How A Computer Nerd Revolutionized Online Dating

VANCOUVER MAGAZINE - Nov 20 - Markus Frind, the founder, CEO, and sole owner of Plentyoffish.com, works as little as he desires. He’s been quoted as saying he works about 10 hours a week. Plentyoffish.com now has 16M users, earns tens of millions of dollars a year, and is growing 80% a year. Alexa ranks Plentyoffish.com the 35th most trafficked site in Canada and the 100th in the US. It serves up 2.4 billion page views/month.

Plentyoffish.com spends far less money than other sites do. Match.com and eHarmony.com each spend $100M a year on advertising. They run thousands of costly servers even though their traffic is smaller than that of Plentyoffish.com, which has only 11 machines.

The next level Frind aims to conquer is the world of “scientific” dating research. Companies like Chemistry.com, eHarmony.com and Genepartner.com hire psychologists and scientists to design questionnaires and even DNA tests to help people find mates. Frind did hire a relationship psychologist to create “compatibility matching tools” for Plentyoffish.com. Now he thinks he can do better with his own algorithms. “The science of dating,” Frind scoffs, “is all BS.” Those much-heralded matching tools are only 15 or 20% better than chance. And does Frind think he can optimize and tweak his way to beat that score? “I could double or triple it.”

See all posts on PlentyofFish

Comments

[Fernando Ardenghi]

If Mr. Frind is so rich, he should begin acquiring the license of the 16PF5 normative personality test, available in different languages to assess personality of daters, or hire a Psychologist to develop a propietary test with exactly the same traits of the 16PF5, to construct a copycat version of the 16PF5 (one per language and norm), but improving it using Item Response Theory / Rasch Scaling (the Gold Standard for Modern Test Theory) with different questionnaires for men and women and all the questions rearranged in a Rasch hierarchy.
The daters, men and women, should take the test at least 3 times, check if they are under hormonal treatment (women taking contraceptive pills to synchronize 7 days before, in their periods and 7 days after) and also synchronize with biorhythms (for future Research purposes).

The ensemble of the 16PF5 is: 10E16, big number as All World Population is nearly 6.7 * 10E9
(WorldWide, there are over 5,000 -five thousand- online dating sites, but no one is using the 16PF5)

Then he will need to develop a high precision quantized pattern comparison method (part of pattern recognition by cross-correlation) to calculate similarity between prospective mates.

For a 1,000,000 persons Database.
Number of Comparisons == (1,000,000)raisedto2 / 2 == 500,000,000,000 each one to predict relationship satisfaction between persons.
If the computer (or high_speed server arrangement) calculates the value of relationship_satisfaction_between_prospective_mates as fast as 1,000,000 per second, it will require 500,000 seconds, nearly 139 hours to sweep the entire database.

Imagine the power calculation for 10,000,000 persons Database!
If Mr. Frind is so rich, perhaps he can buy a USD10millions CRAY supercomputer.

-----------------------------------------

The entire Online Dating Industry for serious daters in 1st World Countries, is a big hoax.

Since the beggining of 2003, I had been testing online dating sites who offer a compatibility matching method to their daters by creating dummy Male/Female profiles and using them as test points for reverse engineering purposes.

All the algorithms used by eHarmony, True, Chemistry, PerfectMatch, Be2, Meetic, PlentyOfFish Chemistry Predictor, Parship, RewardingLove, MyType, etc. are like placebo, because they will show, to any member, 3 or 4 persons as highly compatible per 1,000 persons screened, so in a 10,000,000 persons database, any member will see 30,000 to 40,000 members as highly compatible; 30,000 persons is the population of an average small city. Any person can achieve 3 or 4 persons as highly compatible per 1,000 persons screened, searching by his/her own or by mutual filtering methods.

Success Rates of those sites are less than 10%. The majority of their members are not going to achieve a long term relationship with commitment (or marriage).

I had discovered "the online dating sound barrier". Breaking "the online dating sound barrier" is to achieve far better precision than searching on one's own or mutual filtering.

Breaking "the online dating sound barrier" is to achieve at least:
3 most compatible persons in a 100,000 persons database.
12 most compatible persons in a 1,000,000 persons database.
48 most compatible persons in a 10,000,000 persons database.

The only way to achieve that is:
- using the 16PF5 normative personality test or a propietary test with exactly the same traits of the 16PF5.

