By using the BlogPulse trending tool, we can have a look at the attention given to links that turn up on memeorandum. The following graph shows the time series (percentages of blog posts containing the URL) for the URL at the top of tech.memeorandum at 12 (noon) on the first of the last 5 months.
Not too surprising, perhaps. We see accute spikes with very low follow on posting.
The next graph shows the same experiment for the politics page on memeorandum.
The key observable difference here is that firstly, the general case shows post volumes approximately 50% of that in the tech page (approx <0.0025 vs 0.005). Secondly, there is a single large story. This last observation is most likely an artifact of trivial sampling used.
We can compare this data with the same for BlogPulse' top links:
A clear difference here is the fact that the peaks in BlogPulse are more diffuse. In other words, there is a leading edge prior to the peaks and two of the sample data show clear long follow on linking.
I believe there are two observations to be made here: Firstly, the algorithm used to promote links/articles to the top of such lists determines their leading age (the trend prior to appearing in a promoted position). Secondly, the type of link/article promoted determines the trailing edge (the trend after the link appears). Both of these features are things that any user would want to control. If we don't get that level of control, we are all going to be reading news in a reactive, knee-jerk, echo chamber. Imagine the value of being shown a post and being told, with a high degree of confidence, that this article is going to have significant staying power. Is that a conversation you'd like to be involved in?
Matthew, memeorandum is a news site. So the issues themselves have staying power even though the particular story links don't. That's how news works. Perhaps you should try your method on CNN or NYTimes top links and see if the results remain "reactive, knee-jerk, echo chamber" (lots of vituperation there, buddy!)
Posted by: Gabe | February 27, 2006 at 04:18 PM