Image by CRASH-candy via FlickrNewsflash #1: anyone claiming proficiency in any aspect of social media is making it up as they go along. Present company included.
Resist the temptation to alter your behavior because Social Media Expert suggested you do so. For every gem presented, the schmee (my word, run with it) probably exhibits one or more traits of deplorable activity. That should indicate to you the schmee (or this schmee) isn't perfect. Which is rather obvious, as this medium is designed imperfection.
But don't ignore their advice. If something is worth the wadded panties they suffered, it probably bears examination your part. They may not be perfect, but they aren't stupid.
Case in point: Chris Brogan's latest rant on "robot behavior". I'm long-time user of an auto reply for new followers on Twitter. Yet I agree with much of what Chris had to say. Counter-intuitive? Maybe. But I see auto replies as a tool; the difference is made in how you wield it. And just by typing the word difference, it is implied that that difference -- a change from state A to state B -- can be measured.
Newsflash #2: Yes, social media can be measured. In a myriad of ways.
In my case, the numbers weren't proving out my intended goal. So I modified my auto reply. I changed how I use the tool. In my case, I abandoned the first goal (not all that important to me).
If your auto reply on Twitter is designed to cause someone to take action; measure those actions. If it's working, great! Keep on keepin' on. But if it's not... well then you have some thinking ahead of you. Maybe your audience doesn't want to take that action. At least not in this environment. Maybe it's just not how they want to interact with you. Or maybe they really do only care about what you tweet. Not your free ebook. Not your blog. Not the "personal" video you made just for them.
Bottom line: not all auto replies are evil. But most are worthless. Because I said.