on the question you're actually asking.
The headline numbers
These are the figures nearly every article on this subject is built from. I've put the population and the survey date next to each one, because that's where the confusion starts.
The same question, five different answers
Share who met their partner online or on a dating app, by the population each study measured.
These bars are not directly comparable and that is the point. Each one counts a different group of people over a different period.
39% of US heterosexual couples who met in 2017 found each other online
Source: Rosenfeld, Thomas & Hausen, Proceedings of the National Academy of Sciences, 2019
This is the most cited statistic in the field, from the How Couples Meet and Stay Together study run by Stanford sociologist Michael Rosenfeld.
Two things matter about it. It counts every kind of online meeting, including social media and forums, not only dating apps. And the survey behind it was fielded in 2017, which makes it nine years old. It's still the best academic measurement available, and it's also the number most likely to understate where things stand now.
Same-sex couples came in at roughly 60% to 65%
Source: same study
The gap is large and it isn't an accident. Rosenfeld's argument is that the people who benefit most from online dating are the ones who were poorly served by the old routes of family, friends, church and neighbourhood. That describes the LGBTQ+ population directly, and the data reflects it.
10% of partnered US adults met their partner on a dating site or app
Source: Pew Research Center, 6,034 US adults, July 2022
Pew counts partnered adults as those who are married, living with a partner, or in a committed relationship, which was 69% of US adults at the time. The 10% works out to about 7% of all Americans.
This is the lowest headline figure and it doesn't contradict the others. It includes couples who met in 1985, when meeting online wasn't possible. It measures every relationship currently standing, not the new ones forming.
20% of partnered adults under 30 met their partner on an app
Source: Pew Research Center, same survey
Same survey, narrower population, and the figure roughly doubles. That's the clearest single illustration of why who is being counted matters more than the number itself.
27% of couples who married in 2025 met through a dating app
Source: The Knot Real Weddings Study
The Knot is a wedding company rather than a research institution, so I'd keep the source attached whenever you quote this one. It's also the only annual measurement of newlyweds anyone publishes, which makes it useful anyway.
Worth knowing how it was built: the study captured 16,956 US couples married during 2024, recruited by email from The Knot and WeddingWire's own membership. That is a survey of their customers, not a national sample, and people who plan a wedding through a wedding website are not a random slice of the country. Among the couples who met on an app, Hinge accounted for 36%, Tinder 25% and Bumble 20%.
Why these numbers disagree so much
Most articles list the figures above and stop, which leaves you with a range that spans six times over and no way to read it. Here's what's actually driving the spread.
"Met online" is a much wider category than "met on a dating app"
Stanford's 39% includes couples who met through social media, forums, online games, anywhere on the internet. The Knot's 27% counts dating apps and sites only. Roughly half the distance between those two figures disappears once you account for that alone.
The most quoted figure was collected in 2017
The 39% has been reprinted as a current statistic every year since 2019. It describes couples who formed in 2017, when Tinder was five years old and Hinge hadn't yet relaunched as a relationship app. Nobody has published an equivalent nationally representative academic study since, which is why it persists.
Newlywed surveys and population surveys count different people
A survey of people who married last year captures the newest relationships in the country. A survey of all partnered adults captures relationships formed across six decades. The first will always produce a higher online share, and neither is more correct than the other.
Two surveys in the same country can differ by ten points
Japan is the clean example. Meiji Yasuda Life's 2023 survey found 25% of couples who had married within the previous year met through a dating app. Recruit's Marriage Activity Survey found 15.4% for couples who married in 2022. Same country, overlapping period, nearly ten points apart, because the sampling and the question wording differ.
People who pay for dating apps report more success
Source: Pew Research Center, July 2022
Among partnered adults who had ever used a dating site, 52% of those who had paid for it met their current partner there, against 34% of non-paying users. Any survey that leans on engaged, paying users will produce a higher number.
