Statistical Mechanics of the Blogosphere

That BaconCat guy has two interesting posts this weekend on the detailed workings of blogdom. The first is a closer look at the blogs on Technorati’s Top 100, and the second is a look at Big Posts and how they affect traffic. I have a few responses to these, which probably aren’t terribly interesting to anyone who isn’t already a blog obsessive, but that’s why I’m posting this on a Sunday…

Two things strike me as interesting about John’s look at the Top 25. The first is just how much movement there is in the Top 100– when I first followed that link, the ScienceBlogs front page was #26. Friday morning, when I first thought about writing something on the subject, we were down to #34 or so; at this writing, ScienceBlogs stands at #29. I’m sure the rankings of other sites in the same range are similarly fluid, but I don’t care enough to track them.

The other surprising thing about looking at the Top 25 John posted is how few of them I read. These are the most-linked blogs in the world, and yet there’s only one blog in the Top 25 that is on my current list of RSS subscriptions (and I don’t really keep up with that one), and four in the Top 100 (including ScienceBlogs) that I read with any regularity.

(Continued below the fold.)

John’s analysis pretty much explains why:

This suggests a number of things. One is that the blog world is already pretty damn corporatized and politicized, as 8 of the top 10 blogs are either paid blogs or political blogs, and one of the other two is also pretty politically active. This should not be news. Another is that if you want to crack the top 100 without writing on a single topic, especially politics or tech, it helps if you are a pretty girl, or someone whose online nickname has become a verb. Short of that, you’re pretty much on your own.

I’m just not enough of a tech fetishist to care about Gizmodo or Engadget, and I’m pretty thoroughly sick of most of the big political blogs. The one blog in the Top 25 that I have in my RSS feeds is BoingBoing, and I don’t even regularly read that (it’s updated too often– if I try to keep up in real time, I can’t get anything else done, and if I let it sit for a day, I come back to 174 articles about people who have added wireless network capability to their toasters). The other three that I follow are ScienceBlogs, Go Fug Yourself and Overheard in New York.

It’s not that I have a problem with single-topic or political blogs– plenty of the smaller blogs I read are narrowly focussed on some aspect of science, and I still follow a handful of political blogs. It’s more a matter of the “corporatized” aspect of things. I’m not fond of people throwing around “corporate” as a sort of catch-all epithet, but it’s not too far off the mark in describing the problem I have with a lot of those blogs: it’s the voice, or the lack thereof.

A while back, either at Boskone or Noreascon, I heard Cory Doctorow talk about blogging, and he offered his advice on how to generate lots of readers. One of the points was technical– full-text RSS feeds– but the others were straight corporate journalism: post titles should be descriptive, not clever, and the entire point of a post should be spelled out in the first paragraph, or better yet, the first sentence.

I thought about that for a while, and decided I couldn’t do it. That’s not how I want to write, and it’s not what I want to read. It pretty much strips all personality out of the articles themselves, other than the selection of topics, and that’s just not that interesting to me. The blogs I like best are the ones with a real sense of personality behind them. That’s why Making Light and the Whatever are prernnial favorites, and why my favorite science blogs are sites like Cosmic Variance and In the Pipeline. And that’s why you get dog pictures and lengthy digressions about weight loss on this site– I post about whatever’s on my mind, because I prefer reading sites by people who post about whatever’s on their mind, rather than narrowly focussing on some specific area.

I could try to follow Doctorow’s advice, and draw a million hits a day from people with the attention span of a mayfly, but that just doesn’t seem like fun to me. And this is first and foremost a hobby, something I do for amusement in my spare time. I’m not harboring a secret desire to be a journalist– college professor is as low as I’m willing to sink on the respect scale.

This sort of indirectly brings me around to John’s second post, which he starts off with his own advice about how to build readership: 1) Update frequently, 2) Enable comments, and 3) Post interesting stuff. It’s sort of sad that that’s the right order– you’d like to think that “post interesting stuff” would be first– but that pretty much fits my experience as well.

John also divides traffic-increasing events into two categories: the “LiveJournal Effect” in which lots of people link to a post, and collectively produce a big increase in traffic; and getting Farked (linked by one of the big high-traffice sites), which produces a huge spike in traffic over a very short time. He even provides graphs of the two effects, and we all know that graphs make it scientific. The main difference between the two, in terms of traffic stats, is that the LiveJournal effect tends to produce a smaller bump in traffic, but a larger percentage of that sticks around after the event that produced it is over. The Fark increase is much more dramatic, but tends to disappear after a day or two. That difference is very clear in the graphs.

I’ve started following my own traffic a lot more carefully since the move to ScienceBlogs (for cynical reasons– the payment I get for the site depends on the traffic level), and I pretty much agree with everything John says. I haven’t had anything picked up by any of the really huge traffic generators, because I lack the vision to attach meat products to domestic animals, but I’ve had a post or two linked by MetaFilter, and the effect is similar to Farking, only the spike is smaller.

The one thing I would add is that the LiveJournal Effect tends to be a little better at picking out the best material. Fark and Slashdot are looking for a quick jolt of weirdness, and then moving on to other things. They’re more about novelty than quality. The LiveJournal Effect, on the other hand, involves lots of small sites picking up a link because of the quality, and passing it along to others who pass it on for the same reason. The examples John picked illustrate this very nicely– the Being Poor post is a great bit of writing, while taping bacon to a cat is just a flash of daft inspiration.

Why am I blathering on about this? Partly because that’s what I do, but also because there’s been a bit of back-channel discussion about these phenomena in the ScienceBlogs community. One of my colleagues expressed some frustration over the essnetially random nature of blog traffic, where posts that you spend hours on sink without a trace, while silly, disposable fluff takes off into the stratosphere. Everybody with a blog has their own story along these lines, and we tried to reassure him that, in the end, quality does win out.

And I do believe that. I’ve had a few spikes here and there, but most of the increase in the traffic to this site (which has risen by 20-30% over the last couple of months) has been associated with posts that I’m fairly happy with.

While killing time waiting for some SAT Challenge material to come in, I poked around a bit, and looked at the traffic statistics for this blog in more detail. The most-viewed post in the history of this blog to date is How to Tell a True Lab Story. That could’ve been better– I slapped it together quickly one morning, and the writing could stand to be tightened up a bit– but I’m generally happy with how it came out.

The second most viewed post in the history of the blog is Top Eleven: Time to Vote, where I asked people to cast votes for the Greatest Physics Experiment of all time. The actual vote post isn’t a stellar piece of work, but it represents the culmination of one of the best things I’ve done on this blog. I put a lot of work into the experiment descriptions linked from that post, and I’m fairly proud of how they came out.

In the end, I think quality blogging does get rewarded with traffic. Not all of the people who clicked over here for “How to Tell a True Lab Story” or to vote for the Top Eleven stuck around, but enough of them did to build the readership of this blog up to a level that I still find a little boggling– I get somewhere between 1,000 and 2,000 “unique visits” a day, which is an order of magnitude down from Scalzi, but an order of magnitue up from where I was a year ago. And I’m glad to see that, over time, the posts that I consider my best are drawing the most views.

There’s still no escaping the random element, though: the third most viewed post to date is a cheap throwaway Chuck Klosterman quote that pissed some people off. Go figure.

(Well, OK, I knew that would get me some hits, which is why I put it up. I wouldn’t put it in my Top Five favorite posts of all time, though…)