What Is *Your* Best Profile Picture

May 6th, 2010 by Christian Rudder

One of our most popular posts ever was about profile pictures and what works and what doesn’t. For example, we examined why two very similar-seeming pictures of two very similar-seeming people often get very different amounts of attention:

As a follow-up, we’ve made something that can generate a personalized analysis of anyone’s profile pictures, something that can make a “blog post” just for you. It’s designed to help you choose your best pictures to put on your OkCupid profile, Facebook page, or whatever—sort of a more scientific (and more useful) Hot-Or-Not. We’re calling it MyBestFace.

It Will Analyze Your Pictures

MyBestFace is very simple: you upload photos of yourself; you contribute a few votes to the system; and after enough other people have done likewise, you get a report about your photos. The report is detailed and very awesome. I’ll show you mine.

It starts, naturally, by telling me which of my pictures was the most popular:

It also tells me what kind of people liked my photos: overachievers, deviants, jocks, whatever:

And I get a nifty map showing where all the people who voted for me are. It looks like I have heretofore undiscovered fan in the Philippines. Sweet.

My favorite part of the report is the end, where it shows the other photos mine were matched up against and exactly which people chose my picture as the best.

If you’d like to examine a complete example report, one of our beta-testers has generously agreed to let you guys browse hers. You can do so here.

Try It Out

MyBestFace is something we thought would be fun. We’re not thinking it’s gonna be huge or anything, but we hope it might be useful to anyone looking to do for themselves what we do here on the blog: figure stuff out. Upload a couple pics and play around with it. And let us know if anything seems busted! Besides your face! Burn!!!

104 Responses to “What Is *Your* Best Profile Picture”

  1. sweetglamour says:

    This looks like a lot of fun! I have a hard time choosing photos myself!

  2. Cynical_nerd says:

    This is awesome. Thanks!

  3. Yay says:

    Yay this is almost exactly what I wanted!

    The only difference is I don’t want a ranking of which of my pictures is liked by everyone. I want a ranking of which of my pictures is liked most BY PEOPLE THAT I LIKE. If I’m only attracted to punk rock girls, then the ranking of my pictures should only include the votes of punk rock girls. I don’t care if some old marm dislikes my picture. In fact, if someone I’m not attracted to likes a picture of mine, that should count against it.

    http://en.wikipedia.org/wiki/Collaborative_filtering

  4. John Fouhy says:

    I’m not convinced by your bar chart … If I look at it casually, I would think: “Liberals and geeks liked this picture the most; Christians and divas the least”.

    But if I’m reading it right, all you’ve done is put the biggest groups at the top. That is, you’ve got a lot of liberals and geeks on your site, so a lot of them have viewed your photo, so their bars are the longest. Their bars will always be the longest (I guess), so they’ll always be on top.

    But if you pay more attention to the relative lengths of the bars, it looks more like overachievers, Christians, and vegetarians were the most likely to prefer that photo.

    So … I’m not sure what information you’re trying to convey with the bar chart, but I don’t think it’s doing so effectively.

  5. Bryan says:

    MyBestFace let me know my current profile photo was universally disliked. Thanks for letting me know!

    On a side note; please tell me you’re keeping track of left vs. right preferences. There’s a chance some interesting data is hiding in there.

  6. champ23575 says:

    This looks cool, but it looks as though the pictures and reports are available to the public. Is there a way to restrict it only to other OkCupid users?

  7. browolf says:

    very cool & clever. it’s interesting how wildly different the demographics between different pics can be. we’ll be expecting a report on any new insights you get from this.

  8. cracktonite says:

    So the real question is whether the results of mybestface reinforce the previous findings of what good profile pictures are. Is it still a good idea to not make eye contact, take picture with pets, etc.? Regardless it’s a pretty cool tool to play with.

  9. Zoya says:

    I tried this but could only rate the photographs of men. Does this mean my photos will only be rated by men? Did I miss I bit where I could tick the gay box?

  10. magicaltrevor says:

    Interesting, but why can’t we just choose our already supplied profile pictures. Now I have to go dig them out off my hard driver, if I have them somewhere. Ugh.

  11. David says:

    This is a fantastic idea. I love that you guys add features and make sure they have some substance — it’s flash and bang, not just flash and fizzle.

  12. Also says:

    Also, on my imaginary site, I was imagining that users would be ranked on qualities, not just general looks.

    It would ask “Which of the two images is sexier?” “Which is classier?” “Which looks like more fun?” “Which looks more intelligent?” etc. Sexiest picture goes on the flirty dating website, while the classiest picture goes on the professional web bio or Facebook page.

    Also, picking the best of 4 users instead of 2 might speed up data gathering and make voting easier.

  13. gn says:

    sweet idea! but why don’t you integrate it with okcupid (and it’s security/privacy settings)?

    You should at least let people know that anything they submit is pretty much public aka ends up in Google and wherever. Just in case they end up seeing their photos in somebody’s profile next week (….can happen anyway) :p

  14. josh_i_am says:

    Wow, this is an insanely good implementation of this type of thing. The only issue is that there are so few users right now! I keep running out of people to choose from, but I want to stats-geek my photo collection to death!

