MediaVision Blog : Search Engine Marketing

Perhaps you’ve heard of siloing, which is a method of categorising content on your site, revolving around a central theme. You determine that theme and its sub-topics by looking at the target key/search phrases for your business and its relevant market. You then optimise your site’s internal PageRank by linking the “silos”, or as Microsoft would say, the “blocks”, together using rel=”nofollow” tags on pages that are unimportant in relation to the main theme. Ideally, this means that Google would crawl only the quality internal links to the site, and derive the PageRank successfully from the upper-tier pages. The thinking goes that on account of the nofollow tags, PageRank is siphoned off into the most “searchable” pages.

Now, in an exciting patent published by Microsoft in April, and reported on by Bill Slawski, search engines are striving from the other end to give pages value using an algorithm that takes into account both features of blocks or topic segments, as well as human input determining their importance among surrounding web pages. In other words, if webmasters silo their sites, Microsoft will combine a human’s pyramid of topic importance with “an algorithm that builds a statistical model based upon features of blocks” during the ranking process. Obviously, in this way, advertising is more effective, link value is more easily determined, top level pages can appear for less enabled mobile search devices, user interaction is improved and by extension, conversion rates rise.

Up until now, and certainly in our experience as an SEO company, the nofollow tag is best used with caution. Originally created to eradicate spam in blogger comment boxes, the tag has evolved into a tool that can (emphasis on can) be used to direct crawlers. SEOs have complained that Google is inadvertently forcing people to use it for improved rankings, while others state that Google still follows the links despite the tags; and in some cases, might not even strip the “link juice” out based on the trust of the landing domain and the quality of the anchor text.

Experiments with siloing and SEO have been inconclusive, but what we do know, is that it generally lessens the dilution of PageRank that occurs between internal pages of a site. If ranking algorithms now take into account interoperability between pages, external links will rise may value. If in the past crawlers have indexed “noise” pages but left them out of a site’s overall rankings, siloing has had a measure of success. Following Microsoft’s patent (and Yahoo! and Google’s corresponding steps) siloing could now be instrumental in the infrastructure of a navigable site.

“Slawski (wisely?) refrains from speculating on the macro impetus behind Microsoft’s venture into Block concepts, other than to say that he suspects they’re doing it in order to index more than just content. In other words, they’re looking to evaluate input fields, images, links connecting blocks and forms in a website, which is certainly avant-garde in terms of search engine capabilities. In a sense, (and if Microsoft pulls if off) it’s to every person’s benefit. In the future, sites may be ranked by taking into account the fundamental information on them, including their design elements, instead of the ad copy/privacy policy/about us/noise information that floats weightlessly on the site. Moreover, searchers can find the pages, because search engines can understand them.