You are right, it means nothing, but winning by such a margin (and it would be true for Romney) is hardly a statistical beating, yet I suspect it will be treated as such.
The issue is that, given the current system, the popular vote isn't necessarily representative of the proportion of people that support Obama vs Romney.
The fact that votes are meaningless in many states (since they are already so heavily in one camp) likely has an impact on participation, and if the decision was made on a true popular vote the current splits may not hold.
I'm not saying the impact is one way or the other, but the popular vote tally under the current system isn't just meaningless because it doesn't determine who the president is, it's meaningless because it's unlikely to be a true indication of the popularity of either candidate.