  15. Sarah says:

    You *guys*. I don’t want to sign up for a dating site. Stop making me want so badly to join!

  16. Eidan_Yoson says:

    Thats a really good idea! Now i can use it to improve my image!

  17. KP says:

    This is interesting but you missed an opportunity to let people tag other people’s photo at the same time. Using tags you could have analysed why one picture was seen more favourably than other. Say, maybe you would have found that pictures with a tag “serious” and “sexy” are better than those with a tag “smile” and showing “teeth”. Who knows?

  18. imakite says:

    haha i would pick the author and have my pic posted on here. hahahah

  19. S says:

    You need to integrate it with OkCupid so ‘my profile pictures’ are one of the photo source options.

  20. science_pirate says:

    Looks fantastic, but it would be even better if one could see links to the profiles of people who have submitted the photographs (if the user wishes to enable this).

    I am biased in both directions here. I’ve seen some awesome photographs, and would love to learn more about the people behind them, and my own photographs are pretty darn incredible, and I wouldn’t mind the extra attention. :)

    As a possible feature, since presumably My Best Face users are already OKC members, a simple “Add from my OKC profile” wizard would save a lot of copying and pasting of URLs.

    Thanks again for a great tool,

    S.P.

  21. Daz says:

    Good Idea! It’s hard to choose if neither one is not your type lol

  22. err says:

    none of this changes the fact that if you have gained 80 pounds since your profile pictures were taken, that you should take new pictures.

  23. Tim says:

    Pretty cool, but has some serious flaws:

    – The number of votes I’m getting is statistically insignificant. I’d like it if I could get another round of votes on the same report.

    – Locality is as important to the results as personality. It would be great if you could restrict being voted on by users of a particular area.

    – I think it was brought up above, but I would like to be able to weight or restrict votes to those I’m actually looking to get dates from. Some girl I might think is a bitch shouldn’t be voting on my picture.

  24. jtm297 says:

    It seems like the votes from charity mode can actually mess up your report getting generated. For example, I did so much in charity mode, I ran out of people to rate. So then, when I wanted to submit a new report to get rated, there weren’t enough people to vote on, therefore I had to wait until enough reports were being made so I could appropriately vote. I believe the votes from charity mode should be transferred over when you create a new report. This is the only bug I saw.

  25. Gabriel says:

    I like the idea, but don’t understand why you don’t just have people choose between all of my photos. That would be so much more useful. When I look down the list of competitors that I “defeated” so many had bad light, not a clear shot of their face, etc. that it doesn’t seem a worthwhile comparison.

  26. alt says:

    This is neat, but as another commenter noted, as a female gay-tending bisexual I was disappointed to find I could only rate and be rated by guys. I even temporarily changed my profile to “gay,” but that didn’t change my results. What gives?

    That said, it is a very nifty and helpful thingamajig.

  27. amd says:

    Christian, your favorite part of this is my LEAST favorite. It would be great to be able to vote anonymously. I don’t mind providing a little data for you, but being personally identified, especially with “against” votes, is not cool. It’s like telling someone I insulted them. Please tell me there will be a way to turn that off.

    I agree with a couple others that it would be nice if we could easily import profile pictures.

  28. Brad says:

    I’m guessing this is some sort of Condorcet ranking going on in the back end? Schulze Method perchance? :)

  29. L Green says:

    Love this! But can you find a way to have older people rating older people’s pics? The 31+ designation here is mostly 31 and 32 year olds. I’m 50, so care more about 40 and 50’s ratings! I KNOW there are less of us here, but I’d be willing to WAIT for the more appropriate older demographic, rather than get “instant” results with the wrong folks. Can you OK wizards make this happen :-)

  30. Just_this_guy-- says:

    Why can’t I just use the photos I already have uploaded on OKCupid? I don’t want to find them on my computer and upload them again. I should be able to just select them from the list of what I already have loaded.

  31. -Ampersand- says:

    How long is it supposed to take to get back the results? I submitted my pictures seven hours ago and still don’t have a report.

  32. Snacky says:

    Please add some explanatory text about what the numbers and charts mean.

    I look at the results page, and I don’t understand, for example:

    * One picture has “51/100″. What does this number actually signify? This particular picture’s bar chart is overwhelmingly in the negative. Why is this so, if the numeric “51/100″ is apparently ‘average’. ??

    * “Normalized” to what? For example, by age? On a normal distribution of all ratings?

    * The final bar chart doesn’t seem to be useful, since the length of each bar seems to represent number of respondents with that tag. E.g., it’s not possible to compare the darker colored “yes” bars:

    if this final bar chart were enhanced to display the “yes” percentages represented by the inner dark colored bars, then that would be extremely useful: We could see, for all our pictures, with which groups they most resonate. As the chart currently stands, no “information” can be derived from it.

  33. Hmm says:

    Ok, after using it a bit:

    * In general, it is awesome. I am happy this exists. I am clueless about my own attractiveness.

    * The number of votes seems too small to be statistically accurate.

    * Allowing us to write why we liked a picture might be good, but could wound people’s self-esteem when we’re forced to choose between two ugly people.

