Given that they measure "success" by response rate I'm not sure I put a whole lot of faith in the below quoted analyses. Perhaps the low "success" rate is because people reply w/ an outside service that OkCupid can't track?
"#4 – Don’t try to take it outside.
Obviously, all successful OkCupid relationships outgrow our in-site messaging feature. But an offer to chat or of an email address right off the bat is a sure turn off. One of the things online dating has going for it is its relative anonymity, and if you start chipping away at that too early, you’ll scare the other person off."
Having used the site, no one puts contact information in their profiles. And a request to use the built-in chat system right away is a pretty sure sign that I'm not going to want to talk to someone.
I'd bet that, to a point, there's a correlation between the volume of text sent and reply rate. However, that might just have to do with the fact that a longer message is more likely to reference specific things in the person's profile that they find interesting (for example, I replied to almost anyone who showed interest in the fact that I bake bread and like kites).
given that okcupid doesn't charge for their services, it seems unlikely that a majority users are switching to out-of-site communications right away. There are no real incentives to do so.
This is really only marginally useful information. Interesting, but not as practical as it seems.
The writers' profile does not seem to have been taken into consideration. Rarely does anyone respond without first checking out the profile of the person who just sent the message. Is saying "hi" really a problem, or is it more that people who typically say "hi" also have a boring profile?
Also, getting a single response doesn't mean much. You might say "sorry" and a woman might respond "that's ok" but when you try to continue the conversation it'll go nowhere. Real life dating advice doesn't discourage saying sorry because you'll get ignored, it's discouraged because it makes you boring and lower status. If you can do it without falling into that trap, it works in real life. (Accidentally bump into her shopping cart and say "sorry" and then smile and change the subject)
My okc stats are: 26 attempts to initiate contact, 9 responses. 1 live meeting and 2 open threads(1 fresh, 1 getting stale)). I don't have numbers for people who tried to iniate with me, except that I've met at least 3 others.
So that means while I get a 35% response rate there's only an 11% meet rate after that (though it may be 22% soon). So 4% total. While getting a little feedback feels good and can be useful, a few polite responses is not generally what a user's goal is. How many messages result in meetings, that's what I want to know.
Not just that, but are people who start with "How's it going" generally writing better? There are a number of ways this data can be misleading... many of which fall under "correlation vs causation."
Some girls get a hundred emails a week from guys looking for hookups. How are the numbers affected by those high-volume shameless creeps who spam every girl who looks easy to them? I suspect that physical compliments are a bad idea, but these numbers don't help confirm that. Saying a girl is "pretty" could actually help, yet score low in this analysis because it's a creep marker. There's no attempt at persuasive analysis here, just like in the other recent post that was featured on HN. Or did I miss the part where they explained that they targeted a certain subset of messages for analysis?
oh yeah the girls get spammed like crazy. They get like 500 emails a week. Hell a chick I know didn't even provide a picture and STILL got over 100 emails the first week.
Fascinating data. The reply rate for messages containing "how's it going?" is almost twice that of messages containing "hi". In fact, "how's it going?" is the only string mentioned (apart from "Zeus") that actually pushes the reply rate up above 50%.
I wonder if "god" faired so poorly because it was in lowercase, as it fails the first advice: "Be literate." Christians are careful to use it as a proper noun, God. So if the word christian faired well, and god did not, it seems like it's a turn off because you're probably not really a Christian as opposed to the fact you mentioned God.
I doubt it. I'm willing to bet that they did a case-insensitive search. Not only does it make more sense for making these charts, it's also the default in SQL and in most full-text search tools.
I wonder how well mentioning religion to a person who puts religion on their profile vs mentioning to someone who doesn't compares. They say "Christin" doesn't help as much as atheist, but I would bet that people who put "Christian" in their profile respond more...
"God" is also sprinkled fairly liberally around some common phrases. "Oh my God" comes to mind, shouldn't be hard to think of a dozen others on your own.
I can't stay if these conditions are accounted for, of course.
Compare "Oh, hey, I see you're a Christian too." and "I hope we can grow and enjoy God's love together.". The latter seems to connote much more devotion than the former. I would bet that "God" is more likely to be used in a devoted sense than "Christian". I would further bet that the devoted can find each other at church and/or devoted language from a man to a woman in the first contact doesn't come off well. (Disclaimer: poster is an atheist who was raised Christian but does not intend to bash religion in this particular post.)
This data can be misleading. Girls probably get way more messasges and guys send more messages. The reply rate for messages sent by a man is probably much lower than for messages sent by a woman.
A man in more likely to call a woman "pretty" than the other way around. Thus the reply rate for "You're pretty" is low (30%, the average rate is 32%). But the reply rate for man-to-woman messages is 27%. So saying "you're pretty" probably helps.
