FORTUNE — Sep 14 — eHarmony's constant tweaking of its matchmaking system appears to be working. The 7-year-old online dating site is "very, very profitable" with annual revenues of almost $200 million. Paid subscribers, revenues, and profits are all up roughly 30% year-over-year. By year-end, the company will have $100 million in cash and is eyeing an IPO in the not-distant future. With an expansion into Canada, and the UK and Chinabeing next, their next challenge is to incorporate local behaviors into their matching software.
The full article was originally published at CNN Money, but is no longer available.
Mark Brooks: I'm surprised eHarmony hasn't had more competition spring up. PerfectMatch and Chemistry are the only true competitors and eHarmony has three time the traffic of the two of them combined.
“By year-end, the company will have $100 million in cash and is eyeing an IPO in the not-distant future.”
If they plan to reach the IPO, they should substantiate their matching algorithm!!!
I suspect most Online Dating Sites offering compatibility_matching_methods like
eHarmony
True
PerfectMatch
Chemistry
YahooPersonals (WeAttract’s tests)
CyberSuitors
Ulteem
Parship
Plentyoffish Compatibility Predictor
MatchWise
and others
are using self_adjustment multiple/simple linear/nonlinear regression_equations to predict compatibility between prospective mates, so the whole precision is less than you could had achieved searching by your own!!!
There are more than 900 “Online Dating & Social Networking Sites” at the United States and Canada, and top 10 Online Dating sites have 80%/85% of actual market.
If you add net paid subscribers of all U.S. dating sites, perhaps the total is less than 5 million persons!!!
What dating sites are doing / will do to court the other U.S. 87 million singles not seriously dating online? They need to offer GREAT INNOVATIONS, but …. they will definitively come from new discoveries on Theories of Romantic Relationships Development with:
* 16PF5 NORMATIVE test as main personality test (available in different languages)
* powerful math equations (not more regression equations!!!!)
specifying:
– Ensemble: the whole set of different valid possibilities.
– Precision: like 0.00000001%
– Main Matching Equation: like an adapted quantum mechanic equation.
– How the compatibility predictor is expresed: like Client#X to Client#Y == 92.55033557%
– Average number of “compatible real persons” for one person over the entire database: like the 3 most compatible in 100,000 persons database or the 12 most compatible in 1,000,000.
Regards,
Fernando Ardenghi.
Buenos Aires.
Argentina.
ardenghifer@gmail.com
“By year-end, the company will have $100 million in cash and is eyeing an IPO in the not-distant future.”
If they plan to reach the IPO, they should substantiate their matching algorithm!!!
I suspect most Online Dating Sites offering compatibility_matching_methods like
eHarmony
True
PerfectMatch
Chemistry
YahooPersonals (WeAttract’s tests)
CyberSuitors
Ulteem
Parship
Plentyoffish Compatibility Predictor
MatchWise
and others
are using self_adjustment multiple/simple linear/nonlinear regression_equations to predict compatibility between prospective mates, so the whole precision is less than you could had achieved searching by your own!!!
There are more than 900 “Online Dating & Social Networking Sites” at the United States and Canada, and top 10 Online Dating sites have 80%/85% of actual market.
If you add net paid subscribers of all U.S. dating sites, perhaps the total is less than 5 million persons!!!
What dating sites are doing / will do to court the other U.S. 87 million singles not seriously dating online? They need to offer GREAT INNOVATIONS, but …. they will definitively come from new discoveries on Theories of Romantic Relationships Development with:
* 16PF5 NORMATIVE test as main personality test (available in different languages)
* powerful math equations (not more regression equations!!!!)
specifying:
– Ensemble: the whole set of different valid possibilities.
– Precision: like 0.00000001%
– Main Matching Equation: like an adapted quantum mechanic equation.
– How the compatibility predictor is expresed: like Client#X to Client#Y == 92.55033557%
– Average number of “compatible real persons” for one person over the entire database: like the 3 most compatible in 100,000 persons database or the 12 most compatible in 1,000,000.
Regards,
Fernando Ardenghi.
Buenos Aires.
Argentina.
ardenghifer@gmail.com
Kirkpatrick says they’re doing $200 Million in revenue…not $100 million
Kirkpatrick says they’re doing $200 Million in revenue…not $100 million
It’s their ad spend, not the constant tweaking of the matchmaking system.
It’s their ad spend, not the constant tweaking of the matchmaking system.
I agree with Dave Evans here — the validity of their “matching system” has been seriously criticized (arguably fatally so) by several academics. Their financial success is due to the sales/marketing illusion of overhwhelming success in creating marriages/relationships. I sincerely doubt they have the research expertise to “tweak” their matchmaking system — no one on their team is a recognized tests and measurements expert.
On a related note, not all compatibility tests or websites use regression modeling for creating matches between people. The issue has never really been the mathematical “accuracy” or precision in matching algorithms as Fernando suggested. The primary issue is what relationship variables those algorithms are using in the first place — and most models focus too heavily on similarity in “personality,” which is an approach that experts know is misguided.
Thanks,
James Houran, Ph.D.
Online Dating Magazine
I agree with Dave Evans here — the validity of their “matching system” has been seriously criticized (arguably fatally so) by several academics. Their financial success is due to the sales/marketing illusion of overhwhelming success in creating marriages/relationships. I sincerely doubt they have the research expertise to “tweak” their matchmaking system — no one on their team is a recognized tests and measurements expert.
On a related note, not all compatibility tests or websites use regression modeling for creating matches between people. The issue has never really been the mathematical “accuracy” or precision in matching algorithms as Fernando suggested. The primary issue is what relationship variables those algorithms are using in the first place — and most models focus too heavily on similarity in “personality,” which is an approach that experts know is misguided.
Thanks,
James Houran, Ph.D.
Online Dating Magazine
“most models focus too heavily on similarity in “personality,” which is an approach that experts know is misguided.”
Hi Dr. Houran!!!
Very pleased to contact you once again!
When you said “an approach that experts know is misguided”
Who are those experts?
and
How they could know the similarity in personality approach is misguided?
misguided == mistaken
Had they used the 16PF normative test in different languages with and advanced quantitative method to evaluate similarity between prospective mates?
How did they manage to calculate/evaluate similarity?
Regards,
Fernando Ardenghi.
Buenos Aires.
Argentina.
ardenghifer@gmail.com
“most models focus too heavily on similarity in “personality,” which is an approach that experts know is misguided.”
Hi Dr. Houran!!!
Very pleased to contact you once again!
When you said “an approach that experts know is misguided”
Who are those experts?
and
How they could know the similarity in personality approach is misguided?
misguided == mistaken
Had they used the 16PF normative test in different languages with and advanced quantitative method to evaluate similarity between prospective mates?
How did they manage to calculate/evaluate similarity?
Regards,
Fernando Ardenghi.
Buenos Aires.
Argentina.
ardenghifer@gmail.com