How couples meet statistics: what the 39% really counts, and why only 10% met online
The 39% figure covers new heterosexual couples in 2017; across existing US couples, only 10% met online. Here’s what the denominator changes.
You are about to write “39% of couples meet online.” Pause before treating that number as a personal forecast: it counts a very specific group of new heterosexual couples, not every couple already together.
The clean answer is this: online became the biggest single way new US heterosexual couples met around 2013, but the latest public data still show that most recent couples met another way. Across all existing partnered adults, only 10% met their current partner on a dating site or app.
Key numbers
- 39% — US heterosexual couples who met in 2017 and reported meeting online, using a smoothed PNAS estimate. Rosenfeld et al.
- 10% — partnered US adults whose current partner was met on a dating site or app in Pew’s 2022 survey; our HCMST computation also gives 10.0% for existing couples. Pew Research Center
- 33% → 20% — the smoothed share of heterosexual couples who met through friends in 1995 versus 2017. PNAS
- 70.0% — different-sex couples who met in 2014–2017 and were not coded as meeting online, our weighted computation, n=359. HCMST 2017 data
- About 89% — online-met heterosexual couples in HCMST 2017 who were strangers beforehand. PNAS
- 1.10 — the adjusted breakup odds ratio for online-met couples after controlling for relationship duration; not significant. Rosenfeld 2017
- 16.1% vs 3.5% — partnered adults aged 18–34 versus 65–74 who had met their partner online in our HCMST 2017 computation. HCMST 2017 data
Table of contents
- How do most couples meet today?
- What percentage of couples meet online?
- Did online dating replace meeting through friends?
- Does age or sexual orientation change the odds?
- Do couples who meet online break up more?
- Where do the famous numbers come from?
- Is it the same outside the US?
- How to cite this page
- The bottom line
- Sources
How do most couples meet today?
The latest high-quality US relationship data do not measure 2026 directly. They show that online became the most popular single route for new heterosexual couples around 2013, while a weighted calculation from the public HCMST 2017 data finds that 70.0% of different-sex couples who met in 2014–2017 were not coded as meeting online.
“Most popular” means the largest individual category, not a majority. In the PNAS study, friends fell from 33% among couples who met in 1995 to 20% among those who met in 2017, while online rose from 2% to 39%. These are smoothed estimates for heterosexual couples, not a count of all relationships alive today. Read the underlying study.
Our raw calculation uses the public data, the w1_weight_combo weight and a 2014–2017 bin. It is lower at about 30% because it averages several rising years and does not combine the 2009 wave with the 2017 wave.
Caveat: HCMST asks people to describe how they met in an open-ended story that researchers later code, so category boundaries are not perfectly mechanical.
What percentage of couples meet online?
The answer depends on whether you mean new couples, existing couples, dating sites and apps only, or the wider Internet. 39% is the flow; 10% is the stock.
| The number | What it actually counts | Source |
|---|---|---|
| 39% | US heterosexual couples who met in 2017; a smoothed PNAS point value, not all couples | Rosenfeld et al. |
| ~30% | Our weighted raw HCMST 2017 computation for different-sex couples who met in 2014–2017, n=359 | Stanford dataset |
| 10% | All partnered US adults whose current partner was met on a dating site or app; our HCMST stock computation also gives 10.0% | Pew · HCMST |
| 12% | US adults who had ever married or had a committed relationship with someone first met on a dating site or app, including past relationships | Pew 2020 |
| More than one-third of marriages | Marriages formed from 2005–2012 in an eHarmony-commissioned opt-in panel; online includes more than dating sites | Cacioppo et al. |
The 39% estimate uses a five-year moving average for online meeting and pooled HCMST waves, so it smooths a rising curve. The raw public-data calculation is not a correction to 39%; it answers a different question with a different window and weight.
The HCMST coding includes dating sites and apps, social networks, games, chat and other Internet sites. Phone apps accounted for at least half of the online growth from 2010 to 2017, but not all of it. Our online dating statistics for men goes deeper into the app-specific denominator.
Caveat: these percentages should never be stacked into a pie chart; couples can be coded in more than one meeting category.
Did online dating replace meeting through friends?
