PCWorld, of all publications, does a great job Analyzing the Algorithms of Attraction, asking the question, Just how do eHarmony, Plenty of Fish, True.com and PerfectMatch.com work under the hood?
I wish reporters would talk to different dating sites. Getting tired of hearing from the same few companies. There are many other interesting stories out there waiting to be told. That said, this is a *fantastic* article chuck-full of interesting factoids, a must-read if you haven’t already. Article has links to many other related topics.
Point of Interest
Eharmony stores 4 terabytes of data on some 20 million registered users on Oracle10G databases. One billion calculations every day. Based on Google’s MapReduce and Hadoop, both open source, nice. They use Netezza to do a lot of offline calculations to try to understand patterns and business intelligence about user behavior.
As an armchair technologist, this sort of talk gets me all hot and bothered. Dating startups, this is part of the reason why eHarmony succeeded. They live and breath on the data they collect on users each and every day. Set it and forget it dating sites are fine for niche sites, but big iron and computational gymnastics are what make the big sites successful.
Eharmony is using iovation to combat scammers.
True.com runs on a 64-bit… who cares. What *is* interesting is that True has staffers who constantly watch banks of security monitors that alternate between the 300 to 700 video chat sessions occurring at any one time. As if they are going to kick people off for more than an hour. Pretend you’re doing a service, but don’t really kick people off. If I show up on camera naked holding a shotgun, I’m back on in 60 minutes.
PlentyofFish operates on just three Web servers, five messaging servers and five database servers (the entire database is just 200GB in size), yet it serves up 200 billion pages a month to some 12 million users. “My entire cost is only a few hundred thousand dollars a year,” says Frind. The biggest piece isn’t the technology, he says, but the bandwidth required to keep traffic to the site flowing smoothly. Nice problem to have.
PerfectMatch.com risks becoming a hacker target, saying that “These are not the sharpest guys out there. They use the same techniques over and over.”
On the lack of traction of background checks in the online dating industry:
Other sites have been hesitant to embrace background checks. “Scammers use stolen credit cards all the time, so what good is a background check [on a stolen identity]? It’s more of a [marketing] gimmick than anything,” says Plenty of Fish’s Frind.
Dahl doesn’t think background checks are reliable. “There are hundreds of law enforcement databases that aren’t communicating with each other,” he says, adding that PerfectMatch does offer its users the option to buy background checks using a third-party service.
There is a big difference between background checks and identity verification, a common misconception reporters often fail to discuss. Identity verification, trust and reputation services are where online dating and social networking are going, give it a few years.
Dan Ariely, who researches ways to improve online dating, says matching algorithms are a placebo. He starts off saying that profile questions should ask about what people normally talk about when they first meet. Then he says people should be playing virtual games. I used to have high hopes for the role of casual gaming in online dating, now, not so much.
Online dating site visitors Snapshot: November 2008
* Total number of visitors to online dating sites: 22,274,000
* Male users: 52.4%
* Female users: 47.6%
Source: comScore Media Metrix
The top 5 types of abuse on online dating sites
1. Identity mining/phishing and/or 1-1 credit card fraud – 61%
2. Spam – 14%
3. Profile misrepresentation – 7.6%
4. General misconduct – 5.9%
5. Solicitation – 2.9%
Source: Iovation compilation of incidents from online dating sites using its security services
My hat goes off to Robert Mitchell. After a month of talking about the effects of the recession on online dating (zzzzz) it’s refreshing to see someone take the time to do a deep dive into what makes online dating tick.
No related posts.
{ 7 comments… read them below or add one }
Robert L. Mitchell from computerworld wrote the article.
“Dan Ariely, who researches ways to improve online dating, says matching algorithms are a placebo”
He is right because many actual online dating sites offering compatibility matching methods are only fueled by big marketing budgets and not by serious scientific evidence. No one had proved its method matches persons who will have more stable and satisfying relationships than couples matched by chance, astrological destiny, personal preferences, searching on one’s own, or other technique as the control group.
Dan Ariely is wrong when he suggests virtual dating to improve online dating sites. Virtual dating is terrible slow and boring, looks like *SpeedDating with Joe 90 puppets*.
“Plentyoffish.com …. it does offer some matching capabilities if users want them, CEO Markus Frind says he doesn’t promote them — and he is disdainful of the complex matching algorithms offered by some competitors.”
Also its owner said “.. On a free site users expect unlimited choice. This is why things like personality testing work so well on paid sites and on free sites are considered more of a joke”
I think
PlentyOfFish tried to be the next ‘free eHarmony’ but made 3 mistakes.
1) PlentyofFish Relationship Chemistry Predictor based on a “46-item relationship test that assesses individuals’ standing on five broad dimensions of personality.”
2) failed on how to validate the test. “Validation of the POFCP was done using a national sample of individuals with the same demographic characteristics as members of the PlentyofFish community” /// “the only tool of its kind to match you to people worldwide”
3) failed on how to calculate “similarity” between prospectives mates.
Regards,
Fernando Ardenghi.
Buenos Aires.
Argentina.
ardenghifer@gmail.com
I’m relieved to see that Chemistry.com — perhaps the most poorly designed of the dozen major services I’ve surveyed — is rarely mentioned in discussion of serious dating sites. I captured the frustrating experience of using Chemistry.com, hoping they’ll be motivated to correct some of the absurd problems… Enjoy!
Hi David,
That in itself has so much of data that I hardly have to go anywhere else to check the authenticity of online dating, however, that does not restrain people like us to dat online. Well all good & evil together make this world existent & I choose to be good & daring.
Some real good elaborate information about some sites & this is another one to be amongst the great one’s.
Thanks for the info & have a nice day.
How to Get a Date – Jack Reed
Last month my colleagues and I organized a Focus Group in San Francisco to get candid comments and insights on impressions and opinions on 3 On Line Dating Sites ( Match.com, Yahoo Personals and E-Harmony).
While we failed to analyze algorithms – we did have an animated discussion regarding
results, cost effectiveness, customer service.
Unanimous – Eharmony was deemed to have zero customer service and great advertising and was the least desirable site.
Match – the monster- .com was rated highest on ease of use, cost, size of membership and Yahoo Personals was rated by several people as a good beginner site. JDate came up in frequent conversation as the most effective and aggressive site in attempts to link singles.
1. On a free site you expect massive amounts of users and options. Giving users 10 matches a month would net you very little revenue/pageviews consumers wouldn’t take it seriously.
2. When users aren’t looking for marriage they purposely pick people they are not compatible with and all those compatibility tests end up being useless. Personality tests allow you to predict who will get together about 10% better than chance for the top 10% of matches… Just filtering out smokers from not seeing non smokers gives you far greater return.
Why do you think people not looking for marriage are looking for incompatible people? This implies you know what makes people compatible and if personality tests are making this prediction better by 10%, then that seems to imply you are based this point on some other criteria. What is that, exactly?
Why did POF drop Myers-Briggs over what it has now?
A problem that makes filtering smokers from non-smokers and the like difficult is that some sites including POF allow indeterminate responses such “Prefer Not to Say” or even no answer at all.
Kerry, can you share additional insights gained from your Focus Group, I’d like to hear more.
Markus, good points. I find the discord between what psychologists say and what actually occurs on dating sites fascinating.
{ 2 trackbacks }