JBidwatcher Metrics Details
Since March, 2012 JBidwatcher now includes the ability to collect anonymous statistics about how JBidwatcher is being used.
All the data is gathered anonymously, and can be disabled at any time if you wish.
What I will use the data for
With this enabled I can start to see better how and where JBidwatcher is being used; I am especially interested in what operating systems, screen resolutions, how often the application is started and stopped, how often snipes are placed, and won, and any errors that crop up during a given session with JBidwatcher. Also exceptional circumstances, like unknown bidding errors and the inability to parse certain critical fields (current price, seller name, high bidder, title, end date) are collected.
Over time I hope to learn more to continuously improve JBidwatcher and the related software and services I'm building out.
How to enable/disable usage statistics
On first startup, a dialog will pop up asking you to send anonymous usage statistics.
If you do not want to send anonymous usage statistics, press 'No'; the data will be flushed and not transmitted on program close.
If you close the dialog instead of making a choice, no usage statistics will be reported but the dialog will ask again on next startup.
If you press 'Yes' then usage statistics for JBidwatcher are enabled.
If you later on decide to not send statistics anymore or want to enable it to help, simply go to the General tab of the Configuration Manager. There is a configuration you can change to control whether or not data is reported. The configuration is more detailed than the default dialog, offering the additional choice of 'Pre-release only', which will only report usage statistics and errors when you are running a pre-release version of JBidwatcher.
How it works
When the user has enabled reporting anonymous statistics the software uses a custom library to report data to my own site, where I run a tool much like Google Analytics but for downloadable applications. The data is sent directly to my server, and processed there.
The information collected is similar in aggregate to the data collected by a typical site while web browsing, including IP address, operating system, version you are running, screen dimensions, start and stop times, plus errors encountered and significant actions. It is more specific to JBidwatcher in that the version number is tracked instead of your browser version. The library I'm using also generates a random id for the installation so multiple startups from the same JBidwatcher are treated as one user, but because the id is random it is not associated with you directly.
The source code to JBidwatcher, and the tracking code being used are available for inspection.
Thank you for your help in making JBidwatcher a better sniping and auction monitoring tool!