Statistical Improbabilities
Back in December, I wrote about the statistics issues Quick Blog is facing and concluded by saying that "you can think of your statistics as gone but not forgotten; soon you will be reunited." I clearly had an optimistic view of "soon" in that entry as we still have not reunited the statistics. I am truly sorry for how long this has taken and for the wonkiness that has been occurring as we start back up the efforts.
Some of you may have noticed that your statistics have been decreasing throughout the last day or so. Don't worry, it's not some sort of anti-entry request that negates an existing one. We're culling out data that should have been removed at the time of that blog entry but wasn't. That should be done by tomorrow and things will at least be stabilized. The data is being removed because we've already archived it and we don't want to have duplication at this point.
"So what's the plan from here?" you're likely wondering. Back in January, I developed and ran an application that aggregated that data into a much smaller database. That data is sitting unused right now and the next task is to modify Quick Blog so it integrates that data in with the current data. Once that's done, you'll have a complete picture of your blog's activity and it will be as if the December archiving never happened. After that, we will run that application I mentioned every night to aggregate the day's worth of requests.
(Incidentally, this aggregation will reduce the number of records we need to store in our database by a factor of 100. This should make statistics much more responsive and we'll be able to do neat things in the future.)
So again I conclude with nothing but gratitude for your patience and understanding. We Quick Blog developers are also users of the product and the statistics issue affects us as well. I promise to keep you, our dear customers, informed of our progress on this issue through Quick Blogging.
Bill, I am thouroughly enjoying using Quick Blog and wanted to invite you to drop by http://activerain.com/groups/quickblog to visit the group of real estate agent Qucik Blog user we have started.
Wow, that's really great! I just joined so I can keep track of the interesting things you guys are doing and will do.
Glad you like it, Bill
I have been noticing that adwords statistics differ siginificantly from Quick Blog´s number of articles viewed.
Maybe, they are not following the same count pattern. Will you inform me a little bit on this?
Thank you,
That's an excellent question. Unfortunately, it's impossible for us to compare the differing statistics. They count their statistics based on a Javascript inclusion of the ads served. We derive our statistics during the actual request so ours should be more reliable because their method could be thwarted by any number of means. If you tell me a little more about what you're seeing, I can possibly address it more specifically.
Thanks for your question, Bill
I assume you were already aware of these count differences. aren´t
you?
Specifically, yesterday I had 19 articles read according to Quick
Blog statistics, as opposed to just 2 visits as shown in adwords.
It is quite a significant gap.
There could another explanation: I have not placed correctly the
Google Java Script.
Thank you
Juan
That may be significant, it may not be. It's hard to know specifically. For example, one of my blogs had 3,890 page impressions so far this month according Google's AdSense reports. According to Google Analytics, there were 5,443 page views. But Analog reports 26,337 page views for that time period. Which is right? In reality, none of them. They're all right for as far as they go, but they all miss the mark.
The difference is that Google's AdSense is Javascript-based. If the Javascript library isn't retrieved from Google, the ad isn't displayed and the page is not counted (that's also how Google Analytics works). There are countless reasons why the library might not be loaded: request timeouts, search engine bots not retrieving Javascript libraries on purpose, browser ad blockers, or browsers with Javascript turned off.
Personally, I would trust Quick Blog's reporting of 19 articles read because the request is logged once the page is displayed without any work required by the visitor's browser. Any kind of error results in no request being logged.
Hope that helps a bit, Bill
Just checking to see how blogs functions. Bill, I find your comments about Google adwords fascinating.
Hey thanks! I try to be helpful.