The search engine optimisation profession is very much an art rather than a science. SEO professionals try things out, comparing things like keyword density between sites and between pages to see what works best. They also spy on successful competitors, follow the advice of experts and perhaps even read up on the very broad details that are known about search engine ranking algorithms. But at the end of the day decisions are based upon gut feel and received wisdom: art rather than science.

In contrast, a scientific approach would have a theory together with data to back up the theory and to inform SEO decisions. Of course, it is impossible for any SEO process to be foolproof, otherwise everyone would be number one in Google, but employing a scientific approach as far as possible clearly gives a massive advantage over those restricted to the traditional art of SEO. This blog advocates a scientific theory of SEO, explains the scientific SEO theory and describes how it can be applied in practice.

How Search Engines Rank

Before introducing the theory of SEO, a little background information about search engines like Google is useful. Search engines try to identify the best pages for each search: the most useful to the searcher. Because they can’t possibly know in advance which pages are the most useful, they employ proxy measures, typically based upon the actions of other web users. One proxy measure is the proportion of users who click on a given search result and then don’t come back to try another: presumably these are mainly satisfied customers. This is a measure that SEO professional can do little about, other than through ensuring that the site they are optimising has good contents so that visitors will be happy when they find it.

Another proxy measure of usefulness is the number of links that web site designers have created to point to a given web page. Since the emergence of Google in the mid-1990s, hyperlinks to a page have been a key indicator of the importance of the page and have been used in search engine ranking algorithms. This is because people tend to mainly link to pages that are useful or respected in some way: perhaps they contain authoritative information or sell a relevant service or product.

In addition to importance, search engines need proxy measures of “aboutness”. For example, this includes assessing how well keyword search terms match the contents of a page. For this, the obvious source of evidence is the words in the pages, but the text in the anchors of links to pages is also important. If many people link to a page and the text of the links mostly report “Microsoft’s home page” then Google can be pretty sure that the target of these links is Microsoft’s home page, even in the page itself does not clearly state what it is.

Naturalness

Most SEO professionals probably know roughly what search engines are looking for and try to deliver it in any way possible. For example, an SEO beginner might stuff a web page with 999 copies of the word “zoo” in order to match a “zoo” keyword search. But search engines do not like this because having so many copies of the same word is unreasonable and unnatural: most searchers would not be impressed. Hence there is probably an optimal proportion of “zoo”s to go in a web page: enough to convince the search engine that the page is really about zoos, but not so many that the search engine or visitor would suspect SEO subterfuge.

The key principle here is naturalness: the best proportion of “zoo”s in a zoo page is that which would naturally occur in a page mainly about zoos. With its billions of web pages Google has a fair idea what the answer is and so do SEO professionals through trial and error. Hence on the web there are many recommendations about optimal “keyword density” that a beginning SEO professional can discover.

The above optimal keyword density issue plays out in a similar form in many different arenas. What is the optimal number of keywords to put in a link anchor text? How many links should one site have to another? What is the optimal length for a home page? The answer to all is the same: whatever is natural. Anything that is unnatural risks being blacklisted or penalised as probably being spam.

Search engines seem to have recently raised the stakes in the search for SEO pages, seeking extra dimensions of unnatural behaviour. Today you may be caught out because a wider analysis of your site reveals an anomaly: most of your links were created in a short time; most of your links have the same keywords; most of your link anchor text has the same length; most of your pages have similar proportions of keywords or similar lengths. This dramatically increases the complexity of effective SEO: now the objective has to be generating a natural environment for your site rather than getting each component (e.g., links, keywords) right.

The scientific SEO theory

Like all the best theories, this is extremely simple: effective SEO comes from ensuring all aspects of a site mimic, as naturally as possible, an important site about the product or service offered.

But how is it possible to follow this theory in a scientific way? Only with data from the web that can reveal what “natural” means for keyword density or any other aspect of SEO activity. With such data, the theory can be applied systematically to guide SEO decisions. This blog will explain in later posts how you can get the data necessary from the new influence mapping initiative Linkdex and how to start using a scientific SEO approach.

Share/Save