    * If my normalized high score is 72/100, does that mean that other pictures that are 72/100 are equally attractive? If I look at a chick’s picture and think “wow she’s hot”, and she has the same score as me, does that mean I’m similarly attractive and she’s in my league? Or do the scores not relate between profiles that way?

  34. Snacky says:

    I have a suggestion for making the per-picture bar chart much more useful:

    Problem: It redundantly incorporates the “overall” score of the picture. The presentation makes it hard to compare, from picture to picture, the relative appeal to the different groups/tags.

    Solution: Change the per-picture bar chart in two ways: (1) each chart is normalized so that the y-scale max is the picture’s overall score. (E.g., 76, 52, whatever). (2) the y-axis is changed to start at zero, instead of going into negative.

    I believe that these changes would make the charts *much* more useful: We could compare pictures better against each other, and see, e.g., “Ah hah — with this angle, I pick up vegetarians, but lose Christians.”

    Currently, that use case isn’t supported because the bar chart does double-duty, incorporating the overall (raw) score for the picture, and these cross-picture comparisons are hard to make.

  35. Brian says:

    I agree with Tim, you need to find a way to disqualify girls I think are bitches:) hehe

    While there are some flaws, this is still pretty neat. I think it should be integrated with OKC though so only registered users can view or participate. Also, I would treat it as a feature, not a novelty. It only becomes truly useful if many users participate.

  36. Greg says:

    How exactly are the scores “normalized?” It seems like virtually every picture ends up in the 50-70 range. Are they being normalized to other pictures in the report? Against other users’ average scores? Some more information about how these graphs get generated and what data I’m actually looking at could be really useful.

  37. danny says:

    so I ran this twice and guess what my number 1 image from last time came in last. Can there be an explanation for this maybe, I noticed that “my best face” from last time scored very highly amongst younger introverts. this time the crowd who was rating seemed to be a tad older and more extroverted. Just a theory at this point. Overall I suppose there is some wiggle room based on who’s rating. I’d like to see further analysis on why the same picture would rank differently.

    Nonetheless a great program

  38. Sean says:

    What about location based reports? US v UK etc.

  39. cbs says:

    The “vs” part of the last page isn’t working very well. The pictures don’t line up at all. Also, I can’t get the facebook connection to work, the small thumbnails will load up, but I can’t actually choose from them.

    Otherwise, this is super fun. I’m also really clueless about which profile pics look good and which don’t.

  40. Roman says:

    Can’t wait to see the data mining on all data generated from MyBestFace!
    I bet, there will be some unobvious findings.

    Would be nice to know how preferences correlate among different groups. Are people able to choose other people from similar self-identified groups by just having a look at the picture?

    Also it would be nice to correlate the numbers you get on photo selection with messaging on real profiles of the same people (anonymize the data, and give us general idea). That would tell how much exactly does the picture contribute to your attractiveness comparing to everything else (text in the profile, messaging skills, etc).

    Another question is – why aren’t there perfect 100s? 82 is the highest I’ve seen for women, and 78 for men. Does that mean, there are no ideal good looks, and there are differential looks preferences?

  41. Bryan (2) says:

    Pretty cool, but one small design flaw: pushing the down key seems to place a vote. I have cast quite a few accidental votes while trying to scroll down.

    Also, don’t you think this is going to push people who are heavy, old or otherwise generally unattractive to pick pictures where they are barely visible? I know when presented with two pictures one of which is someone I find unattractive, and the other is arty or blurry or from a distance, I’ve picked the blurry/arty/from-far-away picture.

  42. Jeff says:

    My report is taking a reeeeally long time. Submitted at 9:30 this morning, I still have nothing. Is this a normal wait time?

  43. goocy says:

    Seconds on the question “are we in the same league?”.

  44. Jim says:

    “Why can’t I just use the photos I already have uploaded on OKCupid?”

    You can, if you select “On the web”, then right-click the images and “Copy image location” and paste it into the field. But yeah, it should be easier.

  45. Ugh says:

    Also, there are SO MANY pictures with multiple chicks in them. I “report” them, but why are people uploading them in the first place? Can you put a big notice “The picture should only contain you” or “It should only contain you and other males”, so we know who we’re voting on.

  46. Jim says:

    Once your report is finished, is that it? or do people continue to cast votes on our photos to continually update your report?

  47. Jay says:

    This is great. Cant wait to see my results.

  48. jay says:

    As far as I can tell the ‘normalising’ is just a linear rescaling of the percentage vote so that the minimum is ~35 and the maximum ~90. A 50% vote works out as about 62.

    I guess it’s just a buffer for our collective egos – no one wants to see a 0 rating. Overall pretty cool though.

  49. Eukaryote says:

    To be quite honest I don’t think this feature is all it is cracked up to be. I don’t think you should be rated/be rating people outside of your age category, as they will no doubt choose someone closer to their age over you.

    On the other hand, the results are repeatable. I got the same score on the same two photos three times.

  50. Scott Valdez says:

    Amazing tool, guys. A business contact and I actually created a business plan around doing something similar about 9 months ago but I have to admit… it would have been inferior. This is brilliant… a great tool for serious online daters!