Analyzing an Experiment in Blogging
Since October I've been experimenting here with some personal blogging. Why, you might ask, when I already blog at my job? Isn't that a busman's holiday? Perhaps. But I had plunked down money for this domain, and I had some ideas and obsessions to explore that didn't fit in with my work. And I also wanted to conduct a few experiments. When a blog is housed within a major news site, the metrics get hard to sort out. With some great content and breaking news, and a huge built-in audience, it is a simple matter to draw millions of views. (Palafo.com has drawn under 5,000 views in its entire existence, with who knows how many hundreds of those clicks attributable to family and friends.)
Blogging alone is a lot tougher, as some smaller news outlets and out of work journalists may be discovering the hard way. You have to rely on tools found in the wild -- basic search, trackbacks, Facebook, Twitter, Delicious, Google Reader, LinkedIn, Digg, reminding friends at parties that you have a blog, etc.
As it turns out, the free host Wordpress.com offers some pretty good measurement tools on the back end. They won't let me use Google analytics -- how irritating -- but the stats they provide are interesting. (No measurement of time spent, unique users and repeat visitors, or other ways to judge engagement, alas,)
Take a look at the chart up top (click to enlarge it). It shows day to day traffic for the last few weeks. Basically, all you need to know is that the peaks are when I blogged. The valleys show up when I took a break. No content, no readers. Simple enough. Without posts, the traffic dives off a cliff. This is one reason big commercial sites (both mainstream and indie) often blog shotgun style, throwing as much content to the search engines and feed readers and social networks as they can, 24 hours a day, seven days a week. Over time, you do get some repeat visitors, but the Web audience is pretty fickle. They come for the content, and they don't care too much who you are.
The peaks and valleys are more obvious in this week to week chart (click to enlarge):
As you can probably guess from this chart, without looking around the blog, my posting dropped off in recent weeks, from about two or three times a week to once a week. Some work projects came to a head, and I found it more rewarding and easier to Twitter in 140 characters for a potentially large audience than research and write complicated posts requiring photos and so forth. So I never did the planned posts about the New York Comic Convention, the trip to the Spa Castle in Queens, and any number of food-oriented posts. (There's something about blogs and food.)
It was particularly labor-intensive because I was mostly writing about podcasts, which required hours of listening to audio,music (ditto), blogging/social media/books (hours of reading and Web surfing) and single-source coffees that required comparison shopping around town.
Before that wore me down I did learn a few things about what drives traffic to a little blog like mine in a far corner of the Web. Let's look at the all-time top posts (click to enlarge):
The all-time top post was my advice on a computer problem I encountered: how to get rid of annoying IM coho bots. More about that later.
The No. 2 spot is taken by a list post of my favorite blogs. Web readers love lists, and bloggers love to be put on lists. I had not quite realized the significance of automatic trackbacks, but a lot of blogs use them, so when you link to them, they link back to you. Bloggers themselves will pay you a visit to see what you are saying about them. It is still a thriving form of social media.
Then there was my bio. Not surprising. Just about everyone landing here probably looks at it once.
My list of iPhone apps, updated a few times, also proved surprisingly popular. I put it out at a time when people were having a lot of trouble figuring out which apps were worth using, and there were hundreds of new ones. Plenty of other bloggers had the same idea. It helps to be an early adopter. That list is probably getting a little stale now. I've lost interest in tracking down every single cool app, now that I've settled on the set I need.
The biggest overall topic is podcasting. There are many directories but few that approach the topic in a systematic fashion. My approach was entirely idiosyncratic, and I would have stopped if I hadn't discovered a small but interested audience out there. Podcasters, even commercially successful ones, are rather unreliable about posting reliable show notes or blog posts about their content. And as much as I love the iTunes store, the podcasting area is a bit of a disorganized mess, perhaps because the content is mostly free. That leaves a search void.
My coffee blogging also proved "popular" in the aggregate, because it was aimed at obsessives who are served by a network of blogs and sites that have been going out of business in the economic downturn. While many coffee experts have tried to blog, their expertise tends to be in making great coffee, not writing or blogging. There's definitely an opportunity out there for a good writer who loves coffee and knows more about it than I do.
Any blog post about Twitter is bound to be a hit, especially if you mention it on Twitter. I know, having clicked through to a bunch of them. (The Jan. 23 one about my rules for following on Twitter is the high starting peak in the chart at the top of this post.)
The only real surprises on this list were the N+1 post, about a slightly obscure literary magazine with Luddite pretensions, and the "thoughtprints" post, about a very obscure theater production. Neither had a particularly good Web presence, so these posts filled a void in search results, apparently.
On to the top referring sites. The results below (click to enlarge) taught me that I was better off depending on the curiosity of strangers than the kindness of my friends. The numbers don't lie. Twitter, an open, public platform, wins hands-down, over Facebook, a mostly closed platform where only my friends see my stuff.
Now, something doesn't quite add up here. These stats don't match the larger views listed by the post. But that's often the case with Web metrics. They are suspect.
During this period I had about the same number of Twitter followers as Facebook friends. I promoted links to my blog on both sites -- probably a little more often on Facebook, thinking people who knew me would show more interest in my stuff. Facebook is a closed system, and only my friends can see my profile. Twitter is open and even shows up in search. But Twitter followers far outperformed Facebook friends on click-throughs. Perhaps they prefer to stay on Facebook, chat and look at each others' pictures. Twitter users seem to be more actively seeking out content.
The biggest surprise may be that Mahalo referral, which keeps on giving. I posted an answer on Mahalo about how to get rid of the instant-message coho bots, with a link to my longer blog post about it. Not only did that answer drive a lot of traffic, but a link to my post has been posted on numerous other blogs. Happy to help.
The rest of the referrers are an assortment of individual Wordpress tags, people clicking links in email, Google reader RSS shares, stumbleupon links, and so forth.
Now, what about search? It doesn't seem to have driven a lot of traffic (click to enlarge):
Not surprisingly, many lazy people just type the name of the blog in the search panel rather than bookmarking the site. I do the same thing. The top searched term on Google has been "Yahoo" for many years. This is one reason I picked a short, unusual name for my blog that (I hope) is easy to remember. The other terms are assorted podcast, coffee and blog topics that I briefly mentioned, including the unusual phrase "janky vegetables" from the "Faire du Camping" episode of You Look Nice Today, which is not janky at all.
Most of the few incoming links were trackbacks from posts or blogs I mentioned, and stuff related to the instant-message coho problem.
Now, of course, it is a truism on the Internet that if you send people away with links, they will come back. Where did this blog send people?
Click to enlarge the chart. The greatest beneficiary here is my own Twitter profile, followed by my Facebook profile.
The other links are mostly blogs from my list, podcast sites from the reviews, and assorted links that have appeared in the feeds at the left of the blog. (Wordpress makes it very easy to share links and feeds from Twitter, Facebook, Delicious, Google Reader and so forth, without having to manually post anything here on the blog.)
The most interesting result had nothing to do with traffic here on the blog. I started posting a lot on Twitter in part to promote this blog, as well as share other links I found while looking for stuff to write about on the blog. Then people started following me there, I became part of a community, and I ended up with a bigger, more reliable audience there than here. Click on this Twittercounter chart, for the last three months:
That's remarkable. I'll be thinking about Twitter some more and eventually share thoughts here on the blog that require more than 140 characters. I could obviously use the traffic. :)
Update: After two blog posts, four hours and some promotion on Facebook and Twitter the chart was happily spiking again (at right). Most of the clicks came from Twitter, followed by Facebook, Google Reader and assorted tags here on Wordpress blogs. Plus one click from Mahalo Answers to the IM coho post.