http://www.moskalyuk.com/blog/yes-50-scientifically-proven-w...
(This link is via HN, naturally: http://news.ycombinator.com/item?id=613402)
Some of the tricks are really narrow and so they aren't all that useful themselves (e.g. to write post it notes and put those on interoffice forms and surveys for a better response), but with the theories you can make better extrapolations.
[EDIT: thanks to ironkeith for the Cialdini link ... it was a book by him and it was in a section devoted to the "Social Proof" research]
I have a similar reaction to "you should follow me on twitter". Before knowing about this test, it possibly made me a little less inclined to like/trust you than otherwise. (I think it did, but I don't know how to measure that.)
I wonder if that's a common reaction: maybe there's an uncanny valley-type effect, where more efficient phrasings become turnoffs? Does the high conversion rate come at a cost of slightly alienating a subset of your readers?
"You're following X, Y, and Z, who are also following me; you should too."
I don't really like the tone I wrote, but you get the idea.
"You should join my 4,615 followers on Twitter _here_. Everybody uses it. All the interesting people follow me. You really need to follow me now. I can't believe you're reading this when you could already be following me. DO IT DO IT DO IT. Fine. Enjoy your miserable life without me and die alone. Don't you want to be popular and interesting?"
Am I doing it right?
But if the info could be tailored more (e.g., "2,437 HN subscribers that follow me") I would be more interested. In other words, not everyone finds following the most massive crowd a sign of value.