An ancient subject is being discussed anew at the moment: cloaking. Cloaking is about giving the searchenginecrawlers different content than the users. Back in the day, when On-page-factors were still extremely important for searchengines like Fireball and AltaVista, cloaking was used quite often and has had a negative aftertaste ever since. Today however, Google is publicly recommending their news-suppliers for Google News to cloak content and is kind enough to directly supply the IP-network of the Googlecrawler – 66.249.64.0/20, meaning the 4.096 IP-addresses from 66.249.64.0 to 66.249.79.255. It is also quite often that we get to hear that “cloaking-away” of session-ids is not only tolerated but is being recommended by Google.

So how should you stand on Cloaking at the moment? As long as the user is basically getting the same content as the searchenginecrawlers, right now, it seems to be practicable as long as there are no other methods available anymore. If I think about an Intershop-installation for example, a SEO often has no other choice than to remove session-ids of these shopmonsters for searchengines so that they are included into the index or that duplicate content from sessions is prevented. However, you should be aware of the possible risks and the effort involved: cloaking is still on the “Don't do it”-list of the Mountain View spamhunters and the upkeep of the IP-lists that are needed for meaningful cloaking can be very time-consuming – some Korean robot that gets stuck in the session-id-loop has the potential to ruin your whole night if you are the one stuck with the cell-phone that gets the server-monitoring messages. So: consider it as your last possibility but only after you have exhausted ever other possible option. I would keep my hands away from the method which is being used by large scientific-publishing-houses – especially in the USA – at the moment: they are feeding PDFs of their publications to the Googlebot but once a real visitor shows up they are asked to pay for the information.