How many people are using dating apps at all
Before couples come users, and the user numbers have started moving in a direction the industry didn't plan for. What follows is the survey side of that. The company side, where paying users are falling at both listed dating businesses while revenue per user rises, is in my online dating statistics page.
37% of US adults have ever used an online dating site or app
Source: SSRS Opinion Panel Omnibus, 2,012 US adults, January 2026, margin of error ±2.5 points
This is the most current nationally representative figure available, and it's the reason this page exists in the form it does. It sits above Pew's 30% from 2022, which is consistent with continued adoption across that period.
Only 6% are using one right now
Source: SSRS, January 2026
The gap between 37% ever and 6% currently is the most underreported statistic on this subject. The overwhelming majority of people who have tried online dating aren't doing it now, either because they found someone or because they stopped.
Most people who have tried online dating are not doing it now
Each dot is one in a hundred US adults, January 2026.
Source: SSRS Opinion Panel Omnibus, 2,012 US adults, January 2026.
Usage edged down between 2025 and 2026
Source: SSRS, January 2025 and January 2026
SSRS ran the same questions a year earlier and found 39% ever used and 7% currently using. Both slipped in 2026. The change sits inside the margin of error, so it's a signal rather than a finding, but it points the same way as falling paying-user counts at the largest apps.
More than six in ten LGBTQIA+ adults have used a dating app
Source: SSRS, January 2026
63% of LGBTQIA+ adults have used online dating, including 14% who are current users, against 34% and 4% for everyone else. It's the same divide that shows up in the couples data, measured at the other end of the funnel.
Usage is concentrated under 50
Source: SSRS, January 2026
51% of adults aged 18 to 29 and 53% of those aged 30 to 49 have used online dating, against 20% of those 50 and over.
Who has ever used a dating app
Share of US adults who have used an online dating site or app at any point.
Source: SSRS Opinion Panel Omnibus, January 2026. The 50 and over group is the only one where a large majority have never tried it.
How online compares with the ways people used to meet
The rise of online dating is only half the story. The other half is the quiet collapse of everything else.
Online overtook meeting through friends around 2013
Source: Rosenfeld et al., PNAS, 2019
The data shows online meeting crossing above friend introductions in the early 2010s and climbing since. It is now the single largest route into a heterosexual relationship in the United States.
In 2009, 22% of heterosexual couples had met online
Source: Rosenfeld et al., PNAS, 2019
22% in 2009, 39% in 2017, and low single digits in the 1990s. Very few social behaviours have changed that fast.
Meeting through family, church and the neighbourhood has been declining since 1940
Source: Rosenfeld et al., PNAS, 2019
This is the longer trend underneath the online numbers, and Rosenfeld's argument is that online dating didn't cause it. Adults became more mobile, married later, and built social circles that were more scattered and less overlapping. The pool of people who could plausibly introduce you to someone shrank first. Online filled the space that left behind.
Who is most likely to meet a partner online
The average hides differences big enough to make it close to useless on its own.
One in five partnered adults under 30
Source: Pew Research Center, July 2022
The highest of any age group, and roughly double the overall figure.
Roughly a quarter of partnered LGB adults
Source: Pew Research Center, July 2022
24% of partnered lesbian, gay and bisexual adults met their partner on a dating app, against 9% of straight partnered adults.
Close to zero among the oldest adults
Source: Pew Research Center, July 2022
Very small among partnered adults aged 65 and over, which is unsurprising when most of those relationships predate the internet entirely.
Online dating outside the United States
Almost every English-language article on this topic reports US figures and quietly presents them as universal. They aren't.
Around a quarter of married Japanese people under 40 met their spouse on an app
Source: survey reported by Newsweek
Japan is worth the space here because it's one of the few countries where the government publishes directly relevant data, so you can see the whole picture rather than one company's survey.
Japan recorded 474,717 marriages in 2023
Source: Japan Ministry of Health, Labour and Welfare
Down from 504,930 the year before, and the fewest since the second world war. The fertility rate stood at 1.2.
