For some years now, I’ve largely ignored the issues of how many people read this blog, where they come from, what technologies they use, and so on. But today I took a side-trip into Google Analytics and I found the numbers interesting, so maybe you will too.
But first, these words · Yes, in this world you always have to listen to a sermon before they roll out the free food. Here it is: Don’t do this. Don’t obsess over your analytics or invest your time in worrying about how many people are coming or where they’re coming from.
The first reason is that it’s futile. If there’s anything that having been a blogger for the last eight years has taught me, it’s that you can’t predict what people are going to like.
The second is that it’s corrupting. If you start obsessing over your follower count, you’ll start writing what you think people want to read, not what needs to be written. Want a big audience? Go try to out-sleaze TMZ or Gawker. Failing that, tell the stories that ask you to tell them.
The Index · The data below cover the range between September 20th and October 20th, 2011. I didn’t have any monster hit pieces that got Fireballed or Reddited in that timeframe, but my DMR appreciation got linked by the BBC. So, a not-untypical one-month window in the life of this blog.
Tip o’ the hat to Harper’s for the format:
119,340 — Total pageviews.
68,907 — Unique visitors.
33,561 — Views of the most popular page.
20 — Number of pages with more than 1,000 views.
55.2%, 32.4%, 12.2% — The proportion of visitors arriving via link from another site, “directly” (bookmark or typed it in I guess), and from a search engine, respectively.
26.2%, 13.6%, 10.0% — The proportion of referrals from Twitter, Hacker News, and the BBC, respectively.
40.9%, 10.6%, 7.3%, 3.5%, 3.4% — The proportion of visitors that come from the US, UK, Canada, Germany, and India, respectively.
38.0%, 27.0%, 11.3%, 8.0%, 7.7%, 5.3% — The proportion of visitors that are using Chrome, Firefox, Safari, IE, “Mozilla compatible agent”, and the Android browser, respectively. Note that iDevices are hiding in the Safari number.
85.4%, 14.6% — The proportion of visits from non-mobile and mobile devices, respectively.
60.2%, 39.2% — The proportion of visits from iOS and Android mobile devices, respectively.
Backdrop · This is not perhaps among the first wave of weblogs but, launched in February of 2003, it’s at least middle-aged. I dive too deep into software nerdery to attract a real mainstream audience, but for what it is — a thematically-unhinged assortment of tech and politics and sports and arts and photographs — it’s about as successful as one could hope for.
The statistics above represent much less than all the traffic here. Feeds may never have really been embraced by the non-geek mainstream, but are a primary mode of consumption for the kind of people who are interested in the kind of thing I write about. I have a full-content Atom feed (photos too!) which has some tens of thousands of subscribers directly and via one intermediary or another. Counting them would be work so I won’t.
The Tool · Google Analytics, it’s pretty cool. But this is not gonna become a repeating feature here.
Contributions
Comment feed for ongoing:
From: Boris Mann (Oct 21 2011, at 23:17)
Twitter visitors in many cases are hiding in "directly" since many / most don't include referrer information.
[link]
From: Henning Hoefer (Oct 22 2011, at 00:27)
If mobile access really is more than 85%, why don't Mobile Safari and the Android browser have a much larger share in the browser stats? Could it be that mobile and non-mobile are mixed up?
[link]
From: UE (Oct 22 2011, at 01:07)
Does "mobile" include laptops? Otherwise I can't really consolidate the proportion of mobile devices with the proportion of the different browsers, i.e., 85% mobile but still 38% Chrome and 27% Firefox.
And if they include laptops: I must admit I am surprised how one can distinguish between laptop and desktop.
Ulrich (having read the article on iOS but commenting from my desktop)
[link]
From: Bud Gibson (Oct 22 2011, at 07:56)
I agree with UE, I'm having trouble figuring the chrome numbers given the mobile, laptop split.
BTW, I often read the feed on my mobile device (now iphone) and then star the link to come through here.
Now, on to a tangentially related point, the line has really blurred between web page as document and as application. Analytics can be invaluable when your web page is really an application, and I'm not sure where the difference is entirely anymore. Take for instance kindle 8 ebook format, which is based on html5.
Presumably, books are documents, but looking at kindle 8 and epub extended in ibooks, books start to seem a lot like applications. It strikes me that analytics could be invaluable in determining the effectiveness of document as application.
[link]
From: Nicholas Sushkin (Oct 22 2011, at 10:02)
Tim, can you check if you got the 85/15 mobile/desktop numbers backwards?
[link]
From: Laurent Courtines (Oct 22 2011, at 10:06)
Couple of things (If you are curious)
1. All RSS data is available if you use feedburner on the account you are set up with Google Analytics - Its listed under traffic sources -> campaigns (don't ask why)
2. All twitter traffic is visible under the t.co domain. Twitter consolidated nearly all twitter shortened links.
3. Great work on the traffic. My philosophy is write for yourself and if people read it? It's a bonus!
[link]
From: vitriolix (Oct 22 2011, at 10:09)
Count me in the RSS reading masses. I almost never actually click through and just read in the context of the reader app, like I imagine most android users I am statistically lost in these counts.
[link]
From: Tim (Oct 22 2011, at 10:15)
Nicholas; you’re right, they were reversed. Thanks.
[link]
From: Liam Quin (Oct 22 2011, at 19:59)
Interesting - I get about 3 times as many pageviews on http://www.fromoldooks.org/ (390,000+ in the same period, and over 120,000 visits)of which approx 2.5% were said to be mobile devices.
It's likely Google Analytics underestimates by not counting devices without JavaScript, or with JavaScript turned off, or using adblock, etc.
But I don't get any significant number of referrers from outside search engines really.
IE is under 25% of visitors, it seems, with firefox and chrome leading, and Opera down in the 2% region.
Years ago I wrote my own log analysis software and still use it sometimes, but it's really designed for sites getting under 1,000 visitors in a day, for troubleshooting problems and seeing where to add links.
[link]
From: Karl (Oct 23 2011, at 02:32)
Hi Tim!
First of all: especially geekish surfers usually reading your blog have disabled Google Analytics either by Script blockers or /etc/hosts dead-link.
Second: probably a large portion of readers are using an aggregator like Google Reader with mobile clients and such. I am not sure if you can get those visits using your method.
Third: if you are interested in web page statistics, I definitely recommend you http://piwik.org/ since it is open source, runs on your site (own your own data) and is not considered that «problematic» as Google Analytics (privacy, legislation, big brother, ...).
Shine on and greetings from Graz/Austria!
[link]
From: Chris Lambacher (Oct 23 2011, at 08:06)
Some of the "directly" are going to be twitter clients. I know that is the case for how I get to your articles.
[link]
From: Preston L. Bannister (Oct 23 2011, at 09:48)
I assume that reader that use aggregating readers (like Google Reader) are generally not counted. Same thing (and more so?) for the Google Reader app on Android. Or does Google Reader report the number of readers on whose behalf it is fetching, and do you multiply your numbers appropriately?
Of course, you can do kludgy things with in-page Javascript and/or 1-pixel images to try and force a more accurate count. Yuck.
[link]
From: Kristiono Setyadi (Oct 26 2011, at 06:34)
What if there are people who browse using lynx/links? Is it "Mozilla compatible agent"?
*just kidding*
I'm one of the readers who came from a bookmark as far as I can tell. Nice blog!
[link]