Updated – 2nd July 2026

The dating sites promise a clever algorithm will find your perfect match. The science says something more interesting, and a good deal more useful.

Nearly every big dating brand has, at some point, promised that its special formula will do the hard part for you. Answer these questions, trust the algorithm, meet your soulmate. It’s a lovely pitch. It’s also, going by a sizeable pile of research, not how any of this really works. The better news is that what the science does say is more useful than the sales version, and it points to a handful of things you can actually do to give yourself a better shot. Let’s get into it.

First, how big is this now?

Online dating stopped being a fringe thing a long time ago. It’s now the single most common way couples meet. Around 39% of heterosexual couples who got together in 2017 met online (about 65% of same-sex couples), which was the first time in recorded history that any method overtook meeting through friends. More recent figures put it above half. Worldwide, something like 320 million people now use dating services, and around a third of US adults have tried one. So if you’re doing this, you’re not on the edge of anything. You’re using the front door most people now walk through.

Can a formula really match you?

Here’s the honest answer, and it’s worth the price of admission: no serious evidence says it can. In the most thorough review of the field, a team led by psychologist Eli Finkel concluded there is “no compelling evidence that any online dating matching algorithm actually works,” and that these formulas are unlikely to work even in principle. The reason is a bit awkward for the industry: algorithms match you on the things that are easy to measure, like similar personalities and shared attitudes, rather than the things relationship science has found actually predict a lasting couple.

If that sounds like old news, the follow-up is the clincher. A few years later, researchers fed more than 100 traits and preferences from real speed-daters into a machine-learning model and asked it to predict who would fancy whom. It managed two things: it could spot who tended to like lots of people, and who tended to be liked by lots of people. What it flatly could not do was predict whether two particular people would hit it off. The specific spark between you and someone else simply wasn’t there in the data before you met. A computer can narrow the room. It can’t pick the person.

Why the matching bit falls down

Three findings explain it, and each one is quietly liberating.

• Your “type” tells fibs. What people say they want in a partner barely predicts who they’re actually drawn to face to face. The checklist in your head is not the person you fall for.
• Similarity is overrated. Sharing hobbies turns out to be almost unrelated to whether two people feel a spark. “We both love hiking” is a nice coincidence, not chemistry.
• Chemistry is something that happens between you, in the room, in the timing, in the way a conversation bounces. A profile just can’t carry it. Finkel calls a dating profile “a feeble substitute for face-to-face contact,” and he’s right.

So why does online dating work at all?

Because the magic was never the matching. It’s the access. What online dating genuinely gives you is a far bigger pool than your street, your office and your mates’ barbecues ever could, and that matters most if you’re in what researchers politely call a “thin” market: dating later in life, looking for something specific, LGBTQ+, or simply new to an area.

And it does work. For all the burnout headlines, couples who meet online report relationship satisfaction that’s as high as, or a touch higher than, those who met the old-fashioned way, with much the same divorce rates. Meeting online isn’t a worse route to a good relationship. It’s just a different front door.

The two traps the “science” quietly sets

If the algorithms don’t do much, the design around them does, and not always in your favour. Two things are worth knowing.

Leagues are real (but reachable). When researchers mapped who messaged whom across four cities, a clear pecking order of desirability emerged. Nearly everyone, men and women alike, messages people about 25% more desirable than themselves, and your odds of a reply drop off the bigger that gap gets. That’s the actual maths behind “I never get any replies.” But don’t lose heart: roughly one in five of those hopeful messages up the ladder still got answered, the people who reach up tend to send fewer, longer, better messages, and the researchers noted that women in particular could afford to aim a little higher.

Endless choice makes you fussier, not happier. You’d think more options would mean more luck. The opposite happens. In one set of studies, the more profiles people saw, the more they rejected, with the odds of a yes falling about 27% from the first option to the last. Psychologists call it a “rejection mindset”: scroll long enough and you start closing off, growing pickier and gloomier as you go. The all-you-can-eat buffet doesn’t leave you satisfied. It leaves you unable to choose.

