The first order of business is getting the traffic that is focused on the niche. There are myriad ways to bring in traffic, some more stable than others - such a valuable content and community. Content, unforunately, has a way of migrating off campus without permission, and you can spend your days tracking it down. Community tends to be a bit more stable, until the "big issue" starts to pull it apart. Directories tend to be subject to ripping, just like content and ranking for any directory subject will forever be a tricky matter. (Which is why there's some small advantage to owning the generic name of the subject matter of the directory. :)
The rule so 1999 still pertains: It's about eyeballs and their demographics. Once you have the eyeballs then monetarizing is about matching up the eyeballs and advertisers. Contextual advertising was a missing piece in 1999 and was behavioral targeting.
Interesting times. In my analysis, I find the more competitors the more opportunity arises.
It's a big world. It may not support a lot of the old behemouths, but the big world may well support a number of hunter gatherers. :)