Split A/B test with a module on and off

[marcopolo]

I never did a split A/B test before on my website and would like to do one now testing a module being on and off, is this possible?

[kobra]

It might depend upon the module - a bit more information about what you are attempting might be helpful

[marcopolo]

The module can be defined with either true for on or false for off.

[kobra]

What module is it?
What does it do?
Where did you get it from?
etc

[marcopolo]

It's Dynamic Price Updater: https://www.zen-cart.com/downloads.php?do=file&id=1301
I want to A/B test it and see how it affect sales.

[kobra]

To make this switchable say from the admin would take a complete re-write

I suggest that you collect your base data without it
Install it and collect your data
Then if you want uninstall the files and collect some more data

[marcopolo]

Thank you for letting me know.

[mc12345678]

There is a plugin that was generated to support a/b template testing at the insistance of another zenner. Considering that the plugin is more than likely template bound, you could duplicate the template to a second name and in one template set the module be installed and in the other it be removed. For any core affected files, a. Test could be added comparing the current session's template against the existence of the/a file and process or bypass that code section as necessary to support testing...

[marcopolo]

There is a plugin that was generated to support a/b template testing at the insistance of another zenner. Considering that the plugin is more than likely template bound, you could duplicate the template to a second name and in one template set the module be installed and in the other it be removed. For any core affected files, a. Test could be added comparing the current session's template against the existence of the/a file and process or bypass that code section as necessary to support testing...

I knew there was a reason why I never did A/B testing

[mc12345678]

I knew there was a reason why I never did A/B testing

Hard to gauge the level of facetiousness, but consider the "normal" way to do A/B testing... Create two different web pages that are to be provided randomly per visitor... Sounds easy, right? What's so hard about reproducing a single webpage or perhaps for that matter an entire site?

It takes some work because afterall, trying to basically setup two different sites each based off of some standard/consistent information.

There's a plugin that shows all the possible locations for a template directory, and since the test is related to a plugin, the plugin install files are already available to identify which areas need to be bypassed or not... To me, if desired to be tested, then possible to be done and a means has been provided. :)