My main problems with this methodology are that:
a) By making exact mirrors of the domains and putting them on the same IP, you're creating a situation where G will pick what it deems to be the "best" site, which will either be the one it discovers first or the one with the most / best backlinks, and bury the rest. By not controlling a significant proportion of the backlinks you're leaving it up to the randomness of the internet to see which site gets the most links (eg. from stats sites, Alexa, random scrapers etc). As we all know, backlinks trump on-site SEO all day.
b) With only one trial and so many confounding factors, it's an anecdote at best. There are way too many factors to conclude that any difference in performance is a result of the tld. But good on you for trying to do something at least semi-scientific.
I agree with this completely, this is why I stated it would be a multi stage process. I am going to have to disagree about backlinks trumping on-site SEO all day. Lately I seen multiple times I push the hell out of a site called say
www.brandmodel.com and it is pushing slowly. On this site I have an article called brandmodel so it would be something like
www.brandmodel.com/brandmodel and I will be pushing links to the main url only to have that inner page pass it without being touched. The on-site of that page for that keyword is always better and when I start pushing it instead it pushes much easier than the other page was. I am also starting to see sites/pages of mine rank well in serps for keywords I never pushed that have good competition, when I pull the pages and review the on-site SEO is always done well and targets the word it is ranking for. It is not hard to find keywords that rank in the top 20 with nothing or virtually nothing done to them.
As far as this particular site went, first round didn't do much and I will be starting the next round soon. The .us was the only site to fully index and the .com only indexed 1 less page, .net and org indexed 1 less than that, .mobi and info only index the about page. The .us is the only one to be seen in serps. However, I still wouldn't put much weight on these results. Reason being is I set the sites up and failed to change
www.site.com and had left it at site.com and that indexed before the site had been setup. I configured it and then changed it to all point to
www.site.com.. When doing this most of the root domains did not index. I went back and look at some other sites of mine along with doing this on a few new ones. When I let it index as site.com and then change it to
www.site.com it is often taking months to index the
http://www.blackhatworld.com/blackhat-seo/white-hat-seo/www.. This was consistent on more than a dozen domains.
Here is my first thought on next rounds to test but it could change depending on what results I see.
1) the above, already done but will continue to watch the sites.. this one was the least likely to show anything to begin with but I will still monitor and share the results
2) sites with exact same articles as each other like last time but sites will be on different ip's this time
3) sites with unique articles, similar word count and keyword density, all sites on same ip
4) same as above but all sites on different ip's
After this I will probably come up with something similar but running on IP's from different countries(us/eu/etc) and see if that has any effect.
I will review those results and take it from there. If sites index but I am not seeing any serp movements then I will create equal backlinks which shouldn't take many as they will target very low competition keywords. but it isn't hard to give them the exact same social bookmarks, profiles, auto approve blog comments on the exact same sites.
As said multiple times I do know this is basic 101 stuff, but all I have done is read and I want to see it with my own eyes. This is more for myself but figured I would post the results in case anyone else found them useful.
