Reflections: Random Generator Pageview Tracking

During December, I modified all of my random generation processes to track when results were generation. I wanted to determine which generators were used more frequently so I could enhance the popular ones and rework those with limited interest or just toss into the bit-bucket. Like many people, I use Google Analytics as one of the core metrics for site statistics.

Since December was the cut-over month, I chose the periods just prior and just after too look at the overall results. The early period includes the months of October and November (61 days). The latter period covers January and February (60 days). Both periods include a major holiday. Generator initial page views were tracked in Oct/Nov but only for the initial page view. Jan/Feb includes stats for every page generated as well as the initial page view.

The site wide statistics are quite revealing:

Oct/Nov: 68% Bounce Rate, 5360 Page views, 1.93 pages/visit, 1:45 minutes on site, and 73% of visitors were new.

Jan/Feb: 31% Bounce Rate, 70732 Page views, 7.26 pages/visit, 5:26 minutes on site, and 79% of visitors were new.

The overall page views increased 10-fold. Some of that is due to natural growth. Mostly, the analytics reports just didn’t capture the multiple generator page loads due to the nature of my coding approach. The raw web server statistics indicated those loads but it is nice to have the information in a singular location.

The reduction in bounce rate and increase in retention time also told me the generators were as useful as I expected. Nice to see my assumptions were valid. Now onward to 2K pageviews per day. Such a lofty goal is going to take some time.