I know they took precautions on protecting user privacy, but there is something a little bit unnerving about analyzing personal messages without written consent.
Can someone explain to me why this is still not dead, while the link to 50 Lectures on Brain Anatomy and Development, Neuroscience, and Psychology is dead?
Someone had to say this: how is this (even remotely) HN material? Let me guess: it's about "hacking" dating. Yeah, right. And rape is about hacking sex?
Data analysis and trending is topical to this audience. There are a lot of startups that could use techniques like this to increase return visits--instruct your users how to be more successful.
[+] [-] jadence|16 years ago|reply
"#4 – Don’t try to take it outside. Obviously, all successful OkCupid relationships outgrow our in-site messaging feature. But an offer to chat or of an email address right off the bat is a sure turn off. One of the things online dating has going for it is its relative anonymity, and if you start chipping away at that too early, you’ll scare the other person off."
[+] [-] JeffJenkins|16 years ago|reply
I'd bet that, to a point, there's a correlation between the volume of text sent and reply rate. However, that might just have to do with the fact that a longer message is more likely to reference specific things in the person's profile that they find interesting (for example, I replied to almost anyone who showed interest in the fact that I bake bread and like kites).
[+] [-] dusklight|16 years ago|reply
[+] [-] noodle|16 years ago|reply
[+] [-] rimantas|16 years ago|reply
[+] [-] Goladus|16 years ago|reply
The writers' profile does not seem to have been taken into consideration. Rarely does anyone respond without first checking out the profile of the person who just sent the message. Is saying "hi" really a problem, or is it more that people who typically say "hi" also have a boring profile?
Also, getting a single response doesn't mean much. You might say "sorry" and a woman might respond "that's ok" but when you try to continue the conversation it'll go nowhere. Real life dating advice doesn't discourage saying sorry because you'll get ignored, it's discouraged because it makes you boring and lower status. If you can do it without falling into that trap, it works in real life. (Accidentally bump into her shopping cart and say "sorry" and then smile and change the subject)
My okc stats are: 26 attempts to initiate contact, 9 responses. 1 live meeting and 2 open threads(1 fresh, 1 getting stale)). I don't have numbers for people who tried to iniate with me, except that I've met at least 3 others.
So that means while I get a 35% response rate there's only an 11% meet rate after that (though it may be 22% soon). So 4% total. While getting a little feedback feels good and can be useful, a few polite responses is not generally what a user's goal is. How many messages result in meetings, that's what I want to know.
[+] [-] roundsquare|16 years ago|reply
[+] [-] dkarl|16 years ago|reply
[+] [-] vaksel|16 years ago|reply
[+] [-] greendestiny|16 years ago|reply
[+] [-] BRadmin|16 years ago|reply
http://blog.okcupid.com/index.php/2009/09/03/how-to-get-peop...
[+] [-] hughprime|16 years ago|reply
[+] [-] tlrobinson|16 years ago|reply
[+] [-] ojbyrne|16 years ago|reply
[+] [-] felideon|16 years ago|reply
[+] [-] blasdel|16 years ago|reply
[+] [-] benatkin|16 years ago|reply
[+] [-] roundsquare|16 years ago|reply
[+] [-] Batsu|16 years ago|reply
I can't stay if these conditions are accounted for, of course.
[+] [-] swolchok|16 years ago|reply
[+] [-] chaosmachine|16 years ago|reply
[+] [-] swolchok|16 years ago|reply
[+] [-] jules|16 years ago|reply
A man in more likely to call a woman "pretty" than the other way around. Thus the reply rate for "You're pretty" is low (30%, the average rate is 32%). But the reply rate for man-to-woman messages is 27%. So saying "you're pretty" probably helps.
[+] [-] eru|16 years ago|reply
[+] [-] physcab|16 years ago|reply
[+] [-] Herring|16 years ago|reply
[+] [-] tlrobinson|16 years ago|reply
[+] [-] bbsabelli|16 years ago|reply
I suspect that incorporating the second date conversion rates would skew this data significantly since vegetarian atheists are rarely alike.
[+] [-] biohacker42|16 years ago|reply
[+] [-] silverlake|16 years ago|reply
[+] [-] dgreensp|16 years ago|reply
[+] [-] flashingpumpkin|16 years ago|reply
[+] [-] xiaoma|16 years ago|reply
[+] [-] jrockway|16 years ago|reply
[+] [-] clistctrl|16 years ago|reply
[+] [-] KonaB|16 years ago|reply
[+] [-] jonknee|16 years ago|reply
[+] [-] olliesaunders|16 years ago|reply