In a recent article in the University World News, Richard Holmes points out some of the anomalous rankings in the Times Higher Education and explains how the bibliometric methodology for citation metrics may be flawed for the purpose of rankings.
Holmes suggests that a university involved in 'multi-collaborator mega-cited' projects will see an improvement in rankings, to a point...
There are also some interesting criticisms of the Thomson Reuter products such as InCites.
Richard Holmes, How to win citations and rise up the university rankings, University World News, 07 November 2014 Issue No:342