For years, the program analysis community has accepted that runtime enforcement is a technique that is in many ways preferable to static analysis and is amenable to deployment, as long as we keep the runtime overhead low enough. In this talk we will try to understand what actually gets deployed and when. Given that the cost of independent runtime enforcement mechanisms is usually additive, we will be confronted with the question: can we afford always-on runtime enforcement? This talk will summarize some of the experiences we have accumulated as a community and propose some insights.