How about this: instead of showing you match %'s, they show you profiles and you rate them. Then use recommender systems [1,2] just like netflix. There are two immediate problems with this: you have the cold start problem [3] and also if you don't show random profiles, then you introduce bias into their rankings. The first problem could be solved with side information (you still answer questions) and the second with a nicely chosen exploration/exploitation trade off. PMF[4] for the win!
OKC does have this functionality, it's called quickmatch. I don't know how they use the ratings you give there, but I doubt they just let it go to waste.
* If you're going to do it properly and give them meaningful data, a 'quick' skim is the exact opposite of what you want to do. Yet it's what the function is set up for; you can't even see a person's questions and answers, which have frequently (to me) proved far more revealing than the profile.
* Again, if you're doing it properly, rating someone highly sends them a mail saying 'someone's interested in you'. Which always looks to me like saying 'Someone's interested in you but doesn't have the guts / energy / enthusiasm to write you a proper message so has gone for the easy way'. Not a great intro.
With a bit of tweaking I agree it could give some useful seed data, but not much more; fundamentally I think the question-based approach is very good, but I think their scoring algorithm could benefit from refinement.