Thursday, August 21, 2008

So This Is What You Want?

Tēnā koutou katoa - Greetings to you all
So this is what you want.
I’m testing a theory. Google Analytics has been running on my blog for a clear month. I've been carefully looking at the data gathered over time and trying to make some sense of it.

Looking for patterns

I look for patterns in terms of what my ‘target audience’ might be interested in reading from my blog. There certainly hasn’t been any
earth-shattering evidence so far, but there are some signs and patterns that suggest a few things might be worth considering to do with what attracts visitors and draws comments.

Admittedly I haven’t got an especially wide range of topics, but within my own interest in education, and blogging about it, I’ve tried to be as diverse as possible with my post topics. Further to this, within the expected variation in frequency of visits over time, the popularity of the site has been fairly constant while the data was gathered and examined. I lay no claim to the statistical significance of data gathered from over 550 visits and 1950 page views from 384 visitors - only that some trends are suggested.

Post statistics table
A popular post

In recent weeks the single most popular post on my blog by far has been 5 explanations of a Zen proverb, an essay based on a saying that relates to my own life experiences. It took off, and topped the charts only a week after it was posted. Day by day, other posts slowly crept up in popularity. It is still a popular post, and this continued popularity meant that I was able to use it as a control of sorts against other posts that could possibly compete with it in popularity.

It was almost like watching a bubble sort in action. Slowly and systematically, every post related to one particular theme (and there were 4 of them) gradually worked its way up the list, each one pushing ‘5 explanations . . .’ down. And today, 21 Aug, that post has been superseded by all four posts. And the other popular theme? Google Analytics, and the analysis and blurb that I’ve written on what I’ve been fishing out from the data.

It would appear that personal stories are popular – nothing new here. But the interest that was focused on statistical data from Google Analytics did make me wonder about what people in my commentsphere are interested in reading.

Why you might be reading this post

My hunch is that if you are reading this, you have an interest in blogging. In particular, your interest stretches beyond the topic, with the possible exception of the one implied in this post.

I have left a space here, which I will update with stats I gather on the popularity of this particular post which, at the time of posting, would be the fifth, out of 62, that is to do with the same theme, namely statistical analysis of data on visitors to posts on this blog.

Study now completed: (stats for this post last updated 29 Aug 2008)

Post data table
Thanks to Sue Waters' and Christy Tucker's tips, I can post some alternative data from AideRSS on the popularity of these posts. Here's my so-called Best Posts list, recorded 28 Aug 2008. I think it is significant that those listed are the first and the last (this one) posts included in this study:

AideRSS data table

It seems my theory is supported by the statistics. This post, specifically created to test a theory, is now in third position in popularity and climbing, according to Google Analytics. It is placed top of the PostRank in AideRSS.

The most GA popular post (29 Aug) is DEANZ 2008, closely followed by Splitting the Knol. These have held their post rank for a couple of days - some posts seem to be like that. For instance, Compassionate Vigilance has held 6
th position throughout the time that this post was studied (21 - 29 Aug) with a favourable bounce rate and a reasonable time-on-the-page for a medium length post of 3:07 minutes.

My conclusion is that Google Analytics can certainly be a useful indicator of post popularity. Its potential for identifying patterns is evident. Used in conjunction with other similar applications, it can gather powerful statistics for the blogger. As this study attests, it can yield tangible information from which valid supposition and useful predictions can be made.

( 9 ) ( 8 ) ( 7 ) ( 6 ) << - related posts - >> ( 4 ) ( 3 ) ( 2 ) ( 1 )

Ka kite anō - Catch ya later


Anonymous said...

I wonder--is your Google Analytics post getting more search engine traffic than your Zen post? (The Zen post is great, BTW, I just can't come up with anything profound enough to justify commenting in reply.)

On my blog, the most popular post right now is about setting up Synergy. That's because I'm on the first page of results for "one keyboard two computers" and related phrases. However, even if that post if useful to many people, that isn't my primary target audience. Most of the people who find that post from the search engines probably aren't interested in instructional design or e-learning.

Does Google Analytics tell you how much traffic for a specific post is from search engines?

Blogger In Middle-earth said...

Kia ora Christy!

And thanks for the thoughtful questions and ideas - much appreciated.

The number of comments against posts is not really what I'm interested in so much. But these popular posts are toting over 100 visits each and accumulating.

At the moment I'm getting about 19% traffic from search engines, 28% direct traffic and 53% referred from sites.

As far as I can make out (I haven't spent a lot of time on this) GA does not tell me what portion of traffic on a specific post is from search engines. But the total number of visits on the site from search engines is just over 100, so I can assume that in the main the visits come from other traffic.

What do you think Christy?

Ka kite

Sue Waters said...

Unfortunately Google Analytics is only providing part of the picture of how people are interacting with your site. This is because Google Analytics only provides you information on people who are visiting your site and doesn't supply you any statistics on the other readers who are reading it in a feed reader like Google Reader or by email.

If you add Feedburner to your blog, add feedburner email subscription and integrate feedburner with your blog it will provide statistics on the number of people who are reading via feed readers. Which is also helpful.

I also like using AidRSS (which I think you Christy shared with us?) since it shows another totally different aspect of how people interact with our blogs.

Other interesting stats for Google Analytics you should check out is the average length of time on your site :)

Britt Watwood said...

I am trying out what Sue just suggested...we'll see what Feedburner analytics add to my understanding.

And BTW, your Zen post is still my personal favorite of yours!!!

: - )

Anonymous said...

Hmm. Does GA tell you what search phrases are most popular? That's really how I'm figuring it out on my Wordpress blog--I know which search phrases go to which posts. If the search phrases are mostly about Google Analytics or tracking stats, then you can guess it's more heavily from search engines.

I will second Sue's suggestions of Feedburner and AideRSS. Because of Feedburner, I know I have more RSS subscribers than actual views on my blog.

And yes, Sue, I talked about AideRSS as the "Quick and Dirty Comment Analysis" during the comment challenge--a post which means I now get search engine hits for "dirty comment." I bet people are quite disappointed when they see my non-racy post about AideRSS. :)

Blogger In Middle-earth said...

Tēnā koutou!

(this is going to be long comment :-)

I really appreciate the support everyone has given me with this study and the great advice. Thank you all!

Taking into consideration the comments that you have made already, this alone tends to confirm what I thought about the popularity of this topic with bloggers. It now remains to be seen how the visitor-count accumulates over all - it's not doing so bad already.

The points everyone's making about GA not counting all readers is perfectly valid, and (thank you) I do intend having a close examination of Feedburner fairly soon. I have suspected for a while now - and Britt, you alerted me to this - that visitors viewing my blog from their RSS wouldn't be counted in GA.

The thing is that any study using instrumentation (an this is what this is about) is valid when looking at comparatives as I'm doing.

The idea is to compare frequencies, rather than to count specific numbers of hits for any posts. While it is true that the 'real' number of visitors will differ from what GA shows me, it's really the trends that I'm interested in, at least at this stage.

@Sue - I have done some comparisons of the average length of time on the site, thanks. I was looking at this before. At present it is at 7:20 minutes, which relates to an average time of 2:15 per page.

This compares well with the average time spent on the posts looked at in the study, which ranges from 1:15 to 8:06 (for the Zen post). The times seem to follow linearly according to the length of text in the pages, suggesting that most people probably read most of any posts that they view, which is heartening.

Thanks for the link to AideRSS - it will be very useful :-) How does it work? I was gobsmacked when I saw all the stats on my posts! I will view this one with interest.

@Britt - thanks for dropping in - I'd be interested to compare notes with you on the use of Feedburner analytics when we've both had time to play around with this app. Thanks also for the support on the Zen post! I had no idea it would be so popular when first I wrote it up.

It was one of those things that I'd had at the back of my mind while commenting on other blog posts. I just jotted ideas down as they came into my head until I had something that could be shaped into a post - what we do I guess.

@Christy - yes GA does show which search phrases are used, though each post, of course, has its own analysis. Thanks also for your support on the Zen post. As I said above, it came as a surprise to me that it was so popular - but then I'd not thought too hard about it. I'd forgotten that personal stories are really what a lot of people ARE interested in reading.

What a hoot about the disappointments reaped from the 'dirty comment' search - I like that!

It's like the joke I was told of two wee girls who were new to primary school:

Sitting at the back of the class, Jenny whispered to Jane, "How old do you think Miss Jones is?"

Jane replied, "I don't know, but I know how to find out. You look inside her knickers."

"No you don't!" said Jenny in disgust.

"Yes you do", Jane asserted. "Inside mine it says ages 5 to 6." :-)

Ka kite

V Yonkers said...

I use Stat counter because it was easy to use and I couldn't get Google Analytics to work. However, I soon discovered that my own visits to my blog (which I am sad to say is the most frequent visitor) as added in. Then I forgot the password and username I used to set the account up!

What I have found interesting is the location where my readers are. But the statistics can only tell you a part of the story. I wish there was a better way to keep track of why people read (or don't read) your blog.

Blogger In Middle-earth said...

Tēnā koe Virginia!

What a bother you've had with Stat counter!

Checking that GA didn't record my own visits was something I needed to know early on, for I access my blog from different locations, and regularly. Luckily I discovered, through a few testing experiments, that it didn't count my visits, unless of course I visited it anonymously, which I never do.

One good feature of GA is its ability to show the geographical location of the visitor. It is specific to the city and even the suburb, which is interesting.

But using a machine to find out why people do what they do is always going to be difficult. I'm quite impressed with what GA does show, and though it has its shortcomings, I think that any method would.

Ka kite

Sue Waters said...

@Christy I'm thinking I need to write a post on dirty comments now :) Thanks for introducing me to AidRSS.

@Ken Here is my post on AidRSS which explains a bit about how to use it. I just love the way it provides such a different other picture of how people interact with your blog. For example here is the top 20 posts on The Edublogger and see the difference between commenting, linking and bookmarking.

Blogger In Middle-earth said...

Tēnā koe Sue!

As you can see, I've been pottering around with AideRSS - thank you!

When I said in my comment to you "How does it work?" I really meant how does it do it? I'm still a novice with this blogging thing :-) as you can tell. Thanks for all the links to these places I've never been ;-)

It's pretty self explanatory, the data from AideRSS. But it puts this post at the top of my list of great posts!

How cool is that?

Ka kite

V Yonkers said...

Actually, I get a map with statcounter also. The problem I have with it is more of my own making (there is an option not to count your own visits, but I forgot my password!). I'd like to try google analytics, but I can't figure out how to get the statistics!

Blogger In Middle-earth said...

Kia ora Virginia!

I guess you know where to go to find out how to set up Google Analytics/ on your blog site. Sue Waters did a good job on that post.

Once you have followed the instructions to put the java script into your template and saved the amended template, you’ll have to leave it a day (usually 24 hr) to see any stats as it starts accumulating, day by day, after that time.

Access the Analytic Settings page and choose ‘View Reports’. You can check the status of the installation. It will tell you if it’s successful, receiving data etc.

The reports page lists four main areas: Visitors, Map, Traffic Sources and Content. Each of these have a raft of separate areas you can access with further report data specific to that area.

Check out Sue Waters' basics too.

Ka kite