These two problems have the same solution: charge men for membership, women join for free. Every college keg party has figured this out, I'm surprised online dating sites haven't.
You might be right, but I 'm not sure which element is more important: sheer numbers or the quality of the people who have joined.
On a site that men pay but women join for free you're going to find a lot of women who expect men to pay for everything for them, and the only men on that site are going to be the kind who expect to get what they pay for.
Now, I'm not saying there's anything wrong with that, everyone's a grownup after all. But that isn't the demographic that match.com et al are after - they measure their success in how many marriages they produce. Marriage-oriented people are also more likely to be at the point in their lives where they have money to spend to, they're a highly desirable market.
The more people on the site who live near to a user, the more likely that user is to find someone compatible with them.
Having said that, "quality" might be important if it's a way of making it likely that users are more compatible with other users, e.g. religious-minded people who want to marry within their religion.
It's not that great for women, either, since they're overwhelmed with contact by large numbers of men.
This hasn't really stopped men from trying, though. The gender ratio doesn't have to be perfectly balanced, all is has to be is not ludicrously out of whack.
I like your thoughts on targeting young professionals, and basing it on an existing network such as LinkedIn also sounds reasonable. Your focus on keeping personal information out of the conversation is not viable - remember, it's not just e-mail and phone numbers, it's also suggestions to meet. I also think that the fee to be hooked up with a second person is going to be a major deterrent.
A little experience with queuing models and Petri nets also helps.
As for the queuing models & Petri nets, assuming everyone understood I was a more than a little tongue-in-cheek:
- Setup your net as a cycle between weekdays and weekends. I only use 1 week of forward planning, so my base net is 3 weekdays (I save a few weekdays for non-dating) and 2 weekend days. 5 vertices.
- Duplicate the net for different stages, roughly equating one stage per date sequence number. E.g stage 1 is first date, stage 2 is the second date, etc.
- Each girl is a token.
- Each token in each stage tells you how your day is likely to go, once you equate the stage # to an expected progress level. Especially early on in the dating sequence, dates can be short, so multiple dates can be stacked into a day (e.g. an afternoon 1st date vs a dinner/movie 3rd date that night).
Nothing fancy by any means. Note, again, that this is pretty tongue-in-cheek, but it does help me keep a rational (and semi-quantitative) focus on what is fundamentally a numbers game. I don't lose my cool on dates b/c each one I understand as a single token, instead of freaking out on how this or that girl could be "the one."
Edit: clearly this needs work, but I'll have to look up Petri nets again before tightening this up.