@Philippe: I agree that the rankings are not perfect and one often knows which are the best journals in the field.
However, pretty much everyone of us (with the possible exception of the Nobel Prize winners, Fields medalists etc.) has encountered the situation when submitting to the very top journals makes no sense (i.e. the chances of getting a particular paper accepted there are slim at best), and so one should choose among the journals at the next level (second tier), and here things become tricky: while people usually agree which are the top journals in a field, there is considerably less agreement about the best second-tier journals, so numerical indicators like IF could come in handy but of course still should be used with extreme care and discretion.
@Philippe @Artur: Yes, no rankings are perfect. I think a good work should have many citations. So, what is a paper is more important than which journal/conference the paper was published.