SEO is a Craft not a Science
Running Bumblebee by wwarby
Recently there has been a backlash in the SEO industry against what is deemed to be a non-scientific approach when it comes to SEO testing or collecting statistical data. Some people argue that you have to treat SEO as part of computer science and be very strict about it. Following this logic you are basically forbidden to do any SEO testing yourself unless you have some actual scientists doing the research.
While I am not sure why some people attempt to push in this direction I can only say it won't work. SEO is a craft not a science. First off there are far too many variables to determine in SEO to be able to treat it scientifically. Also, if anything, SEO would have to be a social science as most things in SEO depend on human behavior. People writing algorithms, people linking, people creating content.
Also what is the subject of this science? It's a black box company not telling what is below the hood of it's magical engine. We know by now that there are at least 200+ signals aka ranking factors that determine whether you end up on number 1 or not. These are shifting permanently.
Matt Cutts tells us that there are hundreds of tweaks to these algorithms per year. So even if you would be able to single out one of the ranking factors to test without interference of all the others you still can't assume the situation yout test stays unchanged during the test. It's like a weather forecast: Science can compute what it wants and tell you that you will have sunshine tomorrow but I strongly advise you to look out of your window before going out. Otherwise you might get wet.
In case it rains you have to rely on your own eyes and react, wear a rain jacket and rubber boots.
That's how SEO works. No matter what pope Matt Cutts tells you or what some computer scientists have computed.
Last but not least science always serves a purpose. Depending on who pays for a study you get different results. Different expectations, even different backgrounds of the scientists lead to bias. Scientists were trying to prove the existence of God for ages. Now they look for a God particle at CERN/LHC. Science is more similar to religion than you think.
Of course I love the scientists and geeks in the SEO industry.
The people who can prove that something works or doesn't work or even better tell me why. On the other hand I don't care what others tell me as long I see results. When I do SEO testing in a DIY manner and I notice that something works or not I don't care for it being scientific. The results in rankings, visitors, conversions and sales are proof enough. Likewise I don't believe Matt Cutts when what he says "is not working" in reality works.
The purpose of SEO testing is finding out what works
not proving you are right it front of some kind of scientific authority. The people who try to tell us that we have to be scientific are in many cases self proclaimed authorities or do not even practice SEO. The purpose of publishing your SEO test results is not peer review in the sense of being bashed for not being scientific, it's finding out what works for others.
SEO these days is not about computing. It's about soft skills. You can't compute human behavior. You can just to some extent, just like the weather. Doing SEO is like knitting, baking bread or writing poetry: When people love it it works. Also remember the scientist who stated that the bumblebee can't fly due to poor aerodynamics. He later found out that he erred. He forgot to take the wing movement into account assuming bumblebees behave like airplanes.
SEO is not just computing.
Scientists should keep to trying to explain the world not telling the world how it should behave. That's exactly what some SEO science advocates are trying to achieve now: "Don't do this, don't do that it's not scientific."
Your task is to explain why something works or not. You are not entitled to give me orders. SEO is a craft not a science so please don't try to tell the bumblebee it can't fly.