- expressing compatibility with eight decimals, like The pattern 6.7.6.8.9.6.7.7.8.7.2.5.8.7.3.4 is 92.55033557% +/- 0.00000001% similar to the pattern 7.7.6.8.8.7.6.5.8.7.4.5.7.7.3.4

That is what is going to revolutionize the Online Dating Industry.

Regards,

Fernando Ardenghi.
Buenos Aires.
Argentina.
[email protected]

[Steve Odom]

Marcus seems to have a good nose for opportunity. I wouldn't bet against him. I agree with him that the existing compatibility tools are only marginally better, if at all. I'm interested to see his approach.

We are working on our own compatibility/recomendation engine at Gelato. The thing that I struggle with is what am I optimizing for? Is it to have two people contact each other? I can measure that but I don't know whether the date is successful. Is it marriage? That's so long term that I can't get any meaningful data. Is it showing a person a profile and asking them whether to rank whether they would be compatible with that person? That seems pretty fragile. Does anyone have any thoughts on this or point me to good research?

Steve
Founder: Gelato Dating
http://ge.la.to

[David Evans]

There is matching based on psychology and the like, and then there is putting people in front of members based on who people have responded positively to, or some other metric based on past actions. The two are completely different. Markus seems to be mining how people use the site and making predictions based on that. Nothing to do with actual matching.

[Fernando Ardenghi]

"I wouldn't bet against him"

Mr. Odom, please read what Mr. Frind had said:

"Frind did hire a relationship psychologist to create 'compatibility matching tools' for Plentyoffish. Now he thinks he can do better with his own algorithms. 'The science of dating', Frind scoffs, 'is all BS.' Those much-heralded matching tools are only 15 or 20% better than chance. And does Frind think he can optimize and tweak his way to beat that score? 'I could double or triple it.'

double or triple?
He is completely wrong!!!

It is not to double or triple better than chance, it is to achieve 100 times better.

Matching based on Self-Reported Data / Bidirectional Recommendation Engines (Collaborative Filtering) will always be in the range of 3 or 4 persons "recommended" per 1,000 persons screened, in exactly the same range of searching on one's own.
and
Compatibility Matching Algorithms used by actual online dating sites (with Big-5 Model or a proprietary model like the ones used by Perfectmatch or Chemistry), are in the range of 3 or 4 persons "highly compatible" per 1,000 persons screened, in exactly the same range of searching on one's own. They are using simplified versions of personality traits, instead of the 16PF5 or similar with the complete inventory (16 variables) and inadequate quantitative methods to calculate compatibility between prospective mates, like eHarmony which uses Dyadic Adjustment Scale or other sites which use multivariate linear / logistic regression equations o other powerless equations.

Breaking "the online dating sound barrier" is to achieve at least:
3 most compatible persons in a 100,000 persons database.
12 most compatible persons in a 1,000,000 persons database.
48 most compatible persons in a 10,000,000 persons database.

100 times better than Compatibility Matching Algorithms used by actual online dating sites!

The only way to achieve that is:
- using the 16PF5 normative personality test.
- expressing compatibility with eight decimals.

Regards,

Fernando Ardenghi.
Buenos Aires.
Argentina.
[email protected]

[Pemo]

Thanks for that clarification Dave, there is a difference between using PlentyofFish data algorithm predicting the odds of people connecting. Matchmaking is another area of expertise altogether. I respect Markus but I don't agree that matching criteria on all sites is 'BS'. Being human none of us have the complete picture but maybe we all bring something to the table to explain the mystery of 'the chemistry of attraction'. It would be great if Markus can improve the odds by 15-20% and I guess because he has used a psychological tool previously he should then be able to measure if he reaches this goal? I look forward to hearing his results. I have reviewed a number of dating sites matching assessments on my blog, check out the archives @ http://www.astramatch.com/blog

[Mark Brooks]

Dating sites should at least try to guide their users away from people who are clearly INCOMPATIBLE for them. The technology and science of compatibility and personality profiling has come at least as far as being able to provide this service with some degree of confidence.

[Mark Brooks]

Ultimately, people don't know what they want. Actual behavior is a far better predictor of future behavior.

[David Evans]

Everyone knows what they want, what we lack is an appropriate way to explain what we want. Give people the tools to express themselves and who they seek, and they have a better chance of finding that person. This is why online dating is so difficult.

