Often, after I give a talk, the feedback I hear back is "I understood your whole talk," or something near that. I take it as a compliment (although I don't think it's always meant that way).
When I present a talk to systems people, I think the statement is actually meant in gratitude. Instead of trying to force some challenging theoretical method or computation on them in the space of an hour, I've tried to give the high-level overview of what we've done (and why), and provide some simple and understandable examples behind the work that went into it. With any luck, there's an idea or technique in there they can use themselves sometime. I have some suspicion that many of them have suffered through a fair number of theoretical talks that have left them behind, and were grateful not to have to sit through one of those. Also, in my experience when systems people say they don't understand a systems talk that's a bad thing; in that case, it's often that there are some significant nagging details that haven't been discussed sufficiently that are making the listener suspicious that there's some flaw or an important side case that hasn't been adequately addressed.
When I present a talk to theory people, it's not always clear to me how to take that comment. There's certainly a subculture in theory CS that seems to think it's important to show how complex your result is (or, perhaps, how smart you are), never mind the audience. On the other hand, some theory results are so complicated it is truly a challenge to try to present a lucid 1-hour talk. In some cases, people saying they understand the talk feels like a real compliment -- thank you for presenting things in an understandable way. In some cases, it feels like a backhanded compliment -- if it's that easy to explain, you're not working on very hard stuff, are you? When people say after a talk they didn't understand it, it's more ambiguous -- did the speaker do a bad job, or is this really exciting new difficult stuff that will take some time to learn?
Generally, when I plan my talks, the aim is to make almost all of it understandable to as large an audience as possible. For specialized audiences, I'm happy to go into details, but for more general audiences -- the very large bulk of my talks -- I'll aim for simple when I can.