Mosaic Cloaking: An Explanation

I realize that posting a technology trademark piece on a political blog is not necessarily on topic, but I’ve got another piece coming up right behind this one, so there.

I am posting this here and now, because I want to set the record straight before anyone else goes off and tries to poach the term. I’ve been talking about mosaic cloaking in the abstract for a while, but the term was actually coined by the Fantomaster himself in September of 2007 when we were having a discussion about it at the Syndk8.

Mosaic Cloaking is essentially componentized IP-Delivery. In other words, you do not cloak the whole page, you cloak portions of it dynamically. In the simplest example, you have a page where you have a blank space. Based on the IP address of the inbound surfer, you may provide keyword rich text, or Adsense, or an affiliate ad – based on the IP address of the surfer. If you do not want Google, for example, to see that you are affiliating for then you might push the code for AdSense out if you sense that it is a Google spider – otherwise you send the affiliate ad.

More interesting mosaic cloaking is sending different javascript, CSS files, or even slightly different components of each that might be used to enhance the “experience” of the surfer, be it a search engine spider or human. Using comments and other tricks, your source files could be exactly the same byte-size as cloaked files, look the same, and only be available in the “spider version” if they came from a validated IP address. It would be extremely difficult to ascertain exactly what was happening, if you were trying to reverse engineer someone’s SEO strategy. Combine that with Google’s Multivariate Testing notions (ie., sending different pages to regular surfers under the guise of market testing) and you have a cloak that is extraordinarily difficult to pierce.

The best mosaics are a combination of surfer experience abstraction, with differing chunks of obfuscated Javascript that has been byte-size balanced, differing CSS external files and IP delivered from the same place dynamically based on a proven spider database such as the Fantomas spiderSpy product.