Mostly, yes. The share of heterosexual couples meeting online rose as introductions through friends and family fell, and the authors describe online dating as displacing those routes.
| Way of meeting | Couples who met in 2017 | Couples who met in 1995 |
|---|---|---|
| Online | 39% | 2% |
| Friends | 20% | 33% |
| Family | 7% | 15% |
| Coworkers | 11% | 19% |
| Bar or restaurant | 27% | 19% |
| Primary or secondary school | 5% | 10% |
| Church | 4% | 7% |
| Neighbors | 3% | 8% |
The apparent post-2010 rise in bar and restaurant meetings came entirely from couples who met online and then had their first in-person meeting there; excluding online couples, the bar share was 16.9% in our 2014–2017 calculation versus 24.1% when they were included.
The social handoff is visible in other details: among online-met heterosexual couples, introduction by a friend or third person fell from 11.2% in HCMST 2009 to 3.7% in HCMST 2017, and about 89% were strangers beforehand. For app-specific context, see dating app statistics and does online dating work for average guys.
Caveat: the table’s categories overlap, so the columns do not add to 100%.
Does age or sexual orientation change the odds?
Yes, sharply—but denominators matter. In the HCMST 2017 stock of existing partnerships, 16.1% of partnered adults aged 18–34 had met their partner online, compared with 3.5% of those aged 65–74; that mostly reflects when those couples formed, not older people’s current performance on apps.
Pew’s 2022 survey found that 20% of partnered adults under 30 met their current partner on a dating site or app. Lifetime use also varied from 53% among adults under 30 to 13% among adults 65 and older. Those are current-partner and lifetime-use measures, not interchangeable “success rates.”
Our weighted HCMST calculation puts online meeting at 51.9% for same-sex couples who met in 2014–2017, n=76, and 50.4% for those who met in 2010–2017, n=127. The published PNAS paper does not report a same-sex percentage; the often-cited 65% appears in an author draft, not the published article. The smaller samples make “about half” the responsible summary.
If apps feel like a wall, you are not imagining it: Pew found 64% of men who used dating platforms in the past year had felt insecure about the lack of messages. Still, most couples meet through people and places, and where to meet women maps those routes.
Caveat: sexual identity and couple type are not identical measures; Pew’s LGB category includes bisexual people in different-sex couples.
Do couples who meet online break up more?
The strongest US longitudinal answer is no demonstrated difference once relationship duration is accounted for. In HCMST 2009–2015, the adjusted breakup odds ratio was 1.10 and not significant; the apparent raw gap came largely from online couples being newer relationships.
Online-met couples had been together an average of 4.4 years at the first HCMST wave, versus 19.1 years for offline couples. Younger relationships have had less time to survive, so comparing their raw breakup rates—9.5% per year versus 3.9%—without duration control is misleading. Among couples who met in 1995 or later, the comparable figures were 7.0% online and 7.3% offline, statistically indistinguishable.
The result is not proof that meeting online causes stable relationships; it shows no demonstrated breakup difference once duration is accounted for. In Switzerland, couples who met on dating apps reported no difference in relationship or life satisfaction.
Caveat: relationship quality and breakup studies measure outcomes after people meet; they cannot tell one individual which channel will produce a relationship.
Where do the famous numbers come from?
Most viral numbers are not fabricated from nothing; they are stripped of their denominator, population or sponsor.
The 39% figure is routinely changed into “39% of all couples” or “39% met through dating apps.” Neither is what the published PNAS study says. It covers heterosexual couples who met in 2017, and “online” includes social networks, chat, games and other Internet routes.
The “one in three marriages start online” line comes from Cacioppo and colleagues’ survey of 19,131 US respondents who married between 2005 and 2012. Its online category was broad, and only 45.01% of online meetings were through dating sites. eHarmony commissioned Harris Interactive to run the survey; the paper discloses that the lead author was a scientific advisor, the second author was his spouse and the third had been eHarmony Labs’ director. The authors themselves described the effect as quite small. News coverage reported a $130,000 payment to field the survey; the paper does not state that amount.
The “one in five relationships” figure came from a 2010 Match.com press release based on industry-commissioned research restricted to adults with at least $30,000 in annual income and Internet access. It also claimed one in six marriages. The familiar “40 million Americans use online dating” line has no survey behind it in the source commonly repeating it; Pew reports shares, such as 30% who had ever used a dating site or app, not a headcount.
Caveat: sponsorship does not by itself settle whether a finding is right; it does tell readers how cautiously to label the result and what population it represents.
Is it the same outside the US?
No single global percentage travels cleanly. In Switzerland’s 2018 national survey, among 3,245 partnered people who met their partner in the previous ten years, 4.12% met through dating apps, 7.33% through dating websites and 5.24% through other online services—about 17% online across that window.
