A-B testing

=Overview=

A/B Testing is a process by which companies can determine which design for a web page is better in relation to visitor experience and performance towards the business goals.

Two or more versions are created (one could be the current design) and served in equal numbers to the incoming visitors. A study of the behavior of the visitors who saw version A in comparison to those who saw version B (and D and E...) may provide insight into the preferred version or elements on the page. Thus, design improvements can be applied and subsequently re-tested against other new designs.

=ClickTale and A/B Testing=

Combining ClickTale with A/B Testing can be very powerful. You can use A/B tests to determine which page version performs better, and then use ClickTale’s reports to see which elements caused the winning version to perform better. This will then help you generate ideas for future improvements and tests.

=A/B testing with ClickTale=

You can perform A/B or multivariate testing with ClickTale in two ways:
 * 1. Using internal ClickTale tagging
 * 2. Using external testing tools with ClickTale integration

1. Testing using internal ClickTale tagging

 * If your tested versions are given different URLs then no tagging is needed. You can filter and compare by the different versions' URLs.


 * If your tested versions are on the same URL and differ only in URL parameters or cookie value, you should follow this procedure:

(I). Add a special HTML tag to the tested pages (in the  part of the page) to allow ClickTale to identify the different versions:


 * '''1 - for version 1
 * or
 * '''2 - for version 2

and so on for any other versions tested.

(II). Contact ClickTale support to set-up the correct page fetching for your test.


 * You will need to provide ClickTale Support with a query string parameter or cookie value to direct to the tested page, in order to force fetching the specific version of a test.


 * For example : http://www.somesite.com/AB_Test_PAGE?force_version=1

(III). After implementing the code, use the Event tags of the different versions to filter and compare visitor behavior on the versions you are testing.

2. Testing tools integrations
ClickTale works well with most A/B testing platforms. If your platform creates separate URLs for different page versions, then ClickTale will work out-of-the-box without any additional coding. If page versions do not receive separate URLs or you are conducting multivariate testing, you will most likely need to use Events, ClickTaleExcludeBlock or another integration.

ClickTale has integration documentation for the following platforms:
 * Google Content Experiments
 * Adobe Test & Target
 * Optimizely
 * Maxymiser
 * Visual Website Optimizer
 * Monetate
 * Convert