# The photo that won wasn't the one the best girls liked. **By André de Almeida** · 2024-08-04 Most advice on the best photos for dating apps is guesswork. I ran another experiment. During COVID, bored and stuck inside, I decided to shave off a beard and a head of hair I'd had for years, as one does. Before I did it, I wanted to use the occasion to compare different looks on Tinder and get three photos out of it, not one: full hair and full beard, bald with the beard kept, and bald with everything gone. I ran all three on Tinder, at the same time, on three separate Tinder Gold profiles, and counted who liked which one. ### [![StoneBone drive and libido support](https://cdn.shopify.com/s/files/1/0572/4333/6764/files/StoneBone_Two_Tablets_Drive_and_Libido_Support_from_Marcos_Grounds_1.webp?v=1784724485)](https://marcosgrounds.com/products/stonebone) ### Table of contents. - [The setup.](#the-setup) - [Why total likes is the wrong number to start with.](#why-total-likes-is-the-wrong-number) - [The result: hair and beard won on volume — barely anything else.](#the-result-volume) - [The result that actually matters: who was doing the liking.](#the-result-that-actually-matters) - [Attention isn't attraction — even on your own face.](#attention-isnt-attraction) - [What this means for you.](#what-this-means-for-you) ### ### The setup. Three Tinder Gold accounts, same face, one photo each — back then it was possible to have profiles with only one pic on Tinder. Gold matters for one reason: it shows you who liked you before you swipe back, not just the number of profiles who liked you, so you can actually count and assess. All three accounts went up in London. I wasn't there. I picked it as a cosmopolitan baseline. And it's also big enough so the likelihood of girls seeing all three profiles is low. All three accounts were built at the same time, one after the other. The only thing that changed was the picture. ### Why total likes is the wrong number to start with. For this piece I'm using two tiers: what I call tier 2 or T2 is a regular hot girl — attractive, not average, but also clearly not tier 1. Tier 1 or T1 is a genuinely super hot girl, i.e., Instagram influencer level. Below both is everyone else: girls I wouldn't personally be interested in talking to. It's the same way I split responses when I tested [body fat percentage against the same question](https://marcosgrounds.com/blogs/news/the-girls-couldnt-tell-the-difference) — different data set, same reason for using it. Total likes tells you how many people found a photo good. It says nothing about who those people were, or how many of them you wouldn't have swiped right on either. This time the source is different from the body fat piece in a useful way. Those were eight T1 girls I know personally. This is the opposite: total strangers on a dating app. ### The result: hair and beard won on volume — barely anything else. Full hair and full beard pulled 47 total likes in 24 hours. Clean-shaven with shaved hair pulled 24 in the same time period, roughly half. Beard with no hair pulled 12, also in the same 24 hours — the lowest of the three. ### ### The result that actually matters: who was doing the liking. Split by tier: Clean-shaven with shaved hair: 7 T2, 5 T1, 12 below threshold. Half the likes on this photo came from T2 or T1. Half came from girls below the tier I'd even count. Beard, no hair: 1 T2, 0 T1, 11 below threshold. This photo was bad across the board — fewest total likes, weakest tier mix of the three, almost entirely noise. Full hair and full beard: 4 T2, 2 T1, 41 below threshold. This is where the volume came from. 87% of this photo's likes were below threshold. The T1 count is real — 2 — but it's less than half of what the clean-shaven with no beard photo pulled, from a photo that got roughly double the total attention. There's a second thing worth naming here, because it's practical, not just interesting: the clean-shaven with no beard didn't just pull a better tier mix — it pulled less noise to sort through in absolute terms. 12 below-threshold likes against 41. Less than a third of the clutter, for a better Tier 1 result. That's not a tradeoff between quality and effort. Here, the same choice won on both. I don't know why the beard with no hair photo cratered the way it did, across every single metric. I have no clean mechanism to offer you. I have the numbers. ### Attention isn't attraction — even on your own face. A "hot girl" stock photo stops the scroll on an ad and still doesn't sell anything. Total Tinder likes work the same way. Full hair and full beard bought the most attention, and most of it came from below the tier that was actually worth having. That's not a win. That's a big number built on the wrong audience. The clean-shaven with no beard photo did the opposite: fewer girls liked it, but a much larger share of the ones who did were exactly the ones worth having a look at. ### What this means for you. Not "go shave your head." What actually transfers is the method: if you're testing anything about how you look, don't stop at "which one got more attention." Ask who gave it, and how much of it you'd have to sort through to find them. For instance, that's why I never use services like Photofeeler. I have no idea who these girls are and therefore I don't know how much weight I can put on their opinions. This is exactly why I tested [body fat percentage against the same question](https://marcosgrounds.com/blogs/news/the-girls-couldnt-tell-the-difference) with Tier 1 girls only. Because I know the quality of the girl giving the opinion and I know if I want to consider it or not. This is also my point of contention with studies on "girls'" preferences. First, it's self-reported in a way that often pressures the girls into giving the opinion they find socially acceptable. Second, and extremely important, I don't know what the girls look like. I don't even know if they have all their teeth. So, putting weight on this kind of data is difficult for me. Anyway, the Tinder experiment costs almost nothing to run yourself — a few photos, a couple of Gold subscriptions for a week, and the discipline to actually look at who's liking you instead of just how many. However, now I don't think Tinder lets you upload only one picture. Also, I haven't read the Terms of Service lately, but I think they disallowed multiple accounts per user a while back. Also, if you want to go further on the face and grooming side of this, that's what [Looksmaxxing for men: maximum ROI](https://marcosgrounds.com/blogs/news/looksmaxxing-for-men) covers. Same reversal, different data set, as the [body fat experiment](https://marcosgrounds.com/blogs/news/the-girls-couldnt-tell-the-difference). Most content built for "girls" in general is really talking about the middle. The top doesn't work the same way. Worth remembering next time a number looks good. --- > Source: [Marco's Grounds](https://www.marcosgrounds.com/blogs/news/best-photos-for-dating-apps)