A 2025 European conference paper found online shares of 12–20% for first co-residential partnerships in several countries, around a quarter in France, the UK, the Netherlands and Norway, and 40% in Finland. That is a different population and a conference paper, not a peer-reviewed journal article. Its separation findings were mixed internationally, with higher risk for online-dating couples in half the countries studied.
Caveat: country comparisons mix survey years, relationship definitions and time lags between meeting and moving in.
How to cite this page
Suggested citation: Real World Appeal. “How couples meet statistics: what the 39% really counts, and why only 10% of couples met online.” realworldappeal.com, September 2026. https://realworldappeal.com/blog/how-couples-meet-statistics
The primary studies and the public HCMST dataset are linked so readers can cite them directly. Sources were retrieved on 13 September 2026; the HCMST figures marked as our own are weighted proportions computed from the public data using w1_weight_combo.
The bottom line
For new US heterosexual couples, online is the largest single route, not a majority route. For all existing partnered adults, dating sites and apps account for 10% meeting their current partner. The right number depends on whether you are measuring a recent flow or an accumulated stock.
39% is the flow; 10% is the stock. If you use apps, the first photo opens the conversation; Real World Appeal’s free photo read shows its first impression before payment, with the clear caveat that it is not a clinical or scientific instrument.
Sources
- Rosenfeld, Michael J., Reuben J. Thomas, and Sonia Hausen. 2019. “Disintermediating your friends: How online dating in the United States displaces other ways of meeting.” PNAS 116(36):17753–17758. Linked source
- Rosenfeld, Michael J., Reuben J. Thomas, and Sonia Hausen. 2023. How Couples Meet and Stay Together 2017-2020-2022 combined dataset. Stanford University Libraries. Linked source
- Pew Research Center. 2023. “From Looking for Love to Swiping the Field: Online Dating in the U.S.” Report, February 2, 2023. Linked source
- Pew Research Center. 2020. “The Virtues and Downsides of Online Dating.” Report, February 6, 2020. Linked source
- Rosenfeld, Michael J. 2017. “Marriage, Choice, and Couplehood in the Age of the Internet.” Sociological Science 4:490–510. Linked source
- Cacioppo, John T., Stephanie Cacioppo, G. Gonzaga, Elizabeth Ogburn, and Tyler VanderWeele. 2013. “Marital satisfaction and break-ups differ across on-line and off-line meeting venues.” PNAS 110(25):10135–10140. Linked source
- WTOP. 2013. “Study: 1 in 3 marriages start online.” Web news article, June 2013. Linked source
- Match.com / Chadwick Martin Bailey. 2010. “Studies Find One in Five New Relationships and One in Six New Marriages Are Now Between People Who Met Through an Online Dating Site.” Press release, April 20, 2010. Linked source
- Potarca, Gina. 2020. “The demography of swiping right. An overview of couples who met through dating apps in Switzerland.” PLOS ONE 15(12):e0243733. Linked source
- Perelli-Harris, Brienna, Biao Kuang, and Ann Berrington. 2025. “How we met: the demography of online and offline partnerships.” International Population Conference 2025 paper. Linked source
- Dating Sites Reviews. Undated, accessed 2026. “Online Dating Industry Statistics, Facts, and History.” Linked source
Frequently asked questions
How do most couples meet in 2026?
The latest high-quality US evidence covered here comes from a 2017 relationship survey and Pew’s 2022 survey, not a 2026 census of couples. Online was the largest single route for new different-sex couples, but most still met elsewhere; where to meet women covers the offline side.
What percentage of marriages start online?
There is no single current percentage that answers every version of that question. The widely repeated more-than-one-third figure concerns marriages formed from 2005–2012 in an eHarmony-commissioned study, while Pew’s 12% figure includes US adults whose past or current committed relationship began on a dating site or app; see online dating statistics for men.
Do relationships that start online last?
US longitudinal evidence found no significant breakup difference after relationship duration was controlled, with an adjusted odds ratio of 1.10. That does not prove online dating causes better or worse outcomes; does online dating work for average guys looks at the practical question.
Is it better to meet someone online or in person?
The data do not establish one universally better route. Meeting context matters less than the people, timing and relationship that follow, so use both channels and learn how to be more approachable in real settings: how to be more approachable.
Where do most couples meet besides apps?
Friends, family, coworkers, bars, school, church and neighbors all appear in the survey, and no single offline route accounts for most couples. Friends were still a major route in recent years; where to meet women turns that finding into practical options.