It used to be that you had to worry about data fragments in your word processor files giving away information relating to early drafts of press materials.
Now online documents have a similar kind of problem with Google. I got a narrow escape when I worked at Yahoo!. We built a site that was supposed to go live in early January live on the net. This made testing easier, but consumers or the media finding the site earlier would have been buzz-kill to a project that eight weeks of PR activity to be squeezed out of it and was in the planning since July the previous year.
I also managed to recover a number of pages from the Mark III version of this blog via the Google cache after the Yahoo! server it was hosted on had gone to the great data centre in the sky.
Google caches regularly updated sources like press rooms and blogs so there is a good chance early content will be slurped up.