Split testing

A/B testing of form variants.

Using variants to test changes

A common pattern is to use variants within Formsort, which each can have a unique URL, together with an upstream testing tool, such as Optimizely, to track experiments. Using a third-party experimentation platform allows for experiments that span across Formsort and upstream or downstream parts of the user journey.

Randomly serving a form variant

Flows have their own URLs, and can be pointed to a custom domain, which does not require a variant to be specified.
For quick experiments within Formsort itself, you can assign weights to individual variants from the flow index page.
If a flow has two variants, and variant A has a weight of 50 and variant B has a weight of 50, then when the flow URL is loaded, half of the responders will be sent to variant A and the other half to variant B.
To generate a pseudo-random variant for a given form, simply do not specific the variant in the url string. Rather, just specify the form, e.g.<example_client_name>/flow/<example_flow_title>/
To split traffic between two variants, both variants must be published and the variants dropdown in the domains tab must not be set to a specific variant. And, the two variants must have the appropriate weight set (such as 50/50 or 90/10), and any other published variants have their weight set to zero.
Returning responders will reload the first variant they initially loaded if Returning responder behavior is left to Start at furthest point. To test if traffic splitting is working make sure to use incognito mode.