Libraries like mechanize and hpricot are shrinking the curve for scraping tasks. That's not to say it is easy, but it should not take a bunch of people working on it. One good developer with proper experience would be ample in my opinion.