Photos carry most of the signal
The classic large-scale evidence comes from economists Günter Hitsch, Ali Hortaçsu, and Dan Ariely, who analyzed the browsing and messaging behavior of thousands of users on a major U.S. dating site. Published in the American Economic Review in 2010, their study found that the attractiveness of the profile photo was the dominant predictor of who received first-contact messages — for both men and women, and by a wide margin over most stated attributes.
Text matters at later stages, but the photo decides whether later stages happen. In swipe-based apps, where the photo occupies nearly the whole screen, its share of the decision has only grown.
The Tinder reality: fast filtering at scale
In 2016, Gareth Tyson and colleagues ran curated test profiles on Tinder and measured what happened. The asymmetry was stark: female test profiles accumulated matches rapidly while male profiles gathered them slowly, and men compensated by liking far more profiles with far less deliberation.
Two practical consequences follow. Your photo is typically judged in a seconds-long, low-attention pass — consistent with the 100-millisecond first-impression research. And in the same study, profile presentation moved the numbers: test profiles with more photos earned substantially more matches than single-photo profiles, one of the clearest levers the data identified.
Same person, different outcomes
None of this would matter if your photos were interchangeable — but they are not. Research on face variability shows that different photos of the same person are judged about as differently as photos of different people, and profile-image studies show people systematically pick worse photos of themselves than strangers would pick for them.
Put those together with the dating data and the conclusion is direct: the gap between your best and worst candidate photo is large, it is invisible to you, and it translates into measurably different results on the other side of the swipe.
What this means for your lineup
Lead with a clear, well-lit solo portrait where your face fills a healthy share of the frame — the first photo does almost all the filtering work. Fill the remaining slots rather than leaving them empty, using them to add context: a full-body shot, an activity, a social scene as a later photo.
Most importantly, choose between candidates with evidence instead of instinct. Compare a shortlist side by side, get an outside signal — friends, or a consistent AI rating — and let the comparison, not familiarity, pick your opener.
The limits of this research
The big datasets are observational and platform-specific: they describe aggregate behavior on particular apps in particular years, not guarantees for any individual profile. Norms shift, algorithms change, and attractiveness ratings are only one channel — bios, prompts, and messaging still shape what happens after the match.
And as with every page in this research series: these studies measure perception and behavior, not your worth. The useful takeaway is simply that photo selection is the highest-leverage, most testable part of the profile — so test it.
Frequently asked questions
How many photos should a dating profile have?
More than one. In field data from Tinder, test profiles with three photos earned substantially more matches than identical profiles with a single photo. Fill your slots: lead with a strong solo portrait, then add variety — full-body, activity, social context.
What should the first photo be?
A clear, well-lit solo portrait where your face is unmistakable and fills a good share of the frame. The first photo does the filtering; group shots, sunglasses, and distant full-body photos belong in later slots if anywhere.
Does a better photo really change match results?
The evidence says yes. Photo attractiveness was the strongest predictor of first contacts in large-scale dating data, and different photos of the same person are judged about as differently as photos of different people — so moving from your weak candidate to your strong one is a real, measurable change.
Can an AI rating tell me which dating photo to use?
It gives you a consistent comparison signal across your candidates — useful precisely because people are poor judges of their own photos. It is not a guarantee of matches; treat it as evidence for choosing between your own options, alongside your judgment of context.
References
- Hitsch, G. J., Hortaçsu, A., & Ariely, D. (2010). Matching and sorting in online dating. American Economic Review, 100(1), 130–163.
- Tyson, G., Perta, V. C., Haddadi, H., & Seto, M. C. (2016). A first look at user activity on Tinder. IEEE/ACM ASONAM 2016.
- Jenkins, R., White, D., Van Montfort, X., & Burton, A. M. (2011). Variability in photos of the same face. Cognition, 121(3), 313–323.
- White, D., Sutherland, C. A. M., & Burton, A. L. (2017). Choosing face: The curse of self in profile image selection. Cognitive Research: Principles and Implications, 2, 23.