1. Pick a blog with some substance. Sure, it's fine to write about TMZ.com, but keep in mind that there may not be many intelligent things to say about Lindsay Lohan's latest arrest. (If you have already decided to write about TMZ, don't worry: I'm not going to punish you! But make it interesting!)
2. Consider dividing the elements of analysis among your 4 posts. For example, here's a typical post from Mashable: Inbound Marketing vs. Outbound Marketing. There's much to say about it: I could address the typical audience for Mashable, the reliance on an "infographic," or the way in which this "story" is really just a summary of another Internet source. Instead of trying to address all of those things in one blog post, you should distribute these elements of the blog among your 4 posts, and by doing so, you ensure that each post is focused and informative and definitely not redundant.
3. A blog's audience is an essential part of your analysis. For the Mashable post above, you should absolutely devote one of your posts to a discussion of how Mashable situates itself as a blog about social media for professionals. In other words, this Mashable post talks about marketing using social media. So, this blog is clearly not devoted to an audience of political radicals, such as the people at Anarchist News, and it is obviously not written in the snarky voice of Wonkette or for the hipster audience of Pitchfork.
4. You can discuss blogs that are primarily visual. It helps to use examples from blogs that are primarily textual, but you don't have to ignore the visual component of blogging. Take the Mashable example, again. Here's the large infographic in that post:
But you could talk about the cartoonish quality of the image, some of the particular statistics provided in the image, or the premium placed on "newness" (in this case, the "new" marketing of social media) in social media circles. Of course, this graphic is rich with textual detail. Pictures of Lindsay Lohan: not so much. Think about the kind of analysis you can perform, before you select certain blog posts for analysis.
