We Had Real Women Rate Men's Dating Photos. Here Is What the Votes Show.
Early findings from our voter panels: the same man's trustworthiness swung 4 points between photos, and his most attractive photo was not his most trusted.
By John, founder of Enamor. Every photo insight here comes from panels of real voters rating real dating photos.
The same man, photographed in the same season, scored 4.2 out of 10 on trustworthiness in one photo and 8.2 in another. Nothing about him changed. Only the photo did. That 4-point swing is the single clearest pattern in our voter panel data so far, and it is why photo selection matters more than most men think.
Most dating photo advice is recycled opinion. We wanted numbers, so we are publishing ours: early findings from the photo panels we run for clients, updated quarterly as the dataset grows. The sample is small and we will say so plainly throughout. It is still more data than a guess.
How the testing works
Every photo we test is rated by 20 real voters recruited from the client’s target dating pool. Voters score each photo on three dimensions, attractiveness, trustworthiness, and intelligence, on a 1-10 scale, and can leave written comments. The photos in this post were all rated by women aged 25 to 34.
The dataset behind this post: 11 fully tested photos across two men, with more than 220 individual votes. One set is nine photos from a single client’s profile review. The other is the before-and-after pair from our founder’s own profile, the same two photos shown on our landing page. Small, real, and growing.
Finding 1: trustworthiness swung 4 points on photo choice alone
Here are the nine photos from one client’s panel, all rated by the same demographic in the same week. No photo or identity details, just the scores.
| Photo | Attractive | Trustworthy | Intelligent |
|---|---|---|---|
| 1 | 5.9 | 7.9 | 5.1 |
| 2 | 5.3 | 6.6 | 2.9 |
| 3 | 4.8 | 7.1 | 5.1 |
| 4 | 4.5 | 6.2 | 3.7 |
| 5 | 4.4 | 6.8 | 6.2 |
| 6 | 4.3 | 8.2 | 5.5 |
| 7 | 4.2 | 4.2 | 5.2 |
| 8 | 3.9 | 7.5 | 4.6 |
| 9 | 3.9 | 6.9 | 6.5 |
The spreads within this one lineup: attractiveness ranged 2.0 points (3.9 to 5.9), trustworthiness ranged 4.0 points (4.2 to 8.2), and intelligence ranged 3.6 points (2.9 to 6.5). Trustworthiness, the dimension most tied to getting from match to date, moved twice as much as attractiveness. Your face constrains your attractiveness score. Your photo choices control your trust score.
Finding 2: the most attractive photo was not the most trusted
Photo 1 scored highest on attractiveness (5.9) with a strong but not top trust score. Photo 6 won trustworthiness outright (8.2) while sitting near the bottom on attractiveness (4.3). Across all nine photos, the rank correlation between attractiveness and trustworthiness was 0.11, which is to say: essentially none. Voters were reading two different signals, not one.
The practical consequence: if you choose your lineup by asking “which photo do I look best in,” you are optimizing one dimension and leaving the other to chance. The photo that earns the swipe and the photo that earns the date may not be the same photo, and a good lineup needs both.
Finding 3: perceived intelligence more than doubled between photos
The same man scored 2.9 on intelligence in one photo and 6.5 in another. Nothing in a photo states your IQ; voters infer it from setting, dress, expression, and framing. A score that can more than double on photo choice alone is not measuring you. It is measuring the photo, which means it is fixable with a camera rather than a degree.
Finding 4: swapping one lead photo moved every score
The founder pair makes the same point from the other direction. Replacing one lead photo with a stronger one moved intelligence from 5.0 to 9.9, trustworthiness from 4.2 to 8.9, and attractiveness from 4.4 to 7.9. Same person, same week, one photo swapped. Those two photos, blurred, are the comparison on our how-it-works page.
What the voters said
Voters can leave written comments, and the comments in this dataset were short, polite, and specific. On a photo showing more skin: “Would prefer less skin showing.” On a posed shot: “It would have been better if the pose, posture and the angle is different.” On a low-scoring expression: “Would prefer a different expression.” None of this is cruel, and all of it is the kind of feedback a friend will not give you.
How does this compare with published research?
Our early numbers land where the published research points. Studies of dating photos have found that context drives trust scores (shirtless bedroom photos scoring 2.1 to 4.2 on trustworthiness while beach activity photos scored 8.8 to 9.4), and our lowest-trust photo followed the same pattern, drawing the “less skin” comment. Research on self-selection shows people pick photos of themselves that strangers rate as less flattering, which is the reason we test with voters instead of instinct. The near-zero relationship between attractiveness and trust echoes the finding that trustworthy-reading photos, not just attractive ones, are what convert matches into dates.
What this means for your profile
Three takeaways travel. First, do not pick your lineup by looks alone: attractiveness varied least of the three dimensions, and the scores you can move most are trust and intelligence. Second, your instinct about your own photos is the least reliable input available; the research on self-selection says strangers pick better than you do, and that is the norm. Third, small photo decisions carry real range: the difference between a 4.2 and an 8.2 was not a different face, it was a different photo.
If you want these numbers for your own photos, this is the exact panel we run for every client: 20 real votes per photo from your target dating pool, scored on all three dimensions, with a recommended lineup in your report. It starts at $59, and the intake takes about five minutes.
Limitations, stated plainly
This is a small dataset: 11 photos, two men, one voter demographic, more than 220 votes. We are not claiming statistical significance, and none of these findings should be read as universal laws. They are early, real numbers from identical testing conditions, published because the alternative in this niche is advice with no numbers at all. The dataset grows with every client, and we will update this post quarterly, including the findings that stop holding up. For the research-backed fundamentals that a larger evidence base does support, see our photo guide and men’s photo breakdown.
Frequently asked questions
Who are the voters rating the photos?
Real people recruited from the client's target dating pool. For the data in this post, every vote came from women aged 25 to 34. Voters rate each photo on attractiveness, trustworthiness, and intelligence on a 1-10 scale.
Is 20 votes per photo enough?
Twenty votes gives a stable directional read on how a photo is perceived, which is what photo selection needs. It is not a lab-grade sample, and we do not claim statistical significance. The point is comparing photos of the same person under identical conditions.
Can I get my own photos scored this way?
Yes. That is the service: every photo you submit is rated by 20 real voters from your target dating pool, and your report shows every score alongside a recommended lineup.
Will this data be updated?
Yes, quarterly. The dataset grows with every client, and we will update the findings and the tables in this post as the numbers accumulate.