Ok, that helps me think a lot!
There are some variables there that were confusing me at first but that eliminates a couple.
With figures for posts like that there is a 3D aspect of it that should allow interpretation.
EX,
posts(x), New_Users
, External_stimulus(z)
External stimulus is vague though, you could correlate that with more than a few things. AdWords, Contests, shares(tweets 'n' stuff)
See what I am getting at? The math might also depend on those variables also. But if you just want a 2D representation of this it really will not show you much good data. As it would just be an average of increase/decrease but not why.
(after all we all know we have grown, but why?)
The problem with this that creates the most inaccuracy is the lack of "control" data. However this is exactly how you would think about forecasting stocks, as you used for example. The thing there is they have the same lack of control data, but use other data to base things on.
So for example, you could obtain data to compare from other iNet websites.
The trick is figuring out what you want your data to say!! Then worry about the math, and validation of such.