@Pemo, you social media maven, love your blog. You need to keep telling people about astro-matching, it's fascinating and just as relevant as the current scientific matching going on. Technology/Internet people don't appreciate the stars, it's their loss.

[Mark Brooks]

The science is still in its infancy, but holds great promise.

In the beginning, circa 1994 (i.e. webpersonals.com), dating sites allowed users to search and more quickly find matches. No more grooming through pages of personals in newspapers. Users could search and find exactly what (they thought) they needed in a match.

The next great advance that dating sites offered over plain old newspaper personals was improved communications. No more mailing in letters to the newspaper. Dating sites offered email, instant messaging, winking, webcam chat, chat rooms, etc.

The 3rd, and by far the biggest area that dating sites will improve over plain old newspaper personals is by helping people find better matches. That's the realm of personality profiling, compatibility matching, and behavioral analysis.

Its nice that users can search quickly, and chat online before they go on an actual date, but wouldn't it be nice if they could avoid getting matched with someone they are completely incompatible with. There's huge value in guiding users away from people that they would be most likely to end up being miserable with.

But who's to say people will be happy, even when they are matched with people who they are highly compatible with, and with whom they have great physical chemistry. Still, most people lack the relationship skills to keep it together for the long haul. And even if they do manage to keep it together long enough to raise a family (if that's their inclination), who's to say they will make good relationship role models for their children. My point is, even when 'good' matches are made, people still screw it up because they don't know how to keep it together.

All in all, there's lots of room for the dating industry to grow and add services, and add value.

[Sam Moorcroft, ChristianCafe.com]

One positive coming out of Canada concerning divorce rates is that they are not as high as commonly thought. The National Post headlined this on its Nov 20, 2009 edition entitled, "Stick with it": http://www.nationalpost.com/news/story.html?id=2245026

"Divorce rates are not as high as we thought. Divorce rates have been coming down since the 1990s and since 1997 have plateaued. In fact, first marriages have a 67% chance of lasting a lifetime."
...
"...there are two kinds of divorce: those resulting from an unhappy marriage, and those resulting from 'a weak commitment to marriage.'..."some divorces are avoidable and unnecessary" and ...'a sizable proportion of marriages that end in divorce were actually quite 'salvageable,' even happy, and that many of these ex-spouses are no better off after."

It will be interesting to what role dating sites play in all of this.

[Fernando Ardenghi]

Hello Mr. Brooks!

What you had written "Ultimately, people don't know what they want. Actual behavior is a far better predictor of future behavior."
and
what Mr. Frind had said to Vancouver Magazine "'Everyone is a liar,' he answers immediately, when asked what his site has taught him about dating"

is only valid for "the short-term mating strategy psychology"

Psychologists and other Academics had discovered that men and women have 2 "psychologies":
"the short-term mating strategy psychology"
and
"the long-term mating strategy psychology",
which is part of the new discovery uncovered by Eastwick and Finkel 2008; also Kurzban and Weeden, 2007; Todd, Penke, Fasolo, and Lenton, 2007 who found that people often report partner preferences that are not compatible with their choices in real life. Also Gangestad & Thornhill 2008 says that [for women] short-term but not long-term partner preferences tend to vary with the menstrual cycle.

Daters approach online dating sites with their "long-term mating strategy psychology", but if they detect they are going to fail in finding someone compatible for long term mating with commitment, they change to "the short-term mating strategy psychology". That is what is happening at PlentyOfFish!!!

PlentyOfFish, a "free" online dating site, where low quality daters "pay" with precious time, instead of paying with money.

--------------------------------------------

"Actual behavior is a far better predictor of future behavior."
Personality analysis with normative tests is the BEST predictor of future behavior.

Regards,

Fernando Ardenghi.
Buenos Aires.
Argentina.
[email protected]

[Christian Dating Australia]

Fernando, I think you are pefectly correct that there should be some form of more organised system, and in fact it sounds like the CEO mentioned doesn't really actually care how well people are matched. That said however many of these types of sites try to match people reasonably well and let people go from there, rather than working to create genuine well suited relationships. As long as there are enough signups, and people find other people roughly the same (enough not to doubt the system), thats all they care about when it's a money making exercise. I suspect many sites are like this when claiming to match, but the others don't admit it like he does!