So I propose an experiment. I recently published in Scientific Reports. This journal's article metrics includes twitter postings. At the moment I show 34 tweets-- yes, it is trivial, but still puts me in the upper ~15%.

I'm going to set up the first two steps of this guy's formula (search for niche users; post interesting content) where the "interesting content" will be the article.

We can check back on my article stats in a week, compare growth in tweets vs growth in accesses/downloads.

This is all a bit silly, the important influence will be if this effort gets the article cited more often, and it will be a year before we have any data on that (plus no way to link the data to twitter activity).

I'll put a writeup on "show hacker news". Oh, the scientific article + metrics is here: http://www.nature.com/srep/2015/150302/srep08665/metrics