Difficulties with scientific method in empirical SEO

SEO is not smoke and mirrors or a black art, as some SEO companies might have you believe.  Far from it, SEO is the application of best practice, taking advantages of fixed logical rules and algorithms to achieve best-value results for effort.  The problem for SEO strategists is that these logical rules are not public knowledge, and more so that they differ for diferent search engines.

 The headline to this article refers to “empirical SEO” which is a term I have coined for the process of determining best practice using what is refered to as the scientific method – i.e. experimenting and using the resulting evidence to postulate the rules by which best practice SEO should be governed.

In the SEO community, there are as many opinions about best practice as there are strategists to implement them.  Some of these opinions are speculative – protagonists assume a certain set of actions, sometimes despite the best evidence.  Other opinions are philosophical – protagonists postulate how search engines might react; they try to think like a search engine and guess how a search engine will react to different practice. A third group (and by far the biggest, it often seems) are the sheep, that blindly follow whatever the current popular trend topic of supposed best practice might be despite any evidence for or against.

However, probably the smallest group, distinct from most of those above are the empirical SEO strategists.  Empirical SEO is the method of theorising how a search engine might work, and then seeking to test the theory to seek to ‘prove’ or rebutt the theory.

For example, an empirical SEO strategist might try to determine the ideal keyword density for a web page by creating almost identical web pages with differing content, and collating the respective search results for each page.

However, the unique circumstances prescribed by Google can make empiricism within SEO a difficult methodology – ‘almost identical’ web pages really can’t compete fairly in Google’s index.  For one thing, if the content is too similar, all but one pages are likely to incur some kind of duplicate content penalty.  If the web pages are on the same doamin, there is the likelihood that Google will group pages into canonical equivalents, prioritising one over the others.  If the pages are on different domains, other aspects can affect performance – domain age, keywords in domain, and so on. 

Additionally, the empirical SEO has no real control over inbound links for the test pages.  It only takes one (well meaning or otherwise) third party to link to one of the test pages to skew the results (or at least invalidate any conclusions that could be drawn).

Unlike most branches of science, empirical SEO never has “laboratory conditions.”  Google can not be “isolated” in a lab and experimented on without any outside factors.

Therefore, it is important for the empirical SEO strategist to design experiments in such a way that it is still possible to draw conclusions, however broad brush they might be.  The proof is in repeating the method and seeing similar evidence of success over and over again, not in finding exact results in one fixed experiment.

This is precisely the method by which the Theory of Evolution has been proven over the years.  Most creationists will tell you “Evolution is just a theory,” which is the argument most quoted by the lobby in America (and growing in Northern Ireland, unfortunately) to teach creationism side by side with evolution in science classes in schools.  However, Evolution is no longer just a theory.  The methods of experiments can repeated over and over again, with the same results each time, building an overwhelming catalogue of evidence.  Creationism is just a story with no evidence – like many of the SEO opinions espoused by the speculative SEO strategists!

• Tweet
• "&title=""">