What the science says actually works

So if the formula can’t save you, what can? Reassuringly ordinary things, most of them.

Meet sooner, message less. Since chemistry only shows up in person, long messaging marathons mostly build an idealised picture that reality can’t live up to. Use the chat to get to a coffee, not to interview someone for a fortnight. If it helps, our piece on why a slightly nerve-racking first date beats dinner is really the same idea in action.

Loosen your filters. Because your stated type doesn’t predict attraction, ruthless dealbreakers mostly screen out people you’d have liked. Widen the net a little, keep an open mind, and yes, be a bit more aspirational than feels comfortable.

Date in small batches. To dodge the rejection mindset, don’t binge. A few considered conversations will always beat a hundred half-glanced profiles.

Sort your photos out. This is one place the folk wisdom holds up. Smile properly, look approachable, and lead with one clear, warm shot of just you, then add a couple that show you actually doing something. Group photos can help too, since they hint that other people enjoy your company, and there’s a curious wrinkle in the research: some studies find a man looks more appealing when other women are pictured smiling at him. We asked Saskia from the dating-photography company Hey Saturday for her rule of thumb:

“Your photos should be light, bright, in focus and have you as the most interesting thing in the shot. Ideally colour, composition and background should add value to your shots, ensuring they stand out from the sea of dating profile photos currently out there. Being captured doing something you love in one or two of the photos helps tell your story.”

Write to be easy and true. Professor Khalid Khan, who led one of the more careful reviews of what turns online contact into an actual date, put it well:

“Steer clear of fiction in your profile: written information could come back to bite you. And provide a 70:30 mix of who you are, and what you are looking for. Bear in mind that likeability is more attractive than academic achievement, and that a profile that appears genuine is more likely to generate interest.”

Two things follow from that. Keep it simple: research suggests we warm to words and information that are easy to read and process, so this is no place to show off your vocabulary. And if you’re funny, be funny, don’t announce it. Telling people you’re hilarious is a proven turn-off. Let them find out.

Go deep, not clever. The fastest route to feeling close to someone isn’t a slick opener, it’s honest back-and-forth. That’s the whole basis of Arthur Aron’s famous “36 questions”, which built real closeness between strangers through escalating, genuine self-disclosure. A question you actually want the answer to beats any chat-up line ever written.

The honest bottom line

Here’s what all the data really adds up to. An algorithm can hand you a stack of plausible candidates, but chemistry is discovered, not calculated, and it’s discovered in person. So the clever move isn’t to trust a formula that can’t deliver. It’s to use the thing online dating is genuinely brilliant at, access to people you’d never otherwise meet, keep an open mind, and get to the coffee.

That, as it happens, is roughly how Find Happy is built. We don’t pretend a magic formula has found your one true match. We give you a real pool of single people near you, then get out of the way and let you work out who you click with. No mystery science. Just a better front door, and your own good judgement, which was always going to be the deciding factor anyway.

Studies referenced: Finkel et al. (2012) and Joel, Eastwick & Finkel (2017) on matching algorithms; Rosenfeld et al. (2019) on how couples meet; Bruch & Newman (2018) on desirability; Pronk & Denissen (2020) on choice overload; Aron et al. (1997) on interpersonal closeness. Usage figures via Pew Research Center. Current at the time of writing.

Sources

• How couples meet: Rosenfeld et al., PNAS (2019): Stanford
• No evidence matching algorithms work: Finkel et al., Psychological Science in the Public Interest (2012): summary
• Machine learning can’t predict a match: Joel, Eastwick & Finkel, Psychological Science (2017): summary
• The desirability hierarchy: Bruch & Newman, Science Advances (2018): paper
• The rejection mindset: Pronk & Denissen, Social Psychological and Personality Science (2020): paper
• The “36 questions”: Aron et al., PSPB (1997): paper