On the plus side, this is the easiest ranking factor to check. When you throw out the inherent subjectivity of a variety of their signals (by how you craft, say, a TITLE tag), and can simply do an Off/On, apples to apples comparison, it's a good thing. A pain for some, yes, but at least it's a potential cheap bump. I have several in a test group I'll be implementing this for next week, so I'll post results when they're done. Right now, the only SSL sites we run are those that handle payment transactions or customer logins. I have a group of standard, no SSL actually required, sites segmented out to test with.