Thoughts on the Dark Art of SEO

Search Engine Optimization, or SEO for short, covers all measurements that aim at making your website appear at a better position among the organic, or unpaid, results of a search engine when particular keywords are entered. In SEO terms, this is called making your website rank higher.

How search engines work

The reason behind the desire to rank higher is that the higher your website ranks, the higher the probability is users will see your listing and click on it to visit your website. Obviously, increasing the number of visitors will in turn hopefully increase the number of sales or signups, called conversions. Increasing the number of conversions usually is the ultima ratio driving SEO efforts.

To understand how SEO works, it is vital to understand how search engines work. A search engine like Google takes one or more search terms and gives a list of results in return. The list of results is ordered by relevance, starting with the results the search engine considers as most relevant. The compile the list of relevant pages, the search engine has to compute the relevance of each page on the web with respect to the search terms entered as a numeric value. It then sorts the list of pages according to their relevance value. Pages with higher values for relevance will rank better.

To be anyhow useful, the way a search engine computes relevance values needs to reflect the user’s common understanding about relevance. For example, when searching for the term ‘apple’, computing the relevance of both the website of Apple, Inc. as well as the Wikipedia article about apples should result in high values because the sole word ‘apple’ could refer to both. Contrary to that, when searching for ‘apple computer’, the user is likely not interested in web pages about gardening but rather in the website of the computer company. The relevance computation should therefore result in a high value for that site and a low value for web pages about gardening.

The computation method (or algorithm) the search engine employs for obtaining relevance should also avoid being spoofed. In the early days of the web search engines looked at text on pages only. Adding the text ‘Apple Computer’ somewhere on your site about gardening was likely to rank your site higher on searches for ‘apple computer’. Algorithms used for computing relevance in modern search engines do not fall for such simple manipulation attempts anymore. Instead, they are able to filter out the main topic a page is about, and to understand the meaning of (at least some) search terms as well.

What to base SEO on

SEO aims at improving rankings for a particular site or web page related to certain keywords. Strictly speaking, this could be viewed as cheating (even though entirely legal). As we all know cheating works best if you know about the system or person you cheat. If how search engines work were to be known publicly, cheating was so easy that search engines would become useless. Therefore, search engines keep their technology secret.

So what to base SEO on if we do not know about the technology to cheat? Luckily, as we have seen, the algorithms in search engines need to compute relevance like a human user would do it. Thinking about what users would prefer and how users would judge page relevance is therefore a vital part of SEO. Furthermore, search engines publish recommendations on how to build sites to allow good indexing.

SEO measurements

SEO measurements based on assumptions about how search engines compute relevance divide into five categories:

  • technical measurements
  • information design measurements
  • editorial measurements
  • networking measurements
  • continuous monitoring and improvement

Technical measurements ensure that the site’s contents are accessible for a search engine in a good way. That means that we do not encode or encrypt text content, that we do not protect any important images from downloading, that we program the website according to web standards, and so on. In recent days, even factors like page loading speed, server response times and so seem to be included in relevance computation.

Information design measurements care for that the site nicely splits into pages that search engines can summarize under a single topic. Furtherly, information design measurements make sure all the pages on a site are linked well along thematic aspects, that sites are split into sections or subdomains if they cover many different topics, and so on.

Editorial measurements affect the contents of the site. Well-written copy that goes nicely with the structure of the site and makes use of keywords your site should rank good on has proven to work best. Amending content with outgoing links to relevant resources on the web seems to work as well. In the times of social media and real-time updates, frequent updates with relevant content are also expected to make pages rank higher.

Networking measurements aim at achieving a higher relevance in search engines through linking your website from others websites, blogs, and social media like Twitter and Facebook. Direct links to your website from other websites (partner sites, for instance) are generally viewed as good ways to increase relevance. Mentionings in social media, on the other hand, direct people to your website. Even more they influence relevance computation indirectly as your happy new visitors might mention you on social media channels as well, which then again will count for relevance.

Not being a real measurement but a continuous task, monitoring and improving the site’s usage statistics on a periodical basis is crucial for keeping the site ranking high. Also, continuous monitoring allows to identify chances to improve rankings.

Desktop: 1025px and above

12 column responsive grid
Tablet: from 768px to 1024px

4 column responsive grid
Mobile: up to 767px

4 column responsive grid