High Variance

My Second Blogoversary

In the past few weeks this blog has achieved two big milestones: December 21 marked the second anniversary of my first post, and on December 30, I published my 100th article. To mark the occasion, I thought it would be fun to reflect on the past two years quantitatively.

In 2012, I published 54 articles with 29,699 words, and in 2013, I posted 44 articles with 27,938 words–a decline, but not a major one. The average length of an article went up (550 words vs. 635 words) as did the standard deviation of the word count per article (296 words vs. 394 words). The increasing variance on this dimension can be seen visually in these two overlaid histograms:

So is this blog high variance in article length? That’s hard to say without scraping a bunch of other blogs and analyzing the data. Matt Gemmell says the distribution of his article lengths is bimodal:

When I’m not writing to a count, I have two natural lengths of piece:

  • 700 to 1,000 words for a compact exposition on a topic. This is also an ideal bite-sized chunk of text for the reader, and fits well on a single page. It’s no coincidence that the majority of commissioned magazine pieces fall into this band.
  • 2,000 to 3,500 words for a more detailed exploration or discussion.

The zone between the two is a death march.

I seem to have one natural length, but its range is a bit broader than Matt’s shorter pieces, as only half my articles fall within 369 and 724 words. If you include Matt’s longer pieces, he clearly wears the higher variance crown.

The yearly numbers mask an interesting monthly pattern. While I wrote a little less in 2013, I was more consistent in my output during the year. In both years I started strong, gradually slowed down, and then picked up the pace again in the fall:

Most article content seems to have settled into a small set of categories (tech, kids, education, and music. You could even make an argument that you’re reading a mixture of four distinct extremely low-traffic blogs with the occasional off-the-wall post thrown in. The below picture shows article counts by tag for each year for each tag that had more than one article. Many articles are tagged more than one way and every article I’ve published had at least one of the tags shown below. I wrote a bit more about education and Conan in 2013 and a bit less about music, but I’m not sure those trends will continue in 2014.

On the demand side, High Variance had 8,504 unique visitors in 2013 (up from 2,017) and 13,809 page views (up from 4,000). Chris Blattmann pointed his firehose my way in mid-April, but even if I ignore that week (961 unique visitors and 1,578 page views), readership is still way up for the year.

To celebrate, I’ve made a few minor tweeks to the site. First, I’ve added an IFTTT rule that tweets every time I post something new. You can follow me at HighVarianceNet. Second, I’ve made it easier to share content on the site by adding links for tweeting articles and sharing them on Facebook. Third, I’ve signed up as an Amazon associate. That means if you follow a link from this site to Amazon and then buy the product, you’ll pay the same price but I’ll get a few cents. Maybe by the end of next year I’ll be able to celebrate with a free cup of coffee.