SVN's model is what everyone is using. Sure you git, but almost nobody is using the distributed parts - they all sync to a central sever (github). SVN just could get user management, or merges right - those should be solvable problems but somehow were not. (I don't know enough about SVN to speculate on why they didn't)
The main reason is that SVN's model of branching/tagging is based on directories in the working directory, whereas git's model (and to a certain extent, commercial SVNs like Clearcase and Perforce) is that branching/tagging is of the entire repository and not related to the file tree structure.
This is a fundamental difference and the reason that git's model works much better when branching/merging.
And then you force people to change from google code to something else just to prove a point since people then where unable to setup svn server ;)
Even today you find death links from google code repos.
If I recall correctly, SVN was also more popular than Git at the time, so migrating hosts was a lot more painful than now...