Robert_Charlton - 8:27 pm on Aug 3, 2010 (gmt 0)
My emphasis added...
To test out how "good" these types of links were, I went out and grabbed a handful (10 or less) of these links to a couple of my new domains. Almost immediately, Google removed the domains from the index for the targeted keywords.
I think there's consensus that Google looks at the overall quality profile of a site when assessing a backlink penalty. It's likely that there's a separate score also for backlink quality. On a large established site, with a great many backlinks, a small percentage of questionable links probably wouldn't hurt.
Your "new" domains, though, generally haven't had much of a chance to establish a high quality profile. I'd guess that these low quality backlinks may have constituted a large enough percentage of the backlinks to your new sites that they have significantly altered the profile balance.
Additionally, if you added the backlinks (perhaps with the same anchor text) all at once, this would look like a coordinated action.
If the links came from pages known to sell links, this too would have raised a flag.
These signals combined might have tripped algorithmic filters, or prompted human review. Hard to say how many times you might have crossed the line. Also, it's hard to rule out other factors and perhaps coincidental timing in your first test.
To continue the test (admittedly on a small and slightly simplistic scale), I would try two things...
- to gradually add some good backlinks with varied anchor text to one of the sites to see what happens...
- and to remove the bad backlinks to the other, without adding new links, and also watch what happens.
I wouldn't perform both tests simultaneously, as you want to rule out concurrent algo changes that might be affecting both sites. Eventually... if there's no change, you might also try filing a separate resconsideration request for each of the sites, but only after you've made the changes and several months have passed.