Our framework also allows fractional selection of blogs, which means that instead of reading a large blog every day, we can read it, e.g., only one day per week. This also allows us to ask: what is the best day of the week to read blogs?
In order to study whether fractional selection allows to achieve better benefit cost ratio, we split the blogs which had at least one post per day into 7 blogs, one for each day of the week. Fig. 7(a) shows, that by splitting big blogs, the population affected (PA) objective function increases for 12% over the setting where only whole blogs can be selected.
Returning to the original question, we performed the following experiment: given a budget of 1000 posts, what is the best day of the week to read posts (optimizing PA)? We found that Friday is the best day to read blogs. The value of PA for Friday is 0.20, while it is 0.13 for the rest of the week. We consider this surprising, since the activity of the blogosphere (number of posts and links created) drops towards the end of the week, and especially over the weekend.
You can read the entire paper in these two PDF files: here and here. The logic here might be that bloggers tend to blog less often on Fridays and so what they do end up posting might be more focused and more informative about what is going on in the blogosphere than what you see during the rest of week. That is providing they aren't Friday Cat Blogging or blogging about their weekend plans.
