This is great - well done. Some feedback:
* (only one point on the design, since design feedback is so subjective - you have lots of a funny red color) - The screenshot of your site included in the site as part of the splashpage clutters the page. My eyes darted around when I first landed on the page trying to work out what was going on. Try cropping them so that it doesn't include the menus in the screenshots or better yet but some movie covers there with your 'action icons' overlayed.
* Your action icons are not consistant. Actions are 'have seen', 'want to see' and 'dont want to see' - and they are spread out. I think a more natural place for them is at the bottom, like pandora does. A lot of people will also not assume that just clickin on the cover takes you to that films record - perhaps another button 'info'
* You also have favorit - which means you have to see it first then fav it. I would just have 'favorit' as an option icon since it implies you have seen it (plus to favor a movie atm is a 3-step process).
* Take me back to the page I was on after signin (pehaps with that help as a lightbox).
* The site doesn't help you watch movies (ie. you cant view them online), which was my first impression - so the tagline is 'helps you find films to watch'.
* You have a revenue model in referral fees to amazon, netflix and the apple store.
* I would drop email (esp the gmail logo :)) and have it tweet out your 'seen's and 'likes'. Also drop email on signup (or that entire signup step) since you don't really need it and sharing over fb/tw is much better than email anyway.
* You need it inside facebook as an app
* An early site that I used to use often was http://www.whattorent.com/. They make recommendations based on the type of movies you like and what mood you are in. I think you should extend this to recommendations like you said, but because the web is so much more social now, and people make decisions based on what their peers/friends say and do, you can replace the 20 questions with common interests and friend recommendations. Instead of getting caught up in a /10 score, rely on ppl either liking or not liking a movie and then recommending based on what friends have liked/not liked. if you keep the app simple (ie. 'steve just watched and loved the godfather [add to mylist]') and not clutter it - this can be very successful/viral and provide good revenue.
* how about movies coming out? bookmark that you want to see a movie when it is released.
* A lot of ppl watch movies based on actors and directors (the studios know this, which is why they are both paid so much and second why so many bad movies are made with good actors). if you allow liking actors or directors, you have an instant recommendation shortcut (ie. new brad pitt movie coming out)
* If you do this right, it could become a very good imdb alternative (you should link to them btw, and possibly boxofficemojo and metacritic) - since imdb are stuck in the dark ages and haven't done a lot with the social web (their forums, where I am/was semi-active, are terrible).
* Flixster was supposed to be imdb 2.0, but I know I gave up on it because on the info end it was trying to be imdb but wasn't, and on the sharing end it was too complicated. I think there is a model where you acknowledge that you can not beat imdb in the info stakes, but you can provide the social layer in recommendations by keeping it simple.
otherwise this is great. I would use this to see what my friends have watched and liked in twitter/facebook and would mark the movies I have seen likes and let my friends know about it as well. a lot of ppl are manually tweeting 'just saw the social network and did/didnt like it' so might as well wrap all that into a nice small mini-app.