Two thirds of young Japanese adults say they don't know how to find a partner
Source: Japan Children and Families Agency, 18,000 respondents aged 15 to 39
67% said they didn't know how to go about finding a life partner, and 66% doubted they ever would. The government's response has included publicly funded matchmaking and, in Tokyo, a government-run dating app that asks for tax records to verify income.
It's the clearest case anywhere of online dating being treated as public policy rather than as a consumer product.
Mutual friends still dominate in parts of Europe and Latin America
Source: The Knot Worldwide, 2025 Global Wedding Report
Friends remain the primary route in much of Europe and Latin America, with 35% of Italian couples meeting that way. Industry-sourced, so keep the name attached.
What happens to couples who meet online
This is the question underneath the question, and it has a better answer than most people expect.
How a couple met doesn't predict whether the relationship lasts
Source: Rosenfeld et al., PNAS, 2019
The most useful finding in the whole literature and the one that gets quoted least. Rosenfeld's team found relationship success did not depend on whether the couple met online. Once a relationship starts, it takes on its own life.
You'll also see competing claims about divorce rates for online and offline couples. The most cited of those comes from a 2013 PNAS paper funded by eHarmony, which is worth knowing before you repeat the figure.
Meeting online produces more mixed couples
Source: Thomas, 2019, on online exogamy
Meeting online is associated with higher rates of couples crossing racial and educational lines than the routes it replaced. Friend and family introductions tend to reproduce the social group you're already in. Search doesn't.
Where this is heading
Two things are pulling in opposite directions, and it's worth watching which one wins.
The market is growing while engagement isn't
Revenue projections keep rising while paying user counts at the largest platforms have been falling. Match Group and Bumble both report these figures quarterly in their SEC filings, which is where to take them from rather than from secondary summaries, and I've pulled the current figures out of both.
Some of that drop-off is people pairing off and leaving, which is the whole point of the product. Some of it is fatigue. And a share of it is going somewhere none of these surveys ask about yet: AI companion apps, which have grown into their own category while the dating platforms have been shrinking. I test those separately, and they are not the same thing as dating, but they are competing for some of the same evenings.
The next academic measurement is overdue
The field is still working from a survey fielded in 2017. Until something comparable is published, every "percentage of couples who met online" figure in circulation is either nine years old or produced by a company with something to sell.
How this page was sourced
Every figure here is attributed to the organisation that collected it, with the date it was collected rather than the date it was published. Where two credible sources disagree, both are shown.
- Rosenfeld, Thomas & Hausen, Proceedings of the National Academy of Sciences, 2019: share of couples meeting online by year and orientation. stanford.edu
- Stanford University Libraries: How Couples Meet and Stay Together 2017, dataset and methodology. data.stanford.edu
- Stanford Report: interview with Michael Rosenfeld on the 2019 findings. news.stanford.edu
- Pew Research Center: From Looking for Love to Swiping the Field, full report, 6,034 US adults surveyed July 2022. pewresearch.org
- Pew Research Center: the who, where and why of online dating, partnered adults and paying users. pewresearch.org
- Pew Research Center: key findings about online dating, demographic breakdowns. pewresearch.org
- SSRS: The Public and Online Dating 2026, 2,012 US adults surveyed January 2026. ssrs.com
- SSRS: The Public and Online Dating in 2025, the comparison year. ssrs.com
- The Knot: Real Weddings Study, 16,956 US couples married in 2024, including how they met. theknot.com
- Newsweek: reporting on Japanese survey data and government marriage initiatives. newsweek.com
- NBC News and the Associated Press: Japan health ministry marriage and birth figures. nbcnews.com
Match Group and Bumble user and revenue figures are reported quarterly in their SEC filings and should be taken from there rather than from secondary summaries. Figures on this page reflect the most recent public data available as of the update date above. Sources occasionally differ by a few points depending on the survey, the population and the year.