John Gruber speculates that SEO-style titles don't work for SEO, and goes on to state that they're useless. This author reads a Google starter guide that says search engines read titles and that you should come up with good ones, and states "I think it's safe to say" that SEO-style titles actually do work.<p>There's little to no science happening here. It's amazing that despite search engines being such an important part of our lives, we still aren't sure about this basic question.<p>In a more ideal version of the world (that is, better for all of us except black-hat SEOs), Google's relevance and anti-gaming algorithms would be strong enough to withstand being opened up. But in reality, Google warns [1] that full knowledge would open the floodgates to search-engine spam, while the black-hat SEO industry already thrives on the meager number of flaws they stumble upon themselves.<p>So barring openness from the inside, the next best thing would be for scientific experiments to figure out what matters, attacking the problem from the outside. Though some SEO experts try to do this, many of their articles just retell folklore, or micro-analyze offhand gospel from Google employees, or run worthless experiments that end with the author reasserting a hunch. (To be fair, setting up a valid experiment is difficult. Algorithms change without notice, different data centers return different results, the rest of the Internet changes constantly, and there's no control Google. Physics, at least, runs the same from different locations.)<p>The best, most rigorous way I can think of to test hypotheses would be a set of carefully controlled pages with no outside competition (Googlewhacks). Something like this must have been tried before. If it has, can anyone point to it?<p>[1] <a href="http://googleblog.blogspot.com/2009/12/meaning-of-open.html" rel="nofollow">http://googleblog.blogspot.com/2009/12/meaning-of-open.html</a>