Is SEO Science or Art?
Recently SEOmoz posted about running a test that proved their thesis that nofollow PageRank sculpting still works (while actually only proving issues with sample size & methodology). But the issue of “proving” things with SEO publicly is typically a misguided one.
It is so hard to control variables in tests, and even if you could set up a control set many test types would be isolated to fictional words. But the issue with that is that the relevancy algorithms can change based on your location, the location where a particular keyword is commonly searched from, how many other competing results there are for that query (and what those other sites are changing while you test), and whatever algorithm shifts happen in Google while your test is going on (like promotion of certain vertical databases, baking in new pieces to the relevancy algorithms, improvements in related vocabularies, introduction of new penalties and filters), etc etc etc
But lets ignore all the above and pretend there is a way you can isolate variables or you notice something new and different and important. What happens when you mention it? Typically people tell you that you are full of crap, even when you are right.
Posted in Search Engine Optimization, SEO