abstract art, a systems view

The Blog as system: a little Statistical Process Control

with 8 comments

Here’s a run chart showing the visits to my blog in May (I know, it would be nice to have more visits).

It shows at a glance just how much variation there is in the system visits per day to my site: although the average (mean) number of visits per day for May was 58, the highest number of visits was 144, and the least was 17.

Plotting the data in a control chart or capability chart (invented by Walter Shewhart and used by W. Edwards Deming) shows that the system is out of statistical control, in that there is special cause variation on day 29,

and the run of twelve days below the mean may also suggest special causes of variation (a run of six or more might be an indication of a special cause).

With special causes it could be meaningful to ask “what happened, specifically?”

Answers: 1) On day 29, I used a poll for the first time, and as it was researching a suggestion made by my son (that some people need help to see optical effects), both my sons were happy to encourage their Facebook friends to visit my site and complete the poll. As a result I got more visitors than usual that day. 2) On days 10 to 21, I may have been less active than usual in looking at other blogs as I was away for some of those days.

All the other data points show common cause variation: the variation that can be expected by the normal behaviour of the system. The chart shows that I could expect to get anywhere between 0 visits (the Lower Control Limit, LCL) and 112 visits (the Upper Control Limit, UCL) on any one day. To be surprised by data points within these limits, to get concerned for example at the 17 visits, would be foolish. To improve performance when the system shows common cause variation one must focus on the common causes rather than on individual data points. I could ask myself “what happens predictably every day, that causes this variation?” I would answer that I post something including a visual image, and that I take a few minutes to look at other blogs, mostly by tag surfing. To get more visits I would have to change this system.


Written by Andy Parkinson

June 9, 2011 at 7:10 am

8 Responses

Subscribe to comments with RSS.

  1. Specific statistics aside, I thing it’s a very fruitful to consider a ‘blog’ as a ‘(complex) system’, and not merely an autonomous agency.


    June 9, 2011 at 8:13 am

  2. […] You can find a good general introduction to SPC here thanks to the NHS in Scotland. Also Andy Parkinson has posted on the subject here. […]

  3. Very interesting post. May was my best month ever but some of our stats are very similar. My best day was around 150, and my low was 26. I also notice that I start and end each month with higher stats, and in the the middle of the month my stats drop off. I also notice that if there is a major event such as a natural disaster (Japan) or Royal wedding my stats drop. One exception was the Canadian Federal election, and I held a win a painting contest to see what would happen, and my stats increased. The next week I held a win a drawing contest and my stats fell. I wonder if this is because people are more interested in colour than in line? Other things I’ve noticed is the effect of weather on my stats. On sunny warm days fewer people seem to be online, and on cloudy rainy days my stats go up.
    My best days for views are Wednesday ( on this day I do a list of ten things I like), and Saturday, and on this day I usually post some of my daughters art. I notice that reviews are also popular. I recently did a beer review and I think it had more comments than any other post.
    Tools that I find useful to drive traffic to my blog are Facebook, Twitter, and lately Stumbleupon.


    June 9, 2011 at 1:35 pm

    • I love this. I’ve been thinking that the people who frequent my blog are 9-5ers because the traffic completely drops over the weekend. It’s good to know that my readers are most likely government workers who read my blog instead of work and then relax with family on the weekend. lol

      Love, Art & Fear

      June 10, 2011 at 7:04 pm

  4. Breaking new records is such a thrill. Stats rock!

    Edward Canavarro

    June 9, 2011 at 2:21 pm

  5. Nice to see run charts applied to something else other than manufacturing for a change. I wonder if anyone has mapped a ‘blog’ as a process? Great post, thanks.

    C Marsden

    August 4, 2011 at 9:35 pm

  6. […] have posted before about the blog as a system, and using Statistical Process Control (SPC) to ask questions about the visits, distinguishing […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: