A joke an econ professor once told me:
"Look under your foot, do you see a twenty? No? Well, thats because someone already took it."
The point was a generalization of the fact that in a free market, arb opportunities don't exist because market inefficiencies are corrected in real-time.
Thats not always true, but the general point is.
The problem with finding good jobs on Monster, CL, etc. is that, more or less, there aren't any. Good jobs fill fast, most often through backdoor channels - networking, campus recruiting, etc.
The jobs that flow down to sites like Monster and CL are those that smart, skillfull people don't want.
Think about it from your own perspective. If you wanted to hire someone for a job, chances are you would turn to your circle of friends and trusted advisors and ask them if they knew anyone. You branch out from there like a ripple, with each ring being less focused and more general, until you hit mass market sites likes Monster.
I think this is why the majority of jobs on these sites are usually low-paying admin type work or one-off small business contractual work.
LinkedIn attempts to solve for this, but they I don't think they have a direct enough of a mechanism for doing so.
Other sites fail too because most operate under the assumption that they can do better by improving usability and don't really address the underlying problems, so they suffer from the same issues (namely no good work).
The sites that I think do a better job are those that are niche industry focused. Niche sites help job seekers find relevant (to them) jobs and job posters find relevant hires to fill those positions. This helps both parties filter down, which I think is their value-add.
Ultimately whatever improvement that hopes to be successful needs to focus on ways to either change or divert the current way most employers find employees - word of mouth and personal networks.
Perhaps better mining tools for LinkedIN would solve. I think some functionality that lists job seekers by skill (profile, resume) weighted by in-network relationships may help.
Another idea is scaling the head-hunter model for public consumption. Something where as a job poster, I could send a request out to my network on linkedin, and there would be some sort of incentive (money or otherwise) for those in my network to recommend people the know. There is already natural incentive to do so... having a solid network is a good skill to have by itself; knowing good people says good things about the type of person you are.
Just some thoughts. I know relatively little about this space, so my opinions